Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Proscriptive features of unilateral cross-cousin marriage Kobrinsky, Vernon Harris 1967

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-UBC_1967_A8 K62.pdf [ 7.73MB ]
JSON: 831-1.0104552.json
JSON-LD: 831-1.0104552-ld.json
RDF/XML (Pretty): 831-1.0104552-rdf.xml
RDF/JSON: 831-1.0104552-rdf.json
Turtle: 831-1.0104552-turtle.txt
N-Triples: 831-1.0104552-rdf-ntriples.txt
Original Record: 831-1.0104552-source.json
Full Text

Full Text

PROSCRIPTIVE FEATURES OF UNILATERAL CROSS-COUSIN MARRIAGE by VERNON HARRIS KOBRINSKY B.A., University of Manitoba, 1956 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in the Department of Anthropology and Sociology We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA In p re sen t i ng t h i s t h e s i s in p a r t i a l f u l f i l m e n t of the requirements f o r an advanced degree at the U n i v e r s i t y of B r i t i s h Columbia, I agree that tha L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r reference and study. I f u r t h e r agree that permiss ion f o r ex ten s i ve copying of t h i s t h e s i s f o r s c h o l a r l y purposes may be granted by the Head of my Department or by h i s representat ives. . It i s understood that copying o r p u b l i c a t i o n of t h i s t h e s i s f o r f i n a n c i a l gain s h a l l not be a l lowed without my w r i t t e n pe rmi s s i on . Department of t^Mrygv&r^ The U n i v e r s i t y o f B r i t i s h Columbia Vancouver 8, Canada ABSTRACT This essay offers an analysis of the phenomena of normative matrilateral and patrilateral cross-cousin marriage. The review of prior literature is confined to studies which appear to be significantly opposed in approach to my own; the "exchange theories" of Levi-Strauss, Leach and Needham. Therefore, the review is conducted in the context of a discussion of some epistemological questions important to contem-porary social anthropology. The major points of departure between the exchange theories and my own views center upon the question as to what we are to take as our analytic units: d_e facto corporate units or idealogically explicit social divisions. Whereas I opt for the former, explanatory propositions employ a concept of whole "groups" as their subject element only in an e l l i p t i c a l sense; "group" means a set of human actors. This view im-plies that, where human actors are our subject-element, i t is not merely reasonable but is desirable to predicate phenomena of a motivational nature to them. Thus, both conscious decision-making and unconscious dispositions, admittedly in some sense "psychological", are valid forms of sociological analysis. Indeed, to my way of thinking they are the most powerful conceptions that we can presently form. In the last analysis I regard the predications of the exchange theories to be of precisely this order. My complaint is that this mode of predication is inconsistent with the confinement of our analytic subject-element to idealogical social divisions: whole-descent-groups. That approach, I argue, severely delimits our explanatory powers, es-pecially in a context of structural change. These arguments are i l l u s -trated in an analysis of Purum data. My own models are based upon a series of concepts having to do with processes of role identification between personnel of adjacent generations. These processes, though not verbalized, find formal ex-pression in the distribution of Ego's sexual cum marital privileges and taboos. This encompasses Ego's relations with many personnel, thereby extending the implications of the models' theoretical premises considerably beyond the cousin relationships taken as the i n i t i a l prob-lem. This extension in the scope of deductions permits the formulation of empirically testable hypotheses of value in the verification of the theoretical premises. The concepts and methods developed are illustrated and tested I think successfully against Murngin data. TABLE OF CONTENTS Page Introduction 1 Section I: Critique of the Exchange Theory Chapter 1: Cross-Cousin Marriage: Prescriptive vs. Preferential 8 Chapter 2: Explanation in Structural Analysis 21 Chapter 3 : The Exchange Model of Purum Society: Salient Issues 59 Chapter 4: An Alternative View of Purum Society 73 Section I I : Counter-Proposals Chapter 5: Basic Concepts: Continuity and Trusteeship 87 Chapter 6: Exogamy and Descent 102 Chapter 7: Matrilateral Cross-Cousin Marriage 119 Chapter 8: Murngin Kinship: An Illustration 139 Chapter 9: Patrilateral Cross-Cousin Marriage 157 Conclusions 162 References Cited 169 Some Additional References Footnotes 171 173 LIST OF FIGURES Page Figure 1 10 Figure 2 27 Figure 3 28 Figure 4 35 Figure 5 35 Figure 6 37 Figure 7 38 Figure 8 52 Figure 9 74 Figure 10 124 Figure 11 127 Figure 12 131 Figure 13 133 Figure 14 134 Figure 15 136 Figure 16 143 Figure 17 144 Figure 18 145 Figure 19 146 Figure 20 147 Figure 21 159 - 1 -INTRODUCTION Institutionalized cross-cousin marriage in i t s various forms has for some time been a focus of theoretical controversy. Currently at the forefront among contributors to the subject are Rodney Needham, Edmund Leach, and Claude Levi-Strauss. The theo-ries advanced by these scholars are fundamentally alike. For ease of future reference, therefore, I shall frequently cite them under the single label group-exchange theory (or simply exchange theory)* Where important differences exist among them, they w i l l be noted. Needham has probably been, at least i n recent ys&rs, th& moet ener-getic advocate of this theory. The exchange theorists have been c r i t i c i s e d by various writers on aspects of their analysis. However, although there have been a number of interesting suggestions offered by their c r i t i c s , none, with the exception of Homans and Schneider, have offered a well developed counter-proposal. The dialogue between the exchange theorists and their c r i t i c s has had two aspects. There has been mutual criticism of substantive issues; i.e., of theoretical and empirical issues con-cerned specifically with the phenomena of cross^cousin marriage* Charles Ackerman, for example, has attempted to show that certain entailments of Needham's model of Purum social structure are not 1 borne out i n s t a t i s t i c a l analysis of actual Purum marriages. Needham has pointed out certain flaws i n the logic of the Homans 2 and Schneider model. - 2 -In addition, adherents to both sides of the dialogue have regarded their respective views as incorporating a fundamental op-position in the epistemology of sociological analysis. The ex-change theorists on the one hand argue that the psychological ap-proach of Homans and Schneider is unsuited to the analysis of institutionalized behaviour; dimensions of individual personality have no bearing on group norms. Homans and Schneider argue that the exchange model requires the assumption of a foresightful social designer. The primary aim of this essay is to offer another analysis of the unilateral forms of cross-cousin marriage. As such the emphasis is upon substantive issues. However, Chapter TWO pro-vides a discussion of what I regard as the main requirements of theories of social norms. Most of the epistemological issues in the chapter are illustrated from literature on theories of cross-cousin marriage. Thus, together with Chapters One and Three, Chapter Two contains an examination of some of the more salient aspects of the exchange theory. Chapters one and three in particular examine some of the implications of Needham's distinction between prescriptive and merely preferential marriage rules. Prescriptive rules as he views them are essentially a means of regulating the status rela-tions between entire social groupings. This interpretation requires that the social system provide mechanisms for equating the member-ship of such groupings with the appropriate kin-term categories. - 3 -In Chapter Three certain empirical implications of the exchange interpretation are examined against Das' : data on the Purum. This analysis causes me to doubt the adequacy of Needham's approach. The examination reveals certain inconsistencies between Das'v findings and Needham's model.. Briefly stated, these in-consistencies derive from the idea that Ego places other persons in various kin-term categories on the basis of their membership in particular explicitly recognized whole descent groups. In Chapter Four, Purum data are re-interpreted from the point of view that the social groupings relevant to Ego's alloca-tions of kin-terms are narrower than the named Purum descent groups. To be sure, the structure of the social groupings which compose a society is a decisive factor in Ego's allocations of kin-terms to the people surrounding him. Further, the structure of the rele-vant groupings is regarded as changeable from time to time as well as cross-culturally variable. Hypothetically this approach may be taken to a point at which Ego's allocations of kin-terms are based upon hi8 location in an orientation group reduced to scarcely more persons than his own parents. In such a situation, the matrilateral cross-cousin marriage rule can scarcely be regarded as a charter to status relations among ideological descent groups. In brief, the relations of whole descent groups and the matrila-teral cross-cousin marriage rule are treated as contingent to one another. These considerations represent a two-fold departure from Needham's approach. The orientation to analysis in terms of struc-- 4 -turally relevant social groupings which need not correspond to the culturally normative groupings introduces de facto elements into models of societies having a matrilateral cross-cousin mar-riage rule. This results in a model of mixed composition ( u t i l i -zing both statistical and ideological norms) as distinguished from Needham's more purely mechanical model (ideological elements only). The characteristics and relative merits of mechanical vs. statis-tical vs. mixed models are discussed in the first section of the essay. Furthermore, the view that understanding the marriage rule may be achieved even while treating the normative group struc-ture of the society as a contingent variable would appear to lead our quest for an explanation of the marriage rule in the direction of what the exchange theorists might term psychological reduction-ism. If we do not explain the marriage rule in terms of the be-haviour of or interests of groups taken as a whole, what entities are to serve as our analytic units? This issue is also discussed in the first section. Suffice to say, for the present, that to eliminate whole groups as analytic units is not, as the exchange theorists appear to believe, tantamount to isolating persons, as analytic units, from an envirnoment of social relations with other persons. The analysis of social relations among persons does not reduce to a theory of personality. It is by no means my aim to discredit the brilliant contributions of the exchange theorists. Rather, I hope merely - 5 -to show in the first section of the essay that on some important points there is sufficient doubt that the new approach to the problem of unilateral cross-cousin marriage offered in section two is not a wasted effort. In Chapters Seven and Nine respectively models are pro-posed for matrilateral and patrilateral cross-cousin marriage sys-tems. The models are based upon a few rather simple principles. These principles, formulated in Chapter Five, are concerned with dimensions of social interaction which surround the regulation of heterosexual behaviour. Marriage rules per se are treated as a man-ifestation of the same factors which ex hypothesi structure sexual behaviour. The regulation of sexual behaviour, in turn, is viewed as deriving from a prepotent disposition among the peoples with whom the analysis is concerned to reproduce themselves as social beings in persons of the generation which succeeds them. This tendency is referred to as a process of identification or role-continuity. In sum, then, the basic propositions of the theory concern a process of inter-generational role continuity, parti-cularly as the process is manifest in the regulation of hetero-sexual behaviour. I must emphasize at the outset that my own concern is with a somewhat different range of social phenomena from those for which the exchange theory is designed. Needham confines his analysis to so-called prescriptive systems, This essay, on the - 6 -other hand, is an attempt to understand important features of pre* ferential rules as well as of prescriptive ones. In the exchange analysis of systems of prescriptive matrilateral cross-cousin marriage the characteristic sexual and marital prohibitions are treated as complementary to the prescrip-tive aspect of the rule: a prescription to one social category directly implies marital prohibition of the remaining ones. A merely preferential rule, however, cannot be handled with the same relative simplicity. A normative marriage preference for a given social category scarcely provides an immediate understanding of the strict prohibition of other specific categories. In this expan-ded universe of discourse the marital prohibitions become events which require a more particular understanding. Thus, the primary focus in this essay is upon the pro-scriptive regulations which comprise a definitive aspect of any system of unilateral cross-cousin marriage. My efforts in handling the positive features of the systems -- prescriptions and preferen-ces — are less satisfactory. This limitation becomes apparent in Chapter Eight which offers an analysis of Murngin kinship based upon the abstract model provided in Chapter Seven. The continuity con-cepts are able to render an explanation of the specific distribution of Ego's sexual privileges and taboos among the numerous female kin-term categories which surround him. They do not account, however for the restriction of his marital rights to only a single one (the matrilateral cross-cousin) among the several terminological - 7 -categories of women with whom a Murngin man is in fact allowed to conduct sexual relations. Chapter Six is somewhat incidental to the mainstream of the essay. However, i t provides another illustration of the pos-sibilities offered by a psychological approach in the analysis of norm systems. Rules of exogamy are viewed as a mechanism for limiting the development of strains in the relations of men who must guard a close mutual trust. Throughout the text I have used a consistent kin-status notation system. The abbreviations are as follows: father = Fa, mother = Mo, brother = Bro, sister = Si, son = Son, daughter = Da, wife - Wi, husband - Hu, elder - a prefixed e, younger - a pre-fixed y. Where a different notation system has been used within a quotation, I have taken the liberty of altering i t to this one. In such cases, squared brackets indicate the alteration. Native kin-terms have a l l letters capitalized, except where they occur within quotations and are otherwise distinguished there. - 8 -CHAPTER 1 Cross-Cousin Marriage: Prescriptive vs. Preferential Anyone familiar with the works of Levi-Strauss, Needham and Leach is well aware of the virtually axiomatic position in their theory of some conception of a corporately acting social group. In connection with the problems of institutionalized mar-riage rules this posture is clearly illustrated in the following statements: First Levi-Strauss, The preferred marriage /matrilateral c.c. marriage among the Kachin7 results, not so much from a prescribed and precise degree of relationship, but from a general rela-tionship between a l l the men of a lineage . . . and /ano-ther/ lineage in respect of a l l of its daughters and sisters.3 and Needham, These alliances through prescribed marriage are estab-lished between corporate groups, specifically lineal descent groups, not merely between individuals, and not between individuals standing in any particular genealo-gical relationship.^ or, again, The relationships of affinity are not merely ties be-tween individuals or families. Descent groups,whether localized or dispersed, are also related as groups by ties of prescriptive alliance.5 and finally Leach, There are two kinds of marriage. The first results from the whims of two persons acting as private individuals; the second is a systematically organized affair which forms a part of a series of contractual obligations between two social groups.** Leach considers the second type of marriage to apply to societies practicing matrilateral cross-cousin marriage. - 9 -In their developments of the foregoing statements these writers reveal one or two important divergences. They do not con-cern us at present. On the other hand, each clearly posits the existence of some form of social grouping which acts somehow cor-porately in the marriages of its individual members. The analysis of social institutions in terms of some conception of a corporately acting social unit is a defining characteristic of structural anth-ropology. As such the concept figures vitally in the exchange theories of prescriptive marriage systems. For the most part, however, the studies with which I am familiar make no effort to provide an empirical justification of their idea of the group. They are concerned rather with elabora-tion of the various effects or implications of this, that or the other form of marriage upon the relationships of the given social 7 groupings. The pursuit of these implications introduces further concepts, noteably the ideas of reciprocity, status and solidarity. However, I consider that the quasi-axicmatic position of social groups in the exchange theories is itself a giant step. I shall devote much of this section to examining its methodology in addition to some consideration of the further concepts named above. An important exception to the lack of empirical consid-eration of the group in a society with a prescriptive marriage rile 8,9 appears in Needham's analysis of Das' data on the Purum tribe - 10 -10 of Manipur. Needham employs Das' data in an effort to show that, among the Purum, most individual marriage choices do in fact add up to movements of women, as between whole descent groups, which parallel the mode of movement (exchange) held to be implied by a rule of prescriptive matrilateral cross-cousin marriage. His effort has stimulated a spate of controversy centered upon the empirical features of this question. Figure 1, taken from Needham, represents the "formal fea-11 tures of prescribed matrilateral cross-cousin marriage." It pre-sents diagramatically what Needham views as the structural impli-cations of a prescriptive (as opposed to preferential) matrilateral cross-cousin marriage rule. C B A r O A = 0 A r i -O A •O T ° A = 0 O A = O A : Figure 1 The meanirg of this diagram, to Needham, i s contained i n the f o l -lowing statements: The central fact in a matrilateral system i s that lines are related fey the same kind of relation from instance to instance and from generation to generation. Viewed as a system for the communication, transfer, or "exchange" of women, a l l the transactions are made in the same - 11 -"direction". Three lines satisfy the structural require-ments of such a system, and the model entails that the movement of women is cyclic. Line A gives women to B, B gives to C, and C in turn gives to A, thus closing the cycle.12 The "lines" referred to are also termed "structural groups": Ego's structural group, the group of "wife-givers" and the group of "wife-takers". He continues: In reality, each of these structural groups may comprise a number of corporate unilineal groups as defined by the particular society: these are descent groups. A descent group of any order may comprise a number of dispersed local groups.!3 A central implication of these statements appears to be that, as between whole descent groups, here depicted as lines or as components of lines, women move in one and only one direc-tion. A corollary to this i s the structural prohibition of sister exchange ("sister" being understood in the classificatory sense). Even Needham's most ardent opponents seem agreed on this point. For example, Charles Ackerman states: When there i s a significant tendency to avoid the direct exchange of women among the descent groups, the bride-giving and bride-taking relationships are ordered by a matrilateral connubium.^ The avoidance of sister exchange between alliance groups in actual Purum marriages has been treated as a crucial "test case" of Need-ham's model. To be sure, Needham himself and some of his suppor-ters appear ambivalent about this issue. It w i l l be taken up later. To accept avoidance of sister-exchange between alliance groups as an implication of a prescriptive rule of matrilateral - 12 -cross-cousin marriage (hereafter the word "cross-cousin" w i l l be abbreviated to "c.c") i s , of course, to accept whatever premises support Needham*s model. If only to fu l l y understand Needham's concepts I shall now attempt to make explicit some of these pre-mises. For purposes of this explication I shall employ a dialec-t i c a l approach; i.e., together with presenting some of Needham's own statements of his premises, I shall offer what seem to me logically reasonable alternatives (or dia l e c t i c a l opposites). The question of the empirical reasonableness of these opposed views, and of their reconciliation, w i l l be taken up later. In a most excellent review of Needham's monograph "Struc-ture and Sentiment", Floydd Lounsbury (A.A., 64, pp. 1302-1310) has considered at length some of the semantic d i f f i c u l t i e s in Needham's use of the word prescriptive. Lounsbury's criticisms need not be repeated here. I shall be satisfied to spell out to the best of my own understanding what Needham means by the notion of a prescriptive rule of matrilateral c.c. marriage. To begin with Needham's own words: The term "prescriptive", on the other hand, /as opposed to "preferentially has quite different connotations. In this case the emphasis i s on the very lack of choice: the category or type of person to be married i s precisely determined, and this marriage i s obligatory. Among the Batak of Sumatra, for instance, marriage i s prescribed with a woman exclusively of the category boru n i tulang, of which one of the genealogical specifications i s "mo-ther's brother's daughter". Contrarily, a woman of the category boru n i namboru, of which one of the specifica-tions i s "father^s steer's daughter", may absolutely not be married. This situation i s characterized by an utter lack of categorical choice.15 - 13 -Needham has repeatedly insisted that a prescriptive system i s to be understood as a system of social classification. Our task, as he views i t is the "comprehension of the l i f e of a society in terms of the classification employed by the people themselves, and 16 an analysis i n terms of relations of the widest generality." Thus, in a system of prescriptive matrilateral c.c. marriage every male Ego divides the women of his generation into (at least) three exclusive categories; sisters, mother's brother's daughters, and father's sister's daughters. These categories cor-respond to the three structural groups mentioned earlier (vide  supra, p. 11). Prototypically, each of these categories of women i s set apart by a distinct term of address and/or reference. This is not always precisely the case, however, and differentiation among the three categories i s sometimes accomplished only through the additional support, to the terminology, of further reference to descent group (or section) membership. Among the Purum, for example, we find a terminological equation between ySi and MoBroDa (KA-NAU-NU). This, however,-«*as Needham views it--presents no classificatory problem to the Purum since (a) ySi (=KA-NAU-NU) i s also referred to as a KA-SARR-NU which "appears to be i n fact the common Kuki-Chin term (sar, far) for siter in general (cf. Shafer 1952: 267-8, s.v. "Schwester")". A MoBroDa, who i s referred to as KA-NAU-NU, would not carry the additional label of KA-SARR-NU. (b) Also, "we have in this case both ySi and MoBroDa denoted by nau, /abbreviation of KA-NAU-NU7 - 14 -but there s t i l l need be no practical difficulty, for the former is a member of Ego's own descent group and is thus prohibited, while the latter is not." (c) Finally, "there is s t i l l the fact that the eSi is known as u whereas the /MoBroDa/ is not (Needham 1962 a: 78)" Thus, a maritally proscribed KA-NAU-NU is distinguished from a mari-tally prescribed KA-NAU-NU by the fact that the former has a sis-ter, older than Ego, whom Ego terms KA-U whereas the latter has a sister, older than Ego, whom he terms KA-NAU-NU. In passing, we should note that "A man addresses a woman of his sisters' husbands' 18 group of sibs as Katunu (if elder) or by name i f younger". I presume that this statement provides the basis for Needham's inclu-sion of the FaSiDa under the term KATUNU (Needham, 1962, p. 76, Table 4 and 1964, p. 1378). Thus, there is no single term in the Purum lexicon of kin-terms which may be directly translated maritally prescribed woman. Using Needham's information, one would characterize a woman of the prescriptive category as any woman whom Ego terms KA-NAU-NU, as long as she is "not of Ego's own descent group", and/or as long as she is not included under the more general term KA-SARR-NU, and/or as long as she has no sister, older than Ego, whom he terms KA-U. It has already been stated that a prescriptive marriage system is primarily a system of social relational classification which is "employed by the people themselves" (vide supra, p.13). Such a prescription may be expressed in the system of kinship - 15 -terminology. This need not be the case, however, as in situations where marriage prescriptions take the form of prescription to either specified descent groups or to named marriage sections. In these cases the structural group to which a man's choice is prescribed is not designated by kin-terms as such, but by the appropriate descent group or sectional name (kin terms for women will cross-cut the prescribed structural group). Where the marriage prescription is expressed in the system of terminology, the kin terms must provide for a demarca-tion, in the mind of the cultural participant, of at least three classes of women (in the case of matrilateral c.c. marriage sys-tems; only two classes of women are required in systems of bilateral c.c. marriage, i.e., sisters vs. bilateral cross-cousins) of Ego's generation. The distinction of these categories, as they appear in the terminology, may be achieved either through single terms, each sufficient by itself to denote an exclusive social category or (as in Needham's argument above regarding Purum terminology) through an interaction of terms. Finally, Ego must be morally restricted to selecting a marriage partner from one and only one of these terminological categories. A merely preferential system is the same in a l l respects save that Ego is not rigidly confined by morality to only one of the categories, but merely prefers to choose his bride from i t . - 16 -Perhaps these remarks will become clearer i f we look at some of Needham.*s characterizations of particular cases: Karadjeri: Not prescriptive in the sense required. They have a four section system, and although marriage with the genealogical father's sister's daughter is forbidden they marry categorically a bilateral cross ecusin.19 or, again, Wik-Munken (Archer River group): Not prescriptive in the sense required. A straightforward two section system with exogamous moieties, plus differentiation of senior and junior affines. The genealogical /FaSiDaT is forbid-den, but marriage is categorically with a bilateral cross-cousin.^ o According to Needham then, the Karadjeri and Wik-Munkan are not prescriptive "in the sense required". I take him to mean by this that such systems are indeed prescriptive ones, but not the kind of prescriptive system he is concerned with. Karadjeri and Wik-Munken prescriptions are to named sections; i.e., a given man is morally restricted to choosing a wife from a single named sec-tion (and morally proscribed from choosing one from the one or three remaining sections). Thus, while these are prescriptive sys-tems they are not unilateral (in these cases matrilateral) cross-cousin marriage prescriptions. This is simply because the kin terminology in both cases puts women of Ego's generation in only two terminological categories, rather than in three categories. In these cases, the two categories are identified as sisters, who are proscribed, and bilateral cross-cousins who are not proscribed. The genealogical FaSiDa and MoBroDa both f a l l within the same single kin-term category. Thus where i t is observed that among the Karadjeri - 17 -"marriage i s allowed with own mother's brother's daughter but not 21 with father's sister's daughter", one must assume that the FaSiDa here referred to i s confined to the l i t e r a l genealogical FaSiDa. In terms of Needham's stipulated meaning of the word, this does not make a prescriptive system of matrilateral c.c. marriage since the proscription against FaSiDa does not mean the proscription of an entire exclusive category of women, which merely includes the gen-ealogical FaSiDa, and prescription to a distinct terminological category which includes the genealogical own MoBroDa. In passing i t should be observed that i t i s unclear whether the Karadjeri should be counted as sectionally prescriptive, term-inologically prescriptive (to the bilateral cross-cousin category), or both. It i s possible in a four-section system that the termino-logical category of bilateral cross-cousin may not denote uniquely the category of women to which Ego i s confined in his marriage; i.e., bilateral cross-cousins may be distributed in more sections than the single one to which Ego is prescribed. a named section may not, then, be tantamount to a terminological prescription. It i s not at a l l clear to me what obtains among the Karadjeri in this respect. Indeed, i t is not even clear that they 22 have a four-section marriage system. The Lanes count six. A general summation of our discussion of the concepts prescriptive/preferential contains three major points: (1) A prescriptive system i s an ideology of social classification. Exclusive categories of people are in some way named by the culture participants. - 18 -( 2 ) We may, for present purposes, recognize three v a r i e t i e s of p r e s c r i p t i o n . These d i f f e r i n the conceptual mechanisms which provide the basis f o r crea t i n g the various exclusive and named categories of people. The three types are: (a) p r e s c r i p t i o n to named marriage sections. We should remark, i n passing, that when dealing with i m p l i c i t section systems we w i l l have, from t h i s viewpoint, descended from the realm of c u l t u r a l ideology per se to modeling a mix-ture of i d e o l o g i c a l and de facto marriage patterns. Where i t i s demonstrated that the i d e o l o g i c a l elements of such a combined model represent f a c t u a l marriage behaviour as w e l l , the model as a whole must be regarded as a representation of actual marriage p r a c t i c e s . (b) p r e s c r i p t i o n to women of named descent groups. (c) p r e s c r i p t i o n to the women of a single kin-term category; for example, p r e s c r i p t i v e m a t r i l a t e r a l , p a t r i l a t e r a l or b i l a t e r a l cross-cousin marriage. Combinations of the above three forms of p r e s c r i p t i o n may coincide i n a p a r t i c u l a r s o c i a l system. The major point of Needham's Purum analysis concerns the coincidence of p r e s c r i p -tions of (b)/(c) v a r i e t y . His theory attempts to explain a r u l e of p r e s c r i p t i v e m a t r i l a t e r a l cross-cousin marriage (a type (c) pr e s c r i p t i o n ) i n terms of type (b) p r e s c r i p t i o n s . (3) We must bear i n mind that the pr e s c r i p t i o n s and pro s c r i p t i o n s both carry moral injunc t i o n s . I t i s morally c o r r e c t to marry - 19 -a woman of one category and morally wrong to marry a woman of another. It is not really clear whether preferential systems diffe r from prescriptive ones in this regard—other than as an issue of degree. The difference seems to center about the idea that no serious moral offence i s committed by marrying into a non-preferred category. Rather, " i t i s simply thought a good thing that a man should marry, i f conveniently possible, 23 a woman in a certain position." The "position" may be stipulated through genealogical description, a terminological category, reference to a descent group, etc. Let us accept, then, that Needham's distinguishing criterion centers, above a l l , in the absence of a s t r i c t moral proscription, in pre-ferential systems, against marriage into a category whose counterpart in a prescriptive system i s morally forbidden. Like Needham's, the views which I present later in this essay are an attempt to explain rules of matrilateral c.c. marriage. An important difference, however, appears i n our respective stipu-lations concerning the situations which we accept as cases of these rules. For my part, I include situations, some of which Needham might class as preferential, in which there i s a significant ten-dency shown to frown upon marriage with a FaSiDa (and/or women of her distinctive terminological category) even i f they are willing to reluctantly tolerate i t . At least I am inclined to include such cases as possible examples of my argument. In addition, I •., might include societies, such as the Karadjeri and Wik-Munken, which - 20 -apparently proscribe the genealogical FaSiDa, but not the term category to which she belongs (which includes the genealogical MoBroDa). We are alike, though, in that I also am concerned with social systems that would qualify clearly as systems of prescrip-tive matrilateral c.c. marriage. I have taken some pains to present as clearly as I am able Needham's use of the concepts prescriptive/preferential. The distinction is a central one to his arguments: Prescriptive rules have certain structural entailments, and societies with such rules can be classified accord-ing to the particular entailments of the different pre-scriptions. 24 whereas, Marriage preferences, on the other hand, have no struc-tural entailments in the total social system comparable to those of a prescriptive system, and are not systema-tically integrated with the marriage prohibitions.^ - 21 -CHAPTER 2  Explanation in Structural Analysis The crucial feature of prescriptive marriage rules in 26 general is that they "entail enduring affinal ties between groups", and of prescriptive matrilateral c.c. marriage rules in particular, that "the ties are formed by the unilateral transference of women 27 from group to group. . ." This concept, which has already been introduced (vide supra, pp. 10-11), will require closer analysis. Especial difficulty, i t seems to me, derives from the connotation of the word group. Needham's meaning is somewhat clarified in a statement made in reply to criticisms put Livingstone (F.B. Livingstone, 1959). He states: A social system is an abstraction relating (in this con-text) to lineal descent groups which are also an abstrac-tion. There is no place in abstraction for substantive "specific groups". It would be an odd and profitless notion of social system to so identify i t with substan-tive reality that every change (segmentation, even births or deaths of individuals) would be said to con-stitute a "breakdown". This statement echoes the form vs. matter, mind vs. body, and persistence vs. change dilemmas which troubled the thought of an-cient Greece. Our purpose is to avoid such dilemmas in 20th Century science. I hope to show how this might be accomplished with the problem in hand. I am prepared to accept the notion of groups as an ab-straction. However, the abstract groupings dealt with in a parti-cular situation must, in the context of Needham's ideas, be ones - 2 2 -that are nevertheless real at least as a part of the ideology of the culture participants themselves. The groups must be perceived, named or not, by the culture participants. If they are not par-ticularly named the anthropologist must be prepared to provide evidence that the group is treated as an exclusive societal divi-sion through some designation occurring in the actions or conver-sations of the culture participants; e.g., designations such as the people who live--(place indicated) or the people who (ac-tivity stereotype), etc. The substantive content of such ideological divisions will doubtless undergo continuous flux--population and composi-tional change, localization, politico-economic unity and struc-ture, etc. Indeed, particular groupings may appear and disappear in substantive fact, their numbers may increase or decrease through time, and this may a l l be clearly recognized by the people them-selves. However, as long as the ideological criteria which defined the nature of the earlier societal subdivisions (such as a rule of patrilineal descent) continue to apply, in the minds of the people, to currently recognized social subdivisions, then the compositional elements of the social system remain ideologically unaltered. Fur-ther, as long as the relations among these elements (albeit speci-fically new ones) remain normatively the same, then the entire ide-ological social system remains unchanged. I do not wish to imply that the system is perceived as such by the culture participants. The ideological system is a - 23 -mode of relating some (as many as possible) of the reliably identi-fied normative elements of a culture in some logical manner. Most often, the culture participant himself, when asked "why" with res-pect to a particular norm w i l l ( i f he i s inclined to answer at a l l ) provide explanation directly in terms of some further norm proper to his own culture. Let us take an example from a situation in which there appears a norm of preferential matrilateral c.c. mar-riage. Hsiao-Tung E s i , in his book "Peasant L i f e In China" ob-serves; In the village, two kinds of "cross-cousin" marriage can be distinguished. A g i r l married to the son of her fa-ther's sister i s called saonseodiu, meaning a g i r l going up the h i l l . "Up the h i l l " signifies the prosperity of the family. A g i r l married to the son of mother's bro-ther i s called wesienodiu, meaning a g i r l going back to the native place. This i s interpreted as a sign of ruin of the family. As the terms show, the people prefer the up-hill type and dislike the "reverting" type. Let us see what i s the real difference between these two types. . . .29 Fei's Chinese peasants explain their avoidance of FaSiDa marriage by quite direct reference to other of their culture norms; in this case having to do with the support of "family prosperity". Anthropologists are seldom satisfied with a culture participant's explanation of his own norms (sometimes refered to as "conscious paradigms"). Fei, for example, goes on to explain the "real d i f -30 ference...". Hugo Nutini explains this dissatisfaction as follows: Conscious paradigms are not models at a l l , i f we take for granted that a model must be an explanatory con-struct. .. .conscious models are not explanatory, that i s , they do not entail c a u s a l i t y . ^ - 24 -The minimal form of "causality" manifest in the anthro-pologist's explanatory models is logical connectedness among i t s constituent propositions. Confronted with a cultural norm taken as a problem (such as normative marriage rules), the anthropologist attempts to provide a series of propositions from which he i s able, in effect, to deduce the proposition which states the problem norm. The problem element is then taken as explained by the logical interaction of premises whence i t i s derived. Where the problem element of the model i s an ideological norm the completed deduc-tion i s often referred to as a mechanical model. Where the problem element i s , empirically, a s t a t i s t i c a l norm the construct i s often called a s t a t i s t i c a l model. This essay, then, is concerned with so-called mechanical models. There are two further questions concerning the anthropol-ogists explanatory models: (1) What i s the epistemological nature of the propositions that compose a completed model? (2) How can we select among equally (logically) valid models of the same social phenomenon? How, that i s , do we verify models? Let us consider the f i r s t question. There seem to be three basic types of elements distributed among the premise pro-positions: (1) Ex p l i c i t , and therefore "conscious" ideological norms. In mechanical models this includes the problem norm i t s e l f . - 25 -(2) So-called s t a t i s t i c a l norms. (3) Theoretical postulates of the Anthropologist. One major objection which may be directed against this view involves my inclusion of s t a t i s t i c a l norms as legitimate elements in a so-called mechanical model. Witness Nutini: The reader must not interpret this to mean that we can only construct mechanical models out of or based on ac-tual behaviour. This i s not what I am saying, for a l l models, be they mechanical or s t a t i s t i c a l , can be con-structed out of any set of empirical data as long as they explain the facts in question. What I am maintain-ing i s that mechanical models constructed out of ideal behaviour, and s t a t i s t i c a l models based on actual behaviour are best; they are the "most explanatory" models that we can build in dealing with social phenomena. . . . mech-anical and s t a t i s t i c a l models, as explanatory constructs, must always be regarded vis-a-vis their proper epistemo-logical referents, i f they are to explain a given set of social facts in the best possible manner.^2 In sum, Nutini seems to be arguing (very ambivalently) that the best mechanical model i s one in which a l l directly empirical ele-ments are ideological norms whereas the opposite is the case in s t a t i s t i c a l models (based on s t a t i s t i c a l norms). I see no reason to endorse this viewpoint. It is a case-in-point of the kind of mind/body dichotomy I referred to earlier. Ideological norms belong to the world of mind. The world of mind cannot, i t seems, be best explained by events of the distinct world of body, (i . e . , statistical norms). The realm of human thought, perception, and emotion is self-contained (except as it: i s causal in body action) and wholly discontinuous with the more rudely corporeal conditions of l i f e . For my part, I view - 26 -this epistemology as highly restrictive and more theological than scientific. Man's psychic world, and in particular his normative definitions of the world in which he lives and dies, cannot be comprehended without reference to the de facto behaviour in which he engages, often (if not usually) unawares. In brief, the "best" models are always "mixed" models in that they say more about MAN. Added to this is the unquestionably greater range of theory-building materials afforded by the mixed model approach. In this respect, I might add, Nutini appears to misunder-stand what Levi-Strauss and Needham actually do in their substantive work. The very explanatory postulates of their theories refer to phenomena which are explicitly not conscious norms but either un-perceived (by the culture participant) patterns of de facto behav-iour and/or (at least) attitude sets which, as such, are unper-ceived by the participant himself. These considerations lead me to conclude, further, that a psycho-biological view of Man enters a l l Anthropological theory, admittedly often unexpressed, whether or not the theorist is prepared to admit i t . The epistemological and empirical nature of what I have termed theoretical postulates is an exceedingly subtle matter. Ultimately, i t is these postulates which are the explanatory ele-ments in a l l of our models; they are what we ordinarily call state* ments of theory. Such propositions, as they apply to particular explanatory models, have a number of important characteristics: - 27 -(1) They are more abstract than the statements of the directly empirical norms which are logically connected through them in the model. This follows from my claim that explanatory models are basically deductive constructs. The concluding propositions are logically subtended by the explanatory, theoretical, propositions. Let me illustrate again from Fei's material. Fei attempts to explain the "real difference" between the two types of cross-cousin. I have provided diagrams beneath to enable the reader to more easily follow Fei's remarks. (Figs. 2 and 3) He states:" In the fj^rst case /matrilateral c.c. marriage/ the girl Ja. bride/ will be the daugher-in-law of_her father's sister, who is from the father's Chia /a. patrilocal ex-tended family/ and s t i l l stands in an intimate relation to her father^ while in the second case /patrilateral c.c. marriage/, the girl will be the daugher-in-law of her mother's brother's wife, who has suffered from her own mother-in-law who is the girl's mother's mother. An intimate relation of a mother towards her married daughter is usually jealously resented by her son's wife. When the girl comes under her control, she will take her as a target for revenge.^ Figure 2 PATRILATERAL - 28 -Fig. 2 shows mother/daughter pairs by loops of broken lines. #1 resents the intimacy between #2 and #3, and expresses this by making her daughter-in-law, #4, who i s daughter of #3, suffer. She i s able to do so in virtue of the normative authority vested in the mother-in-law. Figure 3 MATRILATERAL Fig. 3 However much #1 may resent the intimacy of #2 and #3, she cannot effectively avenge herself be-cause her own daughter-in-law, #5, i s not the daughter of #3 as in Fig. 2. Although this argument, as i t stands, has d i f f i c u l t i e s , i t i s not my present purpose to offer a thorough c r i t i c a l analysis. For purposes of il l u s t r a t i n g the abstractness of theoretical postu-lates i t w i l l be sufficient to examine a single segment of the argument: (a) A woman resents the "intimacy" shared between her mother-in-law and the daughter of her mother-in-law. I take this to be based on some direct observation by Fei of de facto behaviour, or, possibly, interview statements. (b) The woman desires "revenge" against them but, in view of her status with respect to them (normative authority of mother-in-- 29 • law over daughter-in-law), is unable to act directly and openly against them. 1 assume this statement would be supported sim-ilarly to (a). (c) Patrilateral c.c. marriage results in the woman's own daughter-in-law being the daughter of the woman's mother-in-law's dau-ghter. (d) Matrilateral c.c. marriage does not have this effect. (e) If the woman's own daughter-in-law is the daughter of her own mother-in-law's daughter, the woman will "take her as a target for revenge"; not so or less so otherwise. (f) Conclusion; A woman will make her daughter-in-law suffer more where the patrilateral form of marriage is practiced than where the matrilateral form is practiced. In the series of premises the proposition which is most uncertain as to its empirical source is (e). I suspect that i t could not be, as such, traced to informants' statements or even from Fei's observations and interpretations of actual patrilateral c.c. marriage situations* Indeed, for Fei to have observed a situ-ation with the entire structural gestalt required by this statement, actual patrilateral marriages would have had to occur through several generations in a family line. Fei's statements make such an event very unlikely. What, then, is the source of the statement? Statement (e) appears to almost follow from (a)-(b); but not quite. They need a logical cement — a statement of more gen-eral cast than any of (a), (b) or (e). Let me suggest, for example, a fourth proposition to be interpolated between (a)-(b) and (e): - 30 -to a third person a mother and her daughter are socially equivalent in the sense that one may serve as a surrogate of the other in the eyes of the third party. This statement anticipates the views I have developed in the second section of this essay. The present formulation of the postulate is certainly inadequate but will serve my immediate pur-pose of illustrating the nature of theoretical postulates. Thus, in a situation of repeated marriage to the literal patrilateral cross-cousin, a woman will be far more tyrannical in her treatment of her daughter-in-law than otherwise because,under- these circumstances, the daughter-in-law is, in the woman's eyes, a sort of social surrogate to persons whom she resents (the woman's mother-in-law and the daughter of the latter) and this is so specifically because she is the daugher of one of these women. It is easy to see from this illustration that our theore-tical postulate is a proposition of more general c a 3 t than the statement of the social phenomena which they purport to explain. Here, for example, our postulate stipulates neither the particular formal status relationship (e.g. Kin-relation) between our third party and either of a mother/daughter pair, nor the qualitative nature of relationship of the outsider toward the pair. Our norm patterns, on the other hand, specify both. (2) A second feature of theoretical postulates concerns the nature of their subject elements and of what functions are appropriately predicated to these. Are we concerned, as the subject elements of - 31 -our theoretical statements, with "people", or "groups", or " i n s t i -tutions", or with something else altogether? Do these "act", or "function", or have "effects", have "attitudes"? A f i r s t and perhaps faci l e answer to these questions i s just that a subject element may be any of a large range of concepts. Which i s employed on a given occasion w i l l depend upon the nature of the problem at hand. Also, the predicate element i s , in a com-mon-sense way, a function of the subject element of the statement. The language of Anthropological theory i s not yet so specialized that we can meaningfully refer to the attitude of an institution, unless we are shown that this can be understood in some e l l i p t i c a l sense. We can say more than this. Homans and Schneider observe that a l l institutional theories have "at one time or another, been called functional theories, but the word function has been used in 34 several different senses." They recognize three "chief senses" of the word. These are: (a) "Malinowskian Functionalism"; An institutuion i s what i s is because i t results from the drives, or meets the immediate needs, of individuals or sub-groups within a society.... interests may be other than economic. (b) "Quasi-Mathematical functionalism"; An institution i s what i t i s because i t meshes with other institutions within a society. (c) "Final Cause" functionalist; An institution i s what i t i s because i t is in some sense good for a society as a whole.^ - 32 -In sum, Homans and Schneider find three types of theoretical postu-lates differing from each other above a l l in what kind of predi-cation each ascribes to its subject element. Since there are cer-tainly many more than three kinds of specific predication to be found in theoretical statements, I take them to mean that a l l such belong to one of these three classes in some ultimate sense beyond the scope of our immediate inquiry. Taken as a totality, the theories of Levi-Strauss and Need-ham contain a mixture of a l l three of these types of functionalism. Let us look first at their "quasi-mathematical" aspect. I have already observed that a central feature of Needham's Purum analysis appears to involve the coincidence or, to use Homans and Schneider's term, "meshing" of the rule of prescriptive matri-lateral c.c. marriage and the alleged normative prescription to women of specified, named, whole descent groups (vide supra, p.18). We have yet to examine this proposal in detail but, for the time being, we shall accept i t at face value. More specifically, the exchange theorists have argued that the ordering of one way giving and taking of brides between whole 36 descent groups is a "structural implication" or, as Needham puts 37 i t , a "structural entailment", of the prescriptive matrilateral marriage rule. As I have already pointed out (vide supra,pp. 21-22) the ordering of descent group relations is to be understood above a l l in a normative sense. In this argument our subject element - 33 -(the marriage rule) i s linked to the object element (descent group relations) by what i s held to be a purely logical connection. The marriage rule explains the group relations as premisses explain their conclusion. In this context nothing of the order of drive or purpose i s predicated to the marriage rule. This argument presents two d i f f i c u l t i e s . F i r s t , i s i t truly the case that the prescriptive marriage rule does "entail" the kinds of whole descent group relations that i t i s purported to entail? I shall show later that i t does not under any but certain highly specific conditions. I shall further examine Purum data i n light of this argument. The second d i f f i c u l t y , of importance i n our present context, i s that the exchange theorists are distinctly concerned with providing an explanation of the marriage rule. Thus, a quasi-mathematical functionalist explanation of the prescriptive matrilateral c.c. marriage rule would demonstrate that ^ t could be derived deductively from the definition of another socially co-existing institutionalized norm; not vice versa. There are points at which Leach and Needham appear to be suggesting just this; i.e., that institutionalized whole group relations (of one-way bride move-ment) imply the rule of marriage. Leach cannot take this position firmly,however, because the Kachin of Burma on whom his analysis i s based simply have no explicit norm of whole-group one-way bride movement. In the case of Needham's Purum this issue i s a moot point (to be examined later) although I suspect, in the f i n a l analysis, that Needham's d i f f i c u l t y i s one and the same as Leach's. - 34 -In view of this d i f f i c u l t y , the exchange theorist lacks a quasi-mathematical explanation of the marriage rule, and although a meshing of the marriage rule and group relations are a part of the theory, this i s not the part which contains the explanatory punch. For this, they must shift to predication of a different kind. The shift i s somewhat obscured by a corresponding shift in the definition of the problem. In Needham's words, the problem now becomes "roughly this: Given that a society practices this form of marriage /unilateral c.c. marriage/, why i s i t contracted 38 with one side rather than the other." As i t stands, this question is remarkably ambiguous. However, I shall take i t to mean what i t appears to mean; viz., why i s matrilateral c.c. marriage (whether understood as prescriptive or merely preferential) considerably more frequent than patrilateral c.c. marriage? In his contrast between the two forms of unilateral c.c. marriage, Levi-Strauss provides what I consider to be the beginning of a good answer to this question. At present I am more concerned with his form of explanation than with i t s detail. Accordingly, I shall reproduce the argument only in broadest outline. The reader w i l l be assisted in following this discussion by the diagram be-neath representing a model of a patrilateral cross-cousin marriage system (from Needham, 1961, p. 15). - 35 -= A O = Figure 4 PATRILATERAL SYSTEM Homans and Schneider make the following summary remark about patri-lateral systems: father's sister's daughter marriage, just like mother's brother's daughter marriage, requires at least three l i n -eages; any one lineage i s linked by marriage to two others in the ring, and the ring can be lengthened indefinitely. On a l l these counts i t meets Levi-Strauss's requirements for generalized exchange /matrilateral system/. The only difference i s that the men of B lineage, defined either pa t r i l i n e a l l y or matrilineally, give women alternately to the As and to the Cs instead of always getting them from one and giving them to the other.^9 This statement may be represented simply by the following diagram. 47 A ^9- 4 E > n Af- i B i t C i D i f -A- ->B E n i 1 -3} n Figure 5 Fig. 5. Schematization of whole descent group relations (for 'n' groups) in a patrilateral system. Letters = men of a descent group; Arrows = direction of movement of women; vertical broken lines = descent. The diagrams given also i l l u s t r a t e an important corollary to Homans and Schneider's observations. Any single descent group, - 36 -such as "C", exchanges women, in both directions with, two others (schematically), but i t is as dependent on the working of the "entire system" as a single group in the matrilateral system. In order to continue the alternating give-take with "C", "D" must persist in alternating give-take with "E", "E" with "F", and so on. In this sense the "organic solidarity** of interdependence appears to be at least as intricate as in the matrilateral sys-tem. Homans and Schneider say that "We might even go on to argue that father's sister's daughter marriage makes for greater organic 40 solidarity...." If I thought that the concept of organic s o l i -darity had any use whatever in this problem (which I do not) I would be obliged to agree with Homans and Schneider. In view of these observations, what can Levi-Strauss possibly mean by his following statement? Instead of constituting a global system, as do both b i -lateral and matrilateral cross-cousin marriage each in i t s own sphere, marriage with father's sister's daughter is not capable of attaining any other form than a multi-tude of l i t t l e closed systems, juxtaposed to one another, without ever realizing a global structure.^ To understand Levi-Strauss's metaphor we must return to an earlier discussion. We have said that a prescriptive rule of matrilateral c.c. marriage, by definition, groups a l l women of marriageable age into three exclusive terminological categories. In addition, any ego w i l l , ex hypothesi, regard a l l such women who are of a single descent group as belonging collectively and exclusively to one or another of these three categories. Thus, from the viewpoint of any Ego, any woman of marriageable age can be judged marriageable - 37 -or proscribed automatically and simply i n virtue of her descent group membership alone. Her kin-status i s determined by her des-cent group status. No system could be more truly "global". Contrast this with the situation generated by a prescrip-tive rule of patrilateral c.c. marriage. Once again l e t us con-front Ego with three terminological categories of women corres-ponding to S i , FaSiDa and MoBroDa. If Ego is a man of descent group "C" (Fig. 5), into which of these terminological categories does he place a given woman of descent group "D"? The women of "D" do not belong, collectively and exclusively, to any one of the three categories; a given woman could belong to any one of the term-categories or at least to either of the FaSiDa or MoBroDa categories. But, i t may be argued, surely Ego (say he i s located in the middle row of Figure 5) calls women of "D" by the term for FaSiDa since the arrow between C and D moves in the direction of "C"? This is true as long as the system includes an explicit means of separating the generations; such as explicit generational moieties or exp l i c i t l y named sections which divide each descent group in half on the basis of alternating generations. These structures are shown in the following diagrams: Figure 6 - 38 -Fig. 6. Patrilateral System with explicit alternating generation moieties (endogamous). Moieties are named "1" and "2"; solid arrows = direction of movement of women; broken arrows = descent. Figure 7 Fig. 7. Patrilateral System with named sections; Num-bers are named sections recruited patrilineally on principle of alternating generations. With either of the above qualifications a patrilateral system can manifest a structure involving prescriptive "circulating connubium" between two sets of sections (which do not themselves intermarry but are connected by descent) or, alternately, circulat-ing connubium among the moiety divisions of whole descent groups (Fig. 6). In either case, furthermore, Ego can readily identify the terminological category of any woman by simple reference to her section or to her descent group and generational moiety. For example, in the hypothetical situation shown in Figure 7, the terms might be distributed roughly as follows (Ego in section 1): women of section 1 = sisters women of section 3 = mothers' brothers' daughters women of section 5 - fathers' sisters' daughters women of section 2 - fathers' sisters and daughters (note: FaSi = mo-in-law) women of section 4 - sons' wives (= sister's daughters) and mothers. women of section 6 = daughters' husbands' sisters and fathers' sisters' husbands' sisters. - 39 -Since Levi-Strauss does not consider this possibility let us ask what Ego would confront without any explicit ordering of generations? The situation would surely be very ambiguous. Once again, i f Ego belongs to descent group "C" (say the middle row in Fig. 5), how does he determine the kin category of a given woman of descent group "D"? She is not categorically one or the other type of cross-cousin simply in virtue of her descent group membership vis-a-vis his own. At its simplest a l l that can be said of the women of "D", collectively, is that in the eyes of the men of "C" some are their individual MoBroDa's while others are their individual FaSiDa's. In the absence of clear-cut gener-ational divisions Ego has very l i t t l e to guide a decision concerning a particular given woman. Nor is relative age a satisfactory c r i -terion since i t does not always fai t h f u l l y reflect the relative generational positions of Ego and the woman in question. Assuming that Ego i s bent upon obeying the FaSiDa prescrip-tion what devices remain to him for determining the given woman's kin category? What about their precise genealogical connexion? Be-yond his f i r s t genealogical cousins Ego w i l l run into very great d i f f i c u l t i e s . Harrison C, Whyte has argued, for example, that f u l l y • eight of the sixteen genealogically distinct second cousins "are 4 2 genuinely new kinds of kin in the classificatory kin system." Of the remaining eight, four are identified as classificatory sisters, two as MoBroDa's and two as FaSiDa's . The upshot of a l l this i s that, in a patrilateral system without explicit odering of generations, Ego i s confined to identi-- 4 0 -fying his undisputed cousins of the various types to a narrow range of genealogical specifications. Matrilateral cross-cousins, for example, would include the l i t e r a l MoBroDa, MoFaBroSoDa, MoMoSiSoDa; patrilateral cross-cousins would include the l i t e r a l FaSiDa, FaFa-BroDaDa, FaMoSiDaDa. Beyond this, we must move to the genealogical third cousin level, and here we may anticipate that an even smaller proportion could be identified as to kin-category than in the case of the second cousins. The remaining second and third cousins can-not be identified through the unaided logic of classificatory kin-ship alone. The reason for this i s clearly, in the last analysis, that a person's terminological category i s characteristically de-termined, for Ego, as a function of that person's descent group (and, sometimes, generational position) and Ego's personal genea-logical connexion with the descent group. Sometimes Ego's personal genealogical connexion with the descent group i s , as i t were, pre-determined by the normative marital relations of his whole-descent group to the other one in question. This need not be the case, however. This discussion w i l l be made clearer at a later point. In sum, Ego can locate unambiguously only a small range of FaSiDa's, MoBroDa's, and Si's, but remains confronted with a relatively large pool of women none of whom can be clearly identi-fied as to kin-category without the aid of an explicit generational principle. Thus, i f Ego obeys the patrilateral prescription he is confined to selecting a bride from a small pool of women who are genealogically traceable through his personal family history. A str i c t patrilateral prescription means, in other words, that the - 41 -repetitious affinal alliances obtain between narrow family lines and not, as i s more clearly possible in a matrilateral system, between larger groupings. This, I take i t , is what Levi-Strauss i s driving at where he refers to "a multitude of l i t t l e closed systems juxta-posed to one another, without ever realizing a global structure." A matrilateral system, by contrast, does not require an explicit generational principle to create such a "global" structure. By simple reference to descent group alone Ego can identify the kin-category of a given woman; at least, that i s , in the hypothetical exchange theory model. But what has a l l of this to do with the different frequen-cies of the two forms of cousin marriage? The answer is that Levi-Strauss predicts in an ex post facto way that the patrilateral rule has a lower survival probability than the matrilateral one. Once again I have had to take interpretive liberties with Levi-Strauss 1s statements. He appears at this juncture to want to move in two direc-tions at once. On the one hand, he has pointed us in the general direction of an important contrast between the two cousin marriage systems. As I w i l l argue more.fully at a later point, there appear to be good reasons to expect that the patrilateral rule has lower survival probability than the matrilateral one. These reasons a l l relate to statements concerning people. It has already been suggested, for example, that Ego (any male culture participant) i s confined to a relatively small pool of women from which to - 42 -choose a wife. One can easily envision situations arising, in the course of human events, as they are played out in a patrilateral system, in which Ego is hard pressed to break his prescription and select a bride from the undefined pool. The situation would simply be one in which Ego confronted a severe scarcity of women who are clearly FaSiDa*s. Should this occur regularly and become widely accepted, the system can no longer be labelled a prescriptive one but must be termed a merely preferential one. The problem is con-siderably more complex than this, however, and I must leave i t un-t i l further analysis prepares us to take i t up again. My present argument concerns the fact that Levi-Strauss has, however impres-sionistically, pointed us i n the direction of important distinctions between the two rules. The distinction devolves about propositions concerning human beings to whom we predicate certain general be-havioural dispositions (e.g. the disposition to mate and reproduce). I intend to locate some of these which are relevant to our present problem and to then show in greater detail that such dispositions are more lik e l y to be frustrated in a patrilateral than in a matri-lateral system. It seems possible then to construct an argument which ex-plains the differential frequency of the two cousin marriage forms on the grounds that the patrilateral form must generate regular frustration of certain important behavioural dispositions where the matrilateral form does not. Such explanatory postulates contain people as their subject elements and predicate to them what I - 43 -have termed behavioural dispositions. Such propositions would f a l l more nearly within the sphere of Malinowskian functionalist than within the other types. At this juncture Levi-Strauss's argument undergoes a subtle shift in ground. To this point he has hinted that there might be good reason why few cultures have a patrilateral (prescriptive) cross-cousin marriage norm. But how does this explain why certain other cultures d_o have a matrilateral cross-cousin marriage norm? The argument virtu a l l y posits an a r t i f i c i a l situation of choice; the matrilateral form i s chosen over the patrilateral. In effect, i f people are in any sense the subject-element of the explanatory postulates implied in this argument, what i s predicated to them i s the action of rational choice; choice, that i s , which is based upon foreknowledge. This i s precisely what Homans and Schneider had in mind when they remarked that "Levi-Strauss's efficient cause 43 is human intelligence." This aspect of Levi-Strauss's explana-tory efforts i s surely an example of f i n a l cause functionalism. In my view, the kinds of propositions which have the greatest and most ultimate explanatory power in the problems dealt with by social anthropologists, are propositions which have people as their subject-element (whether a l l people, Purum people, males, etc.) and which predicate to people various dispositions (sets?) for acting toward and reacting to other people (also further specified) in corresponding ways. Mauss's concept of "total prestation", of great importance in the exchange theories of prescriptive marriage, - 44 -i s an outstanding example of this type of theoretical proposition. Witness Mauss: Total prestation not only carries with i t the obligation to repay gifts received, but i t implies two others equally important: the obligation to give presents and the ob-ligation to receive them.^ "Obligations" are examples of behavioural dispositions predicated to people acting individually or in groups. ( 3 ) The verification of explanatory postulates: I shall return anon to a further exploration of the epistemological nature of dis-positional postulates. For the moment let us return to an earlier perspective. Anthropological models are an attempt to explain explicit norms ( s t a t i s t i c a l or ideological). The deductive sequence which yields the problem norm ( i . e . , the model) requires, as I have ar-gued, certain generalizing postulates taken as axiomatic in the model i t s e l f . As scientists, however, we are inclined to attempt to verify our axioms. Speaking for myself, I regard particular models as ultimately of less importance to our discipline than the general postulates upon which they rest. How, then, do we support the postulates? I have stated that the postulates are general proposi-tions and operate axiomatically within particular models. By this I mean that they logically subtend propositions derived from them. Let me illustrate this point from Leach. In the concluding sec-tion of his essay, "The Structural Implications of Matrilateral Cross-Cousin Marriage", he states that - 45 -In every relationship between individuals and between groups, items in the above l i s t are exchanged. It i s in the nature of most such 'exchanges' that, as regards the tangible items a, b, c, d, e there is always an imbalance on one side or the other. The exchange ac-count i s balanced by the intangible items f and g . ^ The l i s t referred to in the citation i s as follows: I. Tangibles (a) women and men (b) labour of men or women (c) consumer goods and money (d) capital goods (e) r i t u a l objects of no intrinsic value. II. Intangibles (f) rights of a t e r r i t o r i a l and p o l i t i c a l nature (g) relative status or prestige.^6 These statements, operating as premises lead to the following con-clusions: With Kachin type marriage the relationship between wife giving and wife receiving groups i s asymmetrical; hence differentiation of status one way or the other i s more li k e l y than not. Such differentiation can be avoided i f d small number of neighboring local groups marry in a ci r c l e , or i f there i s a system of balancing rights and obligations — as with the Murngin; but any such system of balances w i l l be unstable.^ At the risk of becoming tiresomely repititious I shall try to restate the general flow of Leach's argument: (a) If tangible g i f t s are not repaid by other tangibles they must be repaid by intangibles. This is clearly a corollary to Mauss's notion of total prestation. (b) A rule of matrilateral c.c. marriage implies one-way bride giving between groups, (in Leach's view specifically "local 48 descent groups"). - 46 -(c) Therefore, in a Kachin type system there i s an implicit im-balance i n the exchange between groups of at least one speci-f i c tangible; i.e., women. (d) The debt incurred as per (c) above i s balanced by an intan-gible; i.e., status differential as between bride giving and bride taking groups. (e) Therefore, Leach predicts, there w i l l be a status differen-t i a l between bride giving/bride taking groups i n a society having a rule of matrilateral c.c. marriage. My present point i s that the concluding proposition, (e), is logically subtended by the major premise, (a), which i s i t s e l f subtended by Mauss's definition of total prestation (vide supra, p.44 ). The relation between the movement of women and of deference shown between groups i s a particular case-in-point of the general kinds of behavioural dispositions postulated by Mauss. What does this model achieve? F i r s t we should note that i t does not, as i t stands, provide an explanation of the mar-riage rule i t s e l f . To accomplish that end with the propositions given above i t i s necessary to reverse their order. Given (empirically) a norm ( s t a t i s t i c a l and/or ideolo-gical) of status differential among the descent groups which com-pose a society, one x«mld argue, roughly, that a one-way movement of women between any two groups (in which one renders deference to the other) i s one means at least of balancing the debt incurred by - 47 -the disequilibrium of status. Further, one would have to reverse proposition (b) to show that the one-way movement of women implies a rule of prescriptive matrilateral c.c. marriage. Again, Mauss's principle of total prestation and Leach's corollary, proposition (a), operate as major premises. I am unclear which of these alternate models was actually intended by Leach. To expedite the present discussion I shall con-cern myself with only the f i r s t of them. This model predicts of societies with a rule of matrila-teral c.c. marriage that the constituent local descent groups ex-change women as brides, i f they intermarry at a l l , i n one and only one direction. Secondly i t predicts that, as between a l l inter-marrying pairs of local descent groups, there w i l l be found a status differential ( i . e . , there i s deference shown by the members of one group to the members of the other) Finally, the direction of deference w i l l be homogeneous among a l l such pairs moving either from bride givers to bride takers (Leach cites the Chinese pea-sants) or consistently in the direction opposite to that in which brides move (e.g. Kachin). The direction of the flow of def-erence i s consistently either the same or opposite to the flow of women; not random among the intermarrying groups of the society in question. These proposals constitute exemplifications of the gen-eral human inclination — ex hypothesi -- to balance giftgiving interaction; Leach remarks: - 4 8 -In any such system of reciprocities one must assume that, overall, both parties — the junior group and the senior group alike — are satisfied with their bargain, and therefore that the exchange account 'balances'. But we cannot predict from first princi-ples how the balance will be achieved because we cannot know how the different categories of 'prestation' will be evaluated in any particular society.^ A model of this kind can be used in only one of two ways at a time. It can be used, on the one hand, as an explanation of a particular situation. In this case, i t would provide explana-tion of an empirically observed pattern of intergroup status relationships of the kind described above. Even in this respect i t would not be employed as a complete explanation because no mo-del i s . A l l good models beg further questions. In the present case, for example, explanation of status relations is given in  terms of the principles of prestation, applied to the particular situation of one-way bride give and take. Thus, the unidirection-ality of bride movement (and, hence, of the matrilateral marriage rule) remains, itself, a begged question. The problem arising from this use of the model is, however, that i t requires axio-matic acceptance of the general theoretical postulates upon which the model is based. Thus, the second, and methodologically prior use of such models is their treatment as devices for providing support of the general postulates. Such a procedure has always been philosophi-cally problematical but scientifically indispensable. In most, i f not a l l , scientific models the "if...then" relationship between - e>9 -the general postulate and the hypothesis cannot be directly reversed; i.e., the relation of a particular conclusion to the general pre-mises whence i t is derived is inductive when moving from particular to general. Thus, for example, i t must always be admitted as pos-sible that, in situations found to faithfully reproduce the predic-tions of Leach's model, some general behaviour disposition other than those contained in Leach's model can generate the same pre-dictions. No verification procedure can ever assign to a the* oretical postulate more than a measure of probability. These two uses of models cannot be applied aimu]ftaoeoualy without circular argument. To maintain that the Kachin status structure (taken as an empirically given norm — ideological and/or statistical) is a result of a disposition to reciprocity in inter-action, we must support the claim that the Kachin in fact have such a disposition. What is more, we must support this claim through empirical observations other than those which we use the postulate to explain. Thus, to complete a particular explanatory model we must introduce data other than that which is itself ex-plained by the model. This feature of explanatory models is i t -self sufficient to account for both of the traditional "cross-cultural" and "whole-cultural" analytic frameworks. The expression "whole-culture" analysis means that, where we employ a model to explain a given norm in a given society, we also predict, deductively, other behaviour patterns on the basis of our dispositional postulates. Thus, for example, we might pre-- 50 -diet that, among the Kachin, interactions in many social contexts other than those specifically of marriage and affinal alliance display specific kinds of patterns which are implied by the con-cept of reciprocity. The difficulty in this particular situation resides in the near impossibility of stipulating a situation which cannot be argued to manifest the dispositions of reciprocity. Be that as i t may the point illustrates the idea of "whole-culturalism" as a methodological procedure. Beginning with an explanatory model of one cultural phe-nomenon we deductively generate models for phenomena of other kinds. Thus, where we are concerned initiall y with a marriage rule, we are led to deduce many other forms of behaviour as likely in the same social situation (i.e., involving the same set of actors ) " i f the postulates of our i n i t i a l model are true". We could be led to predictions about behaviour in political, economic, ceremonial, mythological, any number of other frames-of-reference. The par-ticular frames of reference selected will be anticipated by the nature of the particular postulate. A given postulate may specify behavioural dispositions of interactions between certain categories of people, say, for example, persons of opposite sex and alternate generation. Persons of these categories may interact extensively in some contexts (e.g. ceremonially, relating of folk-wisdom, etc.) but very l i t t l e in others (e.g. co-operative work, sexual intercourse). However impressionistic i t may be general ethnography is a completely indispensable guide to such a procedure. We can only afford to - 51 -snub general ethnography when we already have It thanks to the labour of our predecessors. In sum, the reliability (i.e. believability) of a parti-cular explanation of a phenomenon is measured by the support given to its theoretical postulates through the substantiation of other models based upon the same postulates. The measure of substan-tiation of a l l such models (including the one centered about our i n i t i a l problem norm), singly and collectively, is a measure of the reallability of our theroetical postulates in the society with which we are concerned. An important corollary to this argument is that whole- * cultural orientation (the word whole must not be taken over-literally) is methodologically prior to cross-cultural analysis. There is no a priori reason to conclude that the theroetical postulates which believably explain a given norm (e.g., prescriptive matrilateral c.c. marriage) in one society also explain a like-appearing norm in another society. For a model to be acceptable in the second society its postulates must work in the same way that they work, whole-culturally, in the f i r s t . I am in whole-hearted agreement with Needham's remark that "you cannot compare what you do not f i r s t 50 understand On the other hand, the reliability of our theroetical postulates relating to phenomena in one society is increased as cross-cultural evidence accrues. The disturbing feature of this methodo-- 52 -logical point, and i t is one which I personally endorse for the most part, is that there is no "logical" reason for i t . It is based on an ontological faith; i f people do the same thing as one another, they do i t for the same reasons. This axiom can be carried to the tyrannical point where we will deny that a case-in-point is really the same thing; i t cannot be, so i t might be ... argued, because i t has required a different explanation. The diagrams below are provided to aid my present argu-ment. Figure 8 "x" = summary statement of the problem norm. "a,b,c, etc." and "1,2,3, etc." = summary statements of norms other than "x". "A" and "B" - theoretical postulates (statements of be-havioural dispositions). " "^ = a deductive sequence linking "A" and "x"; " the sequence will include other norms not shown and perhaps other postulates (e.g. Leach's models). "......y = "inductive" inference, (vide supra, p. 49 ) Each of these diagrams symbolizes an attempt to provide a reliable explanation of problem norm "x" in each of two types of society in which i t occurs. The social systems differ, obviously, in that "x" coincides with a distinct set of other norms in each - 53 -case. The diagrams are to be understood as purely hypothetical sit-uations since I am here concerned only with the logic of crces-cultural analysis. The surrounding features of each of these social systems (i.e., the other norms which coincide with "x") support, in each case, different theoretical postulates. Remember, i t is given that both "A" and "B" can operate as major premises in deductive models for problem norm "x". Now, given, further, that we have a relatively large number of cases of Type-I, but only a small number of Tyre-II, must we conclude that "B" is not a reliable explanation of the cases we have of Type-II because i t is not reliably supported in cases of Type-I, or vice versa? To be sure, we would be inclined to reject that either of the two theoretical postulates provides the basis of a reliable general (i.e., cross-cultural) theory of norm "x". But this does not mean that they are not each the most reliable theories for their particular social systems. Nor, on the other hand, does this mean that we must surrender the attempt to formulate a more general theory. The paradox in this situation is, however, that the lar-ger the number of cases of Type-I we are able to demonstrate, the more we are ready to believe that "A" is a reliable explanation of norm "x" in a single given case of Type-I. This sense of re-liability derives from the satisfaction of our (i.e., anthropolo-gists') expectation that people are alike, and we tend, therdbre, to be satisfied only with explanations which suggests that they - 54 -are alike. As I have stated, I share this general expectation i f only because i t is the keenest Occam's razor that we possess. On the other hand, we must never conclude that there can be no persons (or peoples) whose social behaviour does not support the same theoretical postulates as those applied to others with whom they share a given problem norm. In summary, while the scholarly orientations and methodology of Anthropology are cross-cultural in part, they are intra-cultural or relativistic first. This is a procedural imperative. (4) There are one or two further methodological points to be made. These will become more apparent following a statement, albeit only a sperse outline, of what I mean by the concept of behavioural disposition. I shall also attempt to show why I believe that such postulates provide the most forceful basis for institutional theory. In the first place, institutions or norms whether stat-istical and/or ideological are manifest (observable) forms of hu-man behaviour; that is, human individuals and/or collections of human individuals are our subject matter. Any theory which purports to explain the observed behaviour of a set of human individuals must predicate something to the human individuals who compose the set. It has been observed that few "conscious models" satisfy the anthropologist, (vide supra, p.23 ) . In some respects Levi-Strauss 's comments on this issue are unclear to me. He remarks, 51 for example, that they "are usually known as "norms" ." As I - 55 -have used the concept, a model is explanatory and, as such, to refer to norms as models of any kind is confusing since, as Levi-Strauss remarks further, norms are "very poor" models "since they are not intended to explain the phenomena but to perpetuate 52 them". Since I agree with this observation, I would, to avoid confusion, simply refrain from identifying the concepts of norms and conscious models. To my way of thinking norms (ideological) may comprise the elements of conscious models where one norm is explained, by the culture participants, in terms of another. The process of linking explains the norms, where the norms are the elements requiring explanation. Fei's report of the Chinese pea-sants' explanation of their marriage system is an example of what I mean by a conscious model"; i t is a genuine attempt at explana-tion albeit, from our viewpoint, an unsatisfactory attempt. The reason for this confusion seems clear enough. The context from which I drew the above citations infers that what is explained by models are de facto social events (what Levi-Strauss terms simply "the phenomena") rather than the ideological norms which perpetuate them. In brief, this posits two states of mind for the social actor; his awareness that he and others of his cul-ture usually do (or always do) something in fact and his conviction that the former is something which he ought to do. Levi-Strauss then argues that the "ought" is not a satisfactory attempt, by the culture participant, to explain what he observes that he does. "I do i t because I ought to do i t " does not satisfy the anthro-pologist as an explanation. - 56 -Of this there is no doubt. However, the important point is that there exist better attempts to explain provided by infor-mants; again refer to the case of the Chinese peasants. In other words, I do not conclude that conscious models are unsatisfactory because they are norms. Rather, the explanations provided by informants as to why they do something they are aware they do, and why they feel they ought to do this are characteristically un-satisfactory because the logical links provided are unsatisfactory in terms of our expectations. (See, yet again, the case of the Chinese peasants). This has nothing whatever to do with the in-telligence of our informants. This will become clear shortly. The elements of conscious models are always, by definition, themselves conscious; i.e., they are ideological norms. Given, that, from the anthropologist's viewpoint, such a model is unsat-isfactory, i t is always because i t is either entirely logically in-valid or because i t tends to validity but has missing links. In either case, the anthropologist, attempting to explain the same problem element, must order the various constituent norms in a logical sequence and complete the sequence by propositions of his own invention. Such propositions, as I have already argued, are more abstract than the statement of the norms which they cement. As such, they are the most explanatory propositions in the sequence. But, to the extent that these statements are of the anthropologist's own making and not of the culture participants', they do not exist on the level of consciousness in the actor. This should scarcely - 57 -surprise anyone since, in a general way, whatever exists as ex-plicitly normative and therefore conscious is defined as a problem by the anthropologist. Now, since most consious models are logi-cally incomplete, — at best — the anthropologist must invent propositions to complete them. Sometimes, on the other hand, there are no conscious models whatever. In this situation the anthropol-ogist must himself do a l l of the combining of various norms (stat-istical and/or ideological) into a deductive sequence which gen-erates the problem norm. Here, again, almost invariably, the anthropologist will have to invent major premises to complete his sequence. Even, in an extremely unusual situation, where the anthropologist should locate a set of statements, reliably sum-marizing norms of a culture, and find himself able to order these into an air-tight deductive sequence he will then ask "why" of the most major premise of the sequence. Again, to try to answer this he must invent further propositions. In a l l such situations the most explanatory statements which the anthropologist achieves are his own inventions. Invar-iably, therefore, the propositions of most forceful explanatory power which we as anthropologists seek are propositions which do not represent explicit verbalized norms, In brief our most power-ful theoretical postulates impute something to actors which does not exist on the level of their verbal consciousness. (There may be other kinds of consciousness such as an eidetic consciousness.) The concept of behavioural dispositions will readily accommodate the predication, to social actors, of characteristics - 58 -of which they have no verbal consciousness. This can include a wide range of psycho-biological concepts such as attitudes, apti-tudes, drives, etc. A l l such can be meaningfully referred to at the level of the unconscious. - 5 9 -CHAPTER 3 The Exchange Model of Purum Society: Salient Issues At this point I wish to return to more direct consideration of prescriptive matrilateral c.c. marriage. Further theoretical and methodological issues will arise out of this discussion. Earlier, I introduced the distinction between mechanical and statistical models. I took Hugo Nutini's formulation as the basis of that discussion. Levi-Strauss's definition of the dis-tinctions seems to be very much more complex. At the same time, I believe that both would agree that Rodney Needham's analysis of Purum social structure is a clear case of a mechanical model. Following his definition of the distinction, Levi-Strauss remarks that: The laws of primitive marriage provide the best i l -lustration of this difference. In primitive societies these laws can be expressed in models calling for ac-tual grouping of the individuals according to kin or clan; these are mechanical models.^3 This statement precisely summarizes Needham's model of Purum soci-ety. More fully, his model argues that: 1. For any Ego, prescription to the kin-category (remember, i t may be characterized by an interaction of terms) which Need-ham translates MoBroDa is tantamount to marrying a woman of any of a strictly limited and somehow pre-established set of descent groups. This set of descent groups will include the one(s) to which his own literal Mother's brother's (s') daugh-ters) belong. At the same time, the set is not exhausted by those group} containing his literal MoBroDas. - 60 -2. For any ego, proscription of women of the kin-category FaSiDa is tantamount to the proscription of marriage to any woman who i s a member of any one of a further set of pre-established descent groups. This set w i l l include, but not be exhausted by the descent group(s) to which ego's l i t e r a l FaSiDas belong. 3. For any ego there i s a further kin-category, sisters, which includes the women of his own descent group and generation, and may be exhausted by these. Equally, i t may include wo-men of other descent groups viewed as related i n "origin" to ego's own. 4. What i s true, above, of Ego i s also true of a l l of the men of Ego's own descent group, 5 . Thus, where a l l men and women of descent group x class a l l men and women of descent groups a, b, and c as matrilateral cross-cousins, a l l of the latter class a l l of the former as patrilateral cross-cousins. Therefore, men of x may marry women of a, b, or c, but men of the latter may not marry wo-men of the former. As between whole-descent-groups there i s only one-way movement of women in marriage. This i s also re-ferred to as the proscription of sister exchange. The f i r s t three points of this model are almost defini-tive of the concept of prescriptive. The definition accords with many societies which proscribe marriage with the l i t e r a l FaSiDa but do not proscribe i t with the MoBroDa. As Needham argues at length (see "Structure and Sentiment", Chapter Three), i t by no - 61 -means accords with a l l such societies. If we confine our analysis to truly prescriptive systems we have to deal more fully with the last two propositions as well as with certain features of the first three. The last proposition is a summary statement of the ex-change model of prescriptive matrilateral c.c. marriage. The fourth proposition is a necessary logical step in the analysis. It seems to me to contain the key element of the theory and yet is not explicitly dealt with in the expositions of any of the exchange theorists. Before we consider these specific questions let us consi-der the model as a whole. It seems, given Levi-Strauss's last statement, to qualify as a mechanical model. A mechanical model is "A model the elements of which are on the same scale as the 54 phenomena". A cryptic definition to be sure; nor do Levi-Strauss's elaborations make its meaning very much clearer. It is often the case with outstanding thinkers that the i n i t i a l formulations of their concepts serve merely to funnel our attention in a general direction. It is clear that the key to the meaning of this defini-tion resides in the word "scale". The "phenomena" referred to are the empirical givens taken as the analytic problem. In Need-ham's analysis of the Purum systems the problem phenomenon is the rule of matrilateral c.c. marriage. The elements of his model are - 62 -whole-descent-groups. The elements of the model and the phenome-non are in this situation alike in that both are components of Purum explicit ideology. At f i r s t glance this similarity does not appear to amount to a likeness of scale which suggests equiva-lence of size or measure. However, ideological norms are categor-i c a l propositions; they impute modes of behaviour to a l l members of a named (though often not expli c i t l y stated) category. Thus,-' for example, we are dealing with a rule which asserts something of a l l Purum men; namely that they must marry women only of the exclu-sive category MoBroDas. If a model i s truly mechanical, the subject element of i t s propositions must (at least) exhaust the membership of the category which comprises the subject-element of the problem norm. The expression to at least exhaust raises d i f f i c u l t i e s in the dis-tinction between the two kinds of model. The point i s that the model's subject-element must, at least, just exhaust the subject-element of the problem norm. Perhaps i t may be more inclusive, but i t may not be less inclusive and yet be a mechanical model. Thus, the Purum marriage rule demands explanation in terms of something which can be predicated to a l l Purum men and from which the marriage rule can be logically derived. To my mind this admits of two broad po s s i b i l i t i e s in the Purum situation: (a) The model may be treated as a purely s t a t i s t i c a l norm. This would require the direct observation of a mode of behaviour of Purum men which, while not actually expressed as such i n explici t - 63 -Purum ideology, nevertheless exists as a behavioural fact which i s , furthermore, a possible logical premise to the marriage rule per se. The d i f f i c u l t y i s , however, that the frequency of this de  facto event must at least just exhaust the frequency with which the ideological norm i s i t s e l f expressed. Before I continue further I must point out that this approach simply cannot be satisfactorily tested i n the Purum case, because we have no idea in fact of the frequency with which the marriage rule i t s e l f i s endorsed by Purum men. We tend to assume that because the statement of an ideology — i.e., a single i n -formant's statement — i s categorical (refers to a l l members of the subject category) a l l possible informants would accept i t as their own ideology. This i s clearly not necessarily true. In the Purum situation, however, we do not have any alternative but to accept this approach at the present time. Bearing i n mind this assumption, let us inquire whether Needham's model can be treated in this way. The required condi-tion i s as follows: If i t i s found that the marriages of a l l Purum men are such that a l l of the men of any given Purum descent group in fact marry into one of an exclusive set of descent groups (which does not exhaust a l l Purum descent groups), and i f the men of the latter do not ever marry women of the former, and i f the women of the marriageable descent groups are referred to by a term (or i n -teraction of terms) which is distinguished from the term applied to a l l women of a l l other descent groups, and i f this term includes - 64 -the actual MoBroDas of each of the Purum men, then we may conclude that a l l Purum men marry only women they term MoBroDas. But, i t might be argued, this model does not explain why the Purum say they ought to marry a MoBroDa; i t explains, rather, only that the women they do marry are termed MoBroDa. This dispute derives from what I referred to earlier as the mind/body problem of Social Anth-ropology. For my part I am prepared to believe that ideologies may be understood in terms of unperceived de facto behavioural tenden-cies even where these are not perceived by the actor in an explicit way. I t appears that Levi-Strauss also holds this position, per-haps in a more extreme form. He states that "When the structure of a certain type of phenomena does not l i e at a great depth, i t i s more lik e l y that some kind of model, standing as a screen to 55 hide i t , w i l l exist in the collective consciousness." I think this means that expressed norms are often screens for unexpressed norms (de facto behaviour or dispositions) which describe the underlying structure. This argues not only that explic i t norms may express ones which are not explicit, they may actually screen them. While I have argued, in the abstract, that this type of a model i s possible, I have now to argue that i t cannot apply to the Purum situation. The reason for this i s simply that the Pur-um evidence f a i l s to support i t . There are two kinds of Purum descent groups. F i r s t , there are five exogamaus patrilineal clans, which Das termed "sibs". - 65 -Four of these clans are further subdivided into named lineages. In a l l , there are twelve such lineages. Most of the lineages and a l l of the clans are dispersed among the four Purum villages. Since the evidence overwhelmingly indicates that one-way bride movement does not occur in any s t a t i s t i c a l l y significant frequency as between clans, Needham, and subsequently a number of other scholars, have attempted to ascertain whether circulating connubium applies to the lineage level of Purum society. This debate appears to have achieved a consensus that there i s a stat-i s t i c a l l y significant tendency among Purum lineages to avoid the direct exchange of women (2-way) between any intermarrying pair. If this conclusion i s accurate i t is indeed an interest-ing and very important feature of Purum society. However, i t i s not useful in the kind of model under present discussion. The most that i t can mean i s that a greater than chance proportion of Purum men marry women of an exclusive set of other Purum lineages; there i s a less, than chance tendency for the men of the latter to marry women of the former; and, f i n a l l y , given that the exclu-sive set of lineages into which men of a given lineage tend to marry, contain the only women whom a l l of the men of the lineage term MoBroDa, there i s a greater than chance frequency of marriage with a MoBroDa. But, is a greater than chance frequency of one mode of behaviour sufficient to explain the 100% frequency of some other mode of behaviour (the ideological norm)? - 66 -At f i r s t glance, such an explanatory effort would appear to be very faulty. How, one might ask, can we explain what 100% of a population do in terms of something which only 40% of that population do? But i f we recall that the 100% refers here to an ideology while the 40% refers to a s t a t i s t i c a l norm, we might en-vision avoiding this dilemma by postulating that i t is possible that the ideology supported by 40% because of something that they do can be spread, as an ideology, to the remaining population in spite of the fact that the latter do not themselves do what in-spired the ideology. Much additional evidence would be required to support such a position. An example of such a model would be one in which the overall tendency to one-way marriage derived from the relationships particularly of lineage nobility, this tendency being expressed by the actual marriage rule which was then, in turn, adopted as a universal ideology. Suggestions of this kind 56 have been made before. In summation, since there i s a considerable disparity in scale between the model (when treated as a s t a t i s t i c a l norm) and the phenomenon (marriage rule), and since no evidence i s provided to account for the disparity in terms of some additional feature of the structure (e.g., interaction of nobility and commoners), we must reject Needham's moSel as being based upon a s t a t i s t i c a l ob-servation concerning inter-lineage marriage relationships. But there are other ways of treating Needham's model and, in fairness to Needham, I should point out that he does not appear - 67 -himself to favour the above approach. He is rather puzzling on this point since he has taken great pains to measure the actual frequency of direct exchange between lineages. At the same time we may recall Needham's remark, cited earlier that "A social system is an abstraction relating (in this context) to lineal descent groups which are also an abstraction. There i s no place in ab-57 straction for substantive "specific groups"." This seems to mean that what is predicated to the groups i s not to be taken l i t -erally. The s t a t i s t i c a l features of whole group behaviour are somehow not of great relevance in his models. This suggests a secnnd interpretation of Needham's Purum model. (b) The model may be regarded as being premised upon propositions which themselves constitute explicit ideological norms. This would mean that the one-way movement of women between whole descent groups exists as an ideological norm whether or not i t i s actually obeyed. In this case, the norm of matrilateral marriage would be viewed as a function of the norms of descent group relations. Here, the subject element of the model would be equivalent in scale to the subject-element of the problem norm. An idedagy prescribing one-way marriage between a l l members of any pair of Purum descent groups, would mean that a l l descent groups just exhausts the membership of the category a l l Purum. In this case we can readily argue from the former to the latter. - 68 -The Purum did indeed have what appears to be such an ideology. It i s contained i n the following statement from Das: The Purum sib was not only an exogamous unit but i t was something more. Purum boys and g i r l s could marry only in one or more selected sibs, Sxh unions between the different sibs were fixed by traditional customs. Be-sides this there was another custom by which the boys and gi r l s of any one of these sibs might not marry into the same sib; that i s , they had to find spouses from two different sibs, the boys from one and the g i r l s from another. Thus brothers and sisters could not marry into the same sib.58 This citation refers to "traditional" marriage alliances between "sibs" taken as-a-whole. Furthermore, the tradition would result in only one-way movement of women between intermarrying sibs. To deduce the marriage rule i t s e l f from these normative sib prescriptions requires only that the system of terminology accord with them. That i s , i f we are able to show that (a) there i s un-anymity among the men of a sib as to what women they c a l l MoBroDa and that (b) these women are the same as those who belong to the maritally prescribed other sibs—there being unanymity among the men of a sib in this prescription as well--then we w i l l have shown that the sib prescriptions are tantamount to the terminological, MoBrcCa, prescription. Das's data do not allow us to conclusively verify either of these statements. In fact, as I have pointed out, he acknowledges 59 that "the sib is no longer the most important exogamous unit" 60 and that i t has been "supplanted by the subsibs" . He proceeds, however, to reconstruct the traditional "marital relations between - 69 -different sibs" on the basis of terms used by the members of a giv-en sib in addressing the members of other sibs. This argument appears in the discussion and appended tables (III to IX inclu-sive) on pages 123-132 of Das's book. In general, Das's argument i s a valuable one. Yet on certain points I remain skeptical. I shall be satisfied to merely il l u s t r a t e the reasons for this skepticism. Das f i r s t provides a l i s t of definitions of Purum kin-terms (terms of address) which, he says "are always used in relation to sibs and never in relation ' 61 to subsibs" . These appear i n table IV, pp. 125-6; the table i s entitled "Terms of Address Used by Men and Women in Respect of Persons Belonging to Their Own As Well As to Other Sibs, When No Direct Relationship Exists." In a l l , he defines twelve term re-lationships according to whether speaker is male or female, whether the person addressed i s male or female, and for each of these four combinations, whether the person addressed is of the speaker's own sib, mother's sib, or of a third category of sibs. I shall cite two examples of such definitions from Das: A man addresses a woman of his sisters' husbands' group' of sibs as Katuna ( i f elder) or by name i f younger. Thus, i f a man addresses a woman as KATUNA, she should belong to a sib from which he may not take a wife since he must not, ideally, marry into a sib from which his sister takes a husband. A woman addresses a man of her mother's group cf sibs as Apu ( i f elder) or by name i f younger. Clearly, a woman should not be allowed to marry an APU. since -her brother takes a wife from Apu's sib(s). - 70 -Tables V-VIII show the terms applied among the members of a l l Purum sibs. Thus, for example, table V l i s t s "Terms of Address used by boys of one sib to the g i r l s of i t s own and other sibs". The remaining tables have the other combinations of sex of speaker and sex of addressee. Combining these tables, Das "deduces" the "traditional" marital relations among the various sibs. Tables III and IX summarize his reconstruction. Das i s well aware that combining tables V-VIII presents several apparent contradictions. He observes that "we find two or three cases where correction was necessary. This correction has been attempted with the help of the internal evidences of the table 62 i t s e l f as well as materials collected in 1932." At f i r s t glance his corrections seem reasonable, but a second look w i l l suggest certain questions and a somewhat d i f f e r -ent interpretation of the meaning of the sib traditions. I shall i l l u s t r a t e the contradictions of his tables with the terms of ad-dress employed by persons of KHEYANG and MARRIM sibs vis-a-vis each other: (1) A KHEYANG man addresses a MARRIM g i r l as KATUNA. This should mean that he may not marry her. Conversely i t suggests that KHEYANG g i r l s may marry men of MARRIM eib. (2) A KHEYANG g i r l addresses a MARRIM man as APU. The definition of APU given above suggests that a KHEYANG g i r l does not marry a MARRIM man. Conversely, i t suggests that a KHEYANG man may marry a MARRIM woman. - 71 -(3) The terms used by MARRIM speakers are more consistent. MARRIM boys and g i r l s seem agreed that women move from MARRIM to KHEYANG. ( 4 ) The women of the two sibs appear in agreement that, when addressing each other, their terms suggest movement of women, from KHEYANG to MARRIM (Table VIII). (5) However, the men of the two sibs disagree, since both term their opposites APU which when male addresses male suggests that the sib of the addressee i s "of his mother's or wife's group of sibs" (Table VII). In sum, the marital relations between KHEYANG and MARRIM, as deduced from the terms of address, would appear to be in great confusion. The same applies, to a lesser extent, in the marital relations of KHEYANG-PARPA and of PARPA-MAKAN. The point I wish to make is not simply that the tables present inconsistencies and should therefore be somehow corrected (as Das has tried to do), but that such inconsistencies are to be expected without any real contradiction to the ideology of sib pre-scription i t s e l f . Perhaps this w i l l become clearer i f we ask who were the informants from whom the information of tables V-VIII was collected? Does the information that "the boys" of KHEYANG address "the g i r l s " of MARRIM as KATUNU derive from only one or from several KHEYANG boys? Indeed,was the informant, or were they, of KHEYANG sib at - 72 -all? If there was more than one KHEYANG male informant and more than one KHEYANG female informant i t i s strange indeed that the boys should a l l agree that a MARRIM female i s KATUNU while the g i r l s should a l l agree, in contradiction to the boys, that a MARRIM male i s an APU. Indeed why should even one informait of each sex disagree? This picture strongly suggests that tables V-VIII were compiled on the basis of information drawn from a relatively small number of informants in each of the sibs, i f there were informants from a l l five sibs. We do not even know that informants from a given sib were always of both sexes. How do we know, in other words, that i f we drew terminology from a substantial cross-section of a l l five sibs, and included informants of both sexes m a l l cases, we wouldn't find that the inconsistencies would abound? It i s my present contention that Das's tables are not drawn from such a wide sweep of Purum informants and, further, that the measure of consis-tency which i t does achieve is a result of the small number of informants. Of course, Das himself i s the only person who can t e l l us whether this i s correct. Until such information i s avail-able I shall assume my hypotheses correct in order to suggest a somewhat different interpretation of the Purum ideologies. - 73 -CHAPTER 4 An Alternative View of Purum Society The following hypothetical situation w i l l i l l u s t r a t e an alternative interpretation of Das' data. Two men, Ego 1 and Ego 2, are both members of the KHEYANG Sib, but Ego 1 i s of the KHULEN village branch of the JULHUNG l i n -eage while Ego 2 i s of the TAMPAK village branch of the AIHUNG l i n -eage. Who knows, the two may never have met? If asked, Ego 1 would claim that he addresses women of MARRIM and perhaps one or two of the other sibs by the terms "U ( i f elder) or as Kanaunu i f younger." It so happens that Ego l's own mother i s a MARRIM. He reports that he w i l l someday marry a KANAUNU. Ego l's sister, on the other hand, reports that she addresses the men of these same lineages as APU and the women "as U ( i f elder) or by name i f youn-ger." She wouldn't dream of marrying an APU. We now speak to Ego 2. He reports that he addresses the women of MARRIM by the term KUTUNU and would not consider taking such a woman as his bride. His mother i s a MAKAN and he addresses women of MAKAN as KANAUNU, or U i f elder. As i t happens Ego 1 and Ego 2 both term MAKAN women as KANAUNU even while they apply opposing terms to MARRIM women. Ego 2's sister i s consistent with her brother; she would marry a MARRIM man, but not a MAKAN. The upshot of such a situation is that two informants of the same sib would provide l i s t s of sibs into which they may and - 74 -and may not marry, both would supply terms of address consistent with their sib prescriptions, and, further, i f each was asked how i t was determined which sibs he may and which he may not take a bride from, each might well answer that the determination was es-tablished by tradition. Furthermore,neither would be mistaken Kheyang Sib Marrim Sib Figure 9 Fig. 9. Large arrows indicate direction of movement of women between sibs. unless he claimed that what applies to him also applied to a l l men of his sib. I t is by no means clear that this i s i n fact what Das's informants said or meant. They have apparently claimed, each, that I (or we) am (are) prescribed to marriage with women of one or more of a s t r i c t l y limited set of other sibs. I t i s not clear to me that the hypothetical'We"would apply to a l l - 75 -other men of his sib. I suspect that the nearest one could come to translating the meaning of whatever Das was told would be best expressed as "we" means "my fathers, my brothers, and I." Under some circumstances this might be conceived as including a l l of the men of a man's sib. Under other circumstances i t may mean a group considerably smaller than a sib; perhaps the Purum subsib (lineage), or, what is even more probable a grouping smaller than either of the named units and one which i s socially more close-knit or corporate. It i s considerations such as these, I suspect, which led Leach to argue that the alliance groups are "local descent groups" of three generations maximum depth. His major point i s that groups which successfully impose marriage prescriptions upon their members, for whatever reason, p o l i t i c a l alliances or other-wise, must be sufficiently corporate not only to be able to sustain (and sanction) the prescriptions, but to be able to successfully communicate them to the members. Indeed, i t is d i f f i c u l t to base a theory upon the assumption of p o l i t i c a l or status interests of groups i f the groups haven't sufficient p o l i t i c a l unity - - o r solidarity -- to have a p o l i t i c a l or status interest. Certainly the Purum sib does not appear to be such a group. The subsib, while more solidary, may not always be the grouping whose members, at a given time, share the same sib pre-scriptions. With respect to sibs at least, the following state-ment from Das bears out these suggestions: The members of a subsib regard themselves to be related by blood and this relation i s more intimate than that - 76 -which subsists between members of different subsibs of the same sib. The subsib i s either an expanded family or an embryonic sib. Perhaps i t i s more the l a t t e r . " This also tends to support one of the implications of my hypothe-t i c a l case. In a sense Ego 2's sister i s Ego l's sister. Her marriage to a MARRIM man would appear, then, to break the norm that a sister does not marry into the same sib as her own brother. Again, a category such as sisters i s — like any ideology — fun-damentally ambiguous. It means only what the culture participants intend, not what the anthropologist would like i t to mean. Indeed, a single term can be used in quite contradictory senses by a sin-gle actor, especially when he uses i t in reference to different persons. This ambiguity in the meaning i s , I think, especially true of kin-terms both in the sense that the people the term i n -cludes and the behaviour such a term implies are subject to a great deal of change through time and variation at a given time. This does not mean that there are no regularities in the use of kin-terms. Rather, i t means that these regularities are of a more general nature than can be ascribed to the non-corporate or only semi-corporate groupings which may constitute an informant's expressed view of his own social structure. I submit only that while i t is certainly possible that such groupings may regulate the distribution of kin-terms employed by their members and, hence, regulate their marriages, i t i s only a special circumstance when they do so. This circumstance requires a measure of corporateness which we do not see in the Purum sibs. - 77 -If my suggestions are correct, there i s no Purum ideo-logy which prescribes uniformity to a l l members of a sib. As a result, there can be no ideology which amounts to one-way bride exchange between whole sibs. In turn, no mechanical model, pre-mised upon an ideology of one-way movement of women between ide-ological Purum groupings, can be made to explain the rule of prescriptive matrilateral cross-cousin marriage. Equally, there i s nothing which even appears to be an ideology of alliances re-lating to the Purum subsibs. Let me pause to summarize the foregoing analysis: (a) Each Purum man is prescribed to one or more sibs and i s proscribed against marriage with any woman from the remaining sibs. (b) Each Purum man classes a l l of the women of his prescribed sib(s) as terminologically uniform (MoBroDas) and as terminologi-cally different from a l l of the women of the remaining sib(s). (c) Each Purum man, therefore, i s prescribed to marrying a woman whom he terms MoBroDa. (d) It is not the case, however, either in fact or in what is unequivocally known of Purum ideology, that a l l men of a given Purum sib are prescribed to the same combination of other sibs and proscribed from the same remainder. (e) Thus we do not find one-way movement of women between entire sibs either as a Purum ideology or as a de facto event (see Fig. 9). (f) We do find a s t a t i s t i c a l tendency toward one-way move-ment of women as between Purum lineages. We do not, however, find a corresponding norm (ideological). - 78 -The question remains, however, as to how a given Ego's sib prescriptions are determined. At this point, the want of suf-fi c i e n t specific data, especially of the kind which traces genea-logies through a span of some generations, makes i t d i f f i c u l t to provide more than a general answer to this question. In the f i r s t place, Ego i s automatically proscribed from marrying his l i t e r a l FaSiDa, and is permitted to marry the l i t e r a l MoBroDa. As we have seen, the Purum tend, apparently, to apply a given kin-term throughout an entire sib. Thus, Ego would treat a l l women of his own mother's patri-sib as terminological MoBroDas, while he would address a l l women of his actual Father's actual sister's own daughters' sib as terminological FaSiDas. Which of the remaining sibs would be placed in which of these categories would depend entirely upon what other people Ego is inclined and/or enjoined (by his socially most near relatives) to regard as fathers of social importance to him. Other MoBroDas w i l l belong to the sibs of other mothers and these are the wives of other important fathers. Likewise, other FaSiDas would belong to the sib(s) of the actual daughters of the actual sisters of Ego's other soc-i a l l y important fathers. It i s clear, then, that a l l decisions pertaining to Ego's sib prescriptions and proscriptions are con-tingent upon what men of his f i r s t ascending generation he regards as important fathers. It i s very easy to imagine that a very large number of variables would affect this arrangement. Under some circumstances, the group of important fathers may include no more than the l i t e r a l - 79 -brothers of Ego's actual father. In a situation where a set of actual brothers are socially close for whatever reasons — po l i -t i c a l , economic, even sentimental -- a set of male patrilateral parallel cousins w i l l share the identical combination of sib pre-scriptions. One can also imagine different circumstances under which the corporate group expands for economic, p o l i t i c a l , m i l i -tary or some such circumstances, to include a larger set of men whose sons are enjoined to treat one another as close brothers and the men of their senior generation as close fathers. The Purum subsib may be the Purums' own ideological approximation of those solidary social units whose members tend, as a function of this solidarity, to share the same sib prescrip-tions and proscriptions. Yet, the de facto corporate groupings may, under various influences, expand and contract around this ideal level even to the point where, through time, the groups which share uniform marriage rules vis-a-vis other groups may be of a very different nature than had once obtained. It could even come to pass that a new group concept might be introduced into the ideology. It should be noted, in passing, that the size of the solidary grouping could be considerably influenced by any phenomen-on which should reduce the number of maritally available women. That i s , reduction of women available for marriage might cause a person to manifestly ignore a relationship -- a classificatory FaSi -- which would preclude a marriage that ego desires. Whether or not this would be regarded as a breach of the FaSiDa proscription - 80 -would be essentially a p o l i t i c a l issue. He would be, i n effect, not acknowledging the social importance of certain classificatory relationships in his marriage choice. By doing so he would, from his viewpoint and that of his supporters, be correctly obeying the marriage rules respecting kin-categry as well as sibs. I t i s conceivable that such decisions could themselves become the source of a group fission. On the other hand i t could even be employed as an intentional device for precipitating or symbolizing such a fission. As a whole, this i s a somewhat different model of Purum society than Needham's. I take as empirically given that each Purum man has a set of sib prescriptions and proscriptions. I also take as empirically green that each Purum man locates clas-sif icatory relationships of the various kinds according to his classification of sibs. It follows from this that since MoBroDas are women of ego's prescribed sibs, he i s prescribed to marriage with a MoBroDa. This i s clearly a mechanical model since both the premises and conclusion are Purum ideological norms whose subject-elements are a l l Purum men. That i s , the scale of the "phenomenon" i s the same as the scale of the elements of the model. This model departs from Needham1 s in several respects. The prescriptions of Purum individuals to a combination of sibs i s not a simple function of his membership in a given sib. I submit that there i s no explicit and clear ideology defining the group of men who must behave uniformly in their marriage decisions v i s -- 81 -a-vis other sibs. To argue the contrary i s to claim that the application of a given kin-term to a number of people means — in the eyes of the Purum that a l l these people must be treated exactly alike. Not only i s such a conclusion false in terms of the actual behaviour of the Purum, there is no reason to believe that the Purum believe this in their ideology. The most that can be said is that, as ideologies, classificatory kin-terms are only very broad and essentially ambiguous frameworks for relationships. It i s quite possible that, to the Purum, the word father (KA-PA) in i t s classificatory usage i s almost as ambiguous as our word "cousin". In some situations there is a clear recognition among culture participants of sharply different degrees of social near-ness among the various persons to whom Ego nevertheless applies a single term. A pointed example of this obtains among the Murngin. Lloyd Warner writes: A definite distinction i s made between actual father and Father's brothers, blood sons and brothers' sons; between father's own brothers and clan brothers of the father. Clan solidarity sharply divides fathers within and out-side of the clan. There may or may not be differences in the emotional attitude of a father and son from a nearby clan and a father and son from a more distant clan. The usual behaviour between a father and son gen-erally does not hold between a distant son and father; frequently because of the opportunities of competition for women among distant groups, there are actual animo-sity and warfare.64 A general consequence of my approach i s that the Purum maintains a norm of matrilateral cross-cousin marriage without any corresponding ideology which can be readily translated to a p h i l -osophy of alliances of a quasi-political nature aligned to one-way movement of brides between whole-groups. - 82 -The measure of deviance from the cross-cousin marriage rule i s estimated, by Needham and his supporters, by the degree of failure to avoid direct exchange between Purum lineages. This assumes that the Purum themselves define their cousin categories in terms of lineage membership. This is manifestly untrue since they define their cousins in terms of SIBS. It also assumes that a l l members of given lineages define their various cousins in the same way. While this i s nearer the mark than the f i r s t assumption we would no doubt find that the whole groups which behave uniform-ly in marriage choices are even smaller than lineages. Here, Leach's concept of the local descent group could prove useful. In sum, a matrilateral cross-cousin i s , to Ego, a person that Ego and perhaps a small set of his near relatives define as such. She jls not the person the anthropologist would like her to be. The upshot of this i s that the measure of deviance from the cousin marriage rule is not reflected by failure to avoid whole-group direct exchange between either sibs or lineages . A l -together i t would be very much lower (the deviance) than Needham, et a l , are obliged to conclude that i t i s . The above modifications of the exchange theory model of matrilateral connubium clearly rob i t of i t s explanatory punch. We are l e f t now only with 6the marriage rule (cross-cousin) i t s e l f and the view that the basis for determining which people belong to what cousin category can be far more individualized than i s allowed for in the exchange theory. Let me now remind you of a remark from Needham with which I introduced this section (vide supya, P. 8 )• - 83 -The relationships of a f f i n i t y are not merely ties be-tween individuals and families. Descent groups, whe-ther localized or dispersed, are also related as groups by ties of prescriptive alliance. I have argued, to the contrary, that among the Purum and perhaps moreso among other cultures having a matrilateral cross-cousin marriage rule the size and nature of the group which influences Ego's marriage choices ( i . e . , the membership of his own cousin categories) can vary greatly from place to place — among the Purum — as well as from time to time. In addition, the bride-giving group may vary not only i n a similar manner, but may be different from the size and nature of the bride-taking group at a given time and place. Thus, the members of a Purum lineage may behave uniformly as a bride-taking group vis-a-vis a combination of Purum sibs. Finally, we can conceive of a model of matrilateral con-nubium in which the uniformly acting bride-taking groups are, in-deed, regularly no larger than small family lines composed of three or even two generations of male siblings and their sons. My conclusion, based upon the argument of this entire section, i s that a theory of matrilateral cross-cousin marriage must be concerned with two related but definitively different issues: (1) F i r s t l y , and most important, why i s i t that a given cul-ture has a norm whereby every man i s permitted to marry his own mother's cwn brother's own daughters while, at the same time, he - 84 -i s proscribed from marrying his own father's own sisters' own daughters? (2) Secondly, what other people, i f any, are identified by a man as his classificatory MoBroDas and FaSiDas, and what are the general structural principles of this identification in a given culture? It w i l l be clear from these questions that my problem defines matrilateral cross-cousin marriage i n such a way as to include not only the classically prescriptive situations of Need-ham which i s confined to situations where classificatory cousins are located in terms of whole-descent-group membership, but also hypothetical preferential systems in which the only members of either cousin category are either f i r s t or at most second genea-logical cousins and where one category i s only avoided as very undesirable. In the following section I shall propose some answers to both of these questions and, in addition, apply the general prin-ciples to a hypothetical model of patrilateral cross-cousin marriage rules as well. Before I leave this section, a f i n a l word concerning the nature of explanation in structural Anthropology. I have tried to make clear, through this section, the kinds of models which I re-gard as most explanatory in our discipline. These are models whose major premises are propositions which predicate behavioural dis-positions to human actors. Any such predication i s taken to - 85 -apply equally to what people actually do as to what they merely say they ought to do. Thus, de facto behaviour patterns and ide-ologies alike f a l l under the aegis of this kind of explanation. In every social situation known to Anthropology there are norms of ideology confronted with only a less than perfect measure of obedience to the corresponding norms. As I have tried to show in the Purum material, however, we must understand what kinds of de facto behaviour these ideological norms imply before we begin to estimate degrees of conformity/deviance. To achieve this end effectively we must make greater efforts to do as Boas had recom-mended many years ago, i.e., "see the culture from the inside." Pursuant to this, explanation of social situations in which de facto behaviour deviates, in whatever measure, from a corresponding explicit cultural ideology requires a twofold ex-planation. Firstly we must explain the most frequent behaviour. In the easiest situation, i t conforms to a corresponding ideology. We are oriented to discover and describe both l i f e circumstances (environmental) and human dispositions which combine to generate these model forms of de facto and ideological behaviour. We are then oriented to treating deviant cases as instances which, like a l l the rest, are also subject to the same dispositions which ex-plain the mode except that they are subject to additional and dif-ferent l i f e circumstances which provoke additional dispositions that negate the former ones. This is somewhat like saying that a l l physical objects are equally subject to the force of gravitation, - 86 -but that some have a greater disposition to oppose i t owing to the shape and size of their surface in the face of air currents. In the case of human beings, the opposition of disposi-tions in a single actor is known as ambivalence. If the reader has not been convinced of the phenomenon of ambivalence through simple introspection perhaps he will be prepared to accept i t for pragmatic reasons; no explanatory hypothesis has been able to ex-plain more in the sciences of human behaviour. Where we argue that one set of actors behaves in a given way as a result of certain kinds of dispositions, and that deviant cases are deviant because of opposed dispositions, we do not con-clude that either set is entirely free of the opposed disposition. To the contrary we will predict that certain aspects of the deviant actors* behaviour do indeed accord with dispositions which they do not predominantly manifest. Likewise, we would be generally oriented to predict that various kinds of changes in their l i f e circumstances would provoke an even wider range of behaviours that accord with the disposition they do not at present predominantly manifest. Both of the above views, that humans manifest contradic-tory aspects in their behaviour and that different circumstances will evoke opposed responses seem to me statements which, when tested, demonstrate the phenomenon of ambivalence. - 87 -CHAPTER 5 Basic Concepts: Continuity and Trusteeship The present section outlines the central features of a theory concerning the sexual relations between a man and his var-ious cousins in societies having one of several cross-cousin marriage norms. Each type of cousin is treated as a distinct relative so that we are dealing with four dyads; MoBroDa/FaSiSo, FaSiDa/MoBroSo, FaBroDa/FaBroSo, MoSiDa/MoSiSo. Each distinctive system of normative cousin relationswith which I shall deal is characterized by a distinctive combination of sexual proscriptions and non-proscriptions among the four dyads. I propose to suggest theories of each of the following systems of unilateral cross-cousin marriage. (1) Matrilateral cross-cousin marriage: in the following argument this means, (a) the proscription of sexual relations and marriage between a man and (at least) his l i t e r a l FaSiDa, FaBroDa, and MoSiDa, and (b) non-proscription of sexual relations and mar-riage between a man and his l i t e r a l (at least) MoBroDa. (2) Patrilateral cross-cousin marriage: this means (a) proscription of sexual relations and marriage with each of a man's l i t e r a l (at least) MoBroDa, MoSiDa, and FaBroDa, and (b) non-proscription of a man's l i t e r a l (at least) FaSiDa. In general, i t is my most basic contention that these various cousin marriage rules are f i r s t and foremost regulations of sexual behaviour. The marriage rules that I am concerned with - 88 -are ones which contain, as an explicit aspect, features which can be termed rules of incest. The marriage rules are treated as secondary to the rules regarding sexual behaviour explicitly. In other words, where a man is proscribed to marriage with his FaSiDa, i t is for the same reason that he is not allowed to have sexual relations with her. This, in turn, is regarded as a prohibition of the kind usually termed a rule of incest; breach of the rule is typically regarded as a somehow contaminating act. We are aware that different kinds of incest are variously repulsive to culture participants — perhaps a mother-son union is frequently felt to be the most horrific. Likewise, varying degrees and kinds of sanction may be imposed in different cultures and on different occasions in a given culture. It is my second contention that various systems of cousin sexual relations are manifestations of a combination of several variables. These are of two general kinds: (1) Intra-familial incest taboos: these are not really var-iables since they are manifestly the same from system to system within our universe of discourse. (2) A set of what I- shall term principles of sex-role iden-tification. These are propositions which predicate general kinds of behavioural dispositions to specified actors in interaction with others of specified status vis-a-vis those actors. I use the expression behavioural dispositions rather than simply roles because many features of the behaviour imputed to the actors do not them-selves represent conscious norms of the actors themselves as is, - 89 -in my opinion, typically the case in usage of the word rule. While many of the acts which the actor manifests may indeed be conscious norms, that they reveal an identification process is not always a clearly conscious norm, i.e., the process of identification itself is a key disposition which is not a clearly elicitable consxous norm. These principles, as I shall show, vary in their particu-lar combinations from system to system. In the following discussion I shall tend to move from general orientation-remarks to more specific definitions. I shall proceed immediately to a discussion of the principles of sex-role identification. I distinguish two varieties of these principles. First, there are principles of sex-role continuity or, for brevity, prin-ciples of continuity. Second, I shall refer to principles of the equivalence of same-sexed siblings or, for brevity, principles of equivalence. Exposition of the principles of continuity will sound at times like a quasi-genetic theory. It must not be interpreted in this sense since I am concerned with a process of the adoption, by persons of one generation, of specific roles (relating to sexual behaviour) revealed in the behaviour of persons of the generation immediately senior. Clearly, a process of role-repetition from generation to generation need not be given a biological rationale (although I personally suspect that this often occurs.) - 90 -In the present context I wish only to propose that, per-haps universally, there i s a tendency for persons of an adult status to treat some persons of their junior generation as an extension of their own self-image. Whatever else such a process might mean, i t i s surely man's principle response to the prospect of his per-sonal death. Conversely, children are moulded, as far forth as i s socially and personally expedient, "after the image" of some one or more individual adults who take i t upon themselves to protect, and train, and nurture a child. I freely admit that the "expediency" i s variable from person to person as from culture to culture. This does not in the least obviate that i t i s a powerful impulse in most persons of a l l societies. Even i n social situations which feature a high measure of individual mobility (status and spatial) this tendency i s clearly recognizable in the "conservative" ele-ments of the society. One further dichotomy; I am concerned with two principles of continuity: these are the sex-role continuity of father and son and the sex-role continuity of mother and daughter. Let me expand a l i t t l e on the concept of sex-role continuity. By sex-role continuity I mean that a child w i l l tend to assume the sexual orientations of some one or more adults. Most especially, the child i s obliged to take upon i t s e l f some or a l l of the sexual taboos proper to that adult. The meaning of "pro-per to" w i l l be c l a r i f i e d later. Likewise, there may be a tendency for the child to assume certain of the adult's sexual privileges, - 91 at least in the sense in which the child is allowed to find a sexual partner in the same category of women as is permitted the adult. There are certain obvious limits to this last generaliza-tion. The adult's permitted category will include women inacces-sible to the child by reason of age difference, and vice versa. In addition, the adult's category will include his or her own spouse who is expressly and specifically proscribed to the child. There is a second aspect to this principle. A dyad which is continuous in the above sense is also continuous in terms of sexual accessibility or non-accessibility from the viewpoint of a given third party. Where the dyad involves a parent and child, this feature of continuity is always broken by the spouse of the parent. From this aspect, the continuous dyad is treated as a single kind of sexual object in the eyes of most others. In the first definition the two of the dyad are treated as sexual actors. For our purposes, the two aspects are mirror images. It will be apparent that where a sex-role continuity can be shown to exist between persons or reciprocal status (e.g. Father/ Son) we will anticipate continuity in roles of other kinds as well as in sexual behaviour; e.g., in work, religious and ceremonial, leadership, and other roles. Our immediate concern however is to explore continuity specifically in the realm of sexual behaviour. Principle of Father/Son Sex-Role Continuity; This describes circumstances in which a man is treated as the sexual equivalent to his own father with the obvious excep-- 92 -tion of their mother and wife respectively. By the same token, Ego assumes the sexual proscriptions proper to his father, and perhaps sexual privileges as well. Certain of these shared proscriptions would be established independently, as i t were, of the process of continuity; notably Ego's proscription to his own l i t e r a l sister. Others, i t shall be argued are a function of the Father/Son iden-t i f i c a t i o n process. As I have indicated the process of sex-role sharing i s one aspect only of a more generalized identification process. The postulates relating specifically to sex-role sharing w i l l be seen to be premises in our cousin-marriage systems models. The pheno-mena of role-sharing in a more generalized sense w i l l be considered corroborative evidence of sex-role continuity in particular. That i s , we do not ascribe a continuity in sexual roles between persons unless there i s evidence of identification in other aspects of their lives. This w i l l be illustrated later. Principle of Mother/Daughter Continuity This situation parallels Fa/So continuity except that the role sharing applies between a woman and her daughter. If, in a given society, i t can be supported that such an identification process i s at a l l operative, i t may be found that the process obtains more between Fa/So than between Mo/Da, or perhaps not at a l l between Mo/Da. Conversely, the Mo/Da principle may obtain in the absence of a parallel Fa/So continuity. What i s the basis for regulation of the sexual l i f e of the child of opposite - 93 sex of a line of continuity where such continuity does not apply, as such, to persons of the child's sex? For example, how i s a g i r l ' s sexual l i f e regulated in a society characterized by Fa/So continuity but not by Mo/Da continuity? I shall refer to such situations as ones in which one line of sex-role continuity pre-dominates over the other line. The lines are identified as either male-continuous or female-continuous according to whether the con-tinuity principle applies to the Fa/So or Mo/Da dyads respectively. As a f i r s t formulation of an answer to the question posed above I suggest that persons of the sex opposite to the predomina-ting line are treated by culture participants as belonging to that line, being linked with i t at, at least, one point and perhaps more. A word of caution must be injected here. The word predom-inating does not necessarily imply either p o l i t i c a l or status pro-minence of the sex of the predominating line. It means only, as I have said, that the process of sex-role continuity applies to one sex, e.g. Mo/Da, conspicuously more than to the opposite sex. Thus, where one line i s conspicuously predominant, mem-bers of the opposite sex are viewed as belonging for purposes of  sexual disposition to some one, or perhaps more, persons of a line of continuity. Here, I expressly avoid the concept of Jural Authority. Through the Needham vs. Homans and Schneider debate, this concept has come to imply more than I am prepared to affix to the notion of belonging for purposes of sexual disposition. Radcliffe-Brown 1s concept of jus in personam strikes nearer the - 94 -mark. At the same time, i t has important features of his jus in 65 rem. This means that where Fa/So continuity predominates (Mo/Da continuity does not obtain), a man treats his own daughter's sexual disposition, vis-a-vis other people, as his own right and duty; as a sexual object. She is his for disposition. This ap-plies not merely to her marriage but to her pre-marital sexual l i f e as well. A man assumes the right to regulate the issues of with whom his daughter may or may not have sexual relations. It means that a man may impose sanctions upon his daughter as well as upon her lover for a sexual union of which he disapproves. This sexual trusteeship continues at least until the time of his daugh-ter's marriage. At this time the rights are at least partly transferred to the g i r l ' s formal husband, and he as a rule con-fines sexual access to the g i r l to himself. The father may be expected to continue his trusteeship in a limited sense beyond marriage. This is well illustrated in the following statement concerning the Murngin: If a daughter runs away from her husband (the father's Waku), the father with the_aid_of his sons brings the runaway back to his Waku /SiSj?/. She receives a beat-ing from him and i s advised to remain faithful to her husband. Quarreling daughters who are the wives of one man are instructed by their father to live in peace, for he considers the welfare of his Waku of the greatest importance.66 And in extreme cases of i n f i d e l i t y : If a wife continues to be unfaithful she might be ki l l e d by her due' /FaSiSo = Hu or Brother of Husband/ and mem-bers of her own family.^ - 95 -Principle of the Sexual Equivalence of Sisters. Sisters, at least g i r l s who are daughters of a single * sire, become equated as sexual objects, and in a sense as actors, as a function of having one and the same sexual trustee — their mutual father. His attitude regarding the proper disposition of one, in terms of the status of what he w i l l consider an acceptable sexual partner for her, w i l l apply likewise to the others. Like-wise, they are viewed as equivalent sexual objects by third par-ties in the sense that a man may deny sexual access to his daugh-ter's categorically to that third party. This principle of sister equivalence does not pretend to countenance no contradiction. The sexual access of a man to his own wife does not imply his automatic access to her sisters. At the same time i t does help to account for such phenomena as the sororate and sororal polygyny. Also, i t does mean that the s t r i c t sexual proscription to a woman implies the same to her sisters . It w i l l also help us to understand sexual taboo of the matrilat-eral parallel cousin. The Sharing of Trusteeship In such systems as these, having predominance of the male-continuous line and father/daughter trusteeship, we anticipate that, as one aspect of the more generalized Fa/So role-continuity, a man w i l l share his own father's role of trustee vis-a-vis their sister-daughter (respectively). This is not implied by the defi-- 96 -nition of Fa/So sex-role continuity but i s predicated as a ten-dency corresponding to the anticipated generalized identification process. I t i s also interesting to speculate that, where such a role-sharing occurs, i t goes far to explain the brother-sister incest taboo. That i s , given the father-daughter incest taboo, and given father/son sex-role continuity, the brother-sister sexual taboo is accounted for as the son's adoption of a proscription proper to his father. This kind of an approach allows us to anticipate c i r -cumstances (lack of Fa/So identification) in which the brother-sister taboo might be ignored even i f the Fa/Da taboo i s stringently res-pected. In addition, I propose that the role of sexual trustee i s , with the important exception of the Husband role, incompatible with sexual union between the trustee and his trust. This w i l l be taken up again at a later point. However, given this, and the sharing by a man of his father's trustee role over their sister/ daughter, we have supplied an additional reason for a brother/bister incest taboo. Just as an example of shared trusteeship, r e c a l l the c i -tation above concerning the runaway wives. Such a g i r l , among the Murngin, i s retrieved and punished by the father "with the aid of his sons." This statement clearly implies primacy of the paternal trusteeship, but just as clearly specifies that his sons must participate in the role. - 97 -Female Ego and her Father's S i s t e r In a s i t u a t i o n i n which, as we l l as sharing h i s father's own sexual o r i e n t a t i o n s , a man al s o shares h i s father's trustee r o l e s (remember that the former must not be taken to n e c e s s a r i l y imply the l a t t e r c o n d i t i o n ) , we are led to a n t i c i p a t e a tendency to the sexual equating of a woman and her Fa S i . A man i n t h i s case w i l l assume a trustee r o l e v i s - a - v i s both h i s own s i s t e r and h i s own daughter (rather than only over h i s daughter). Thus, a g i r l and her FaSi are i d e n t i f i e d as belonging f o r purposes of sex-ua l d i s p o s i t i o n to a single given man at l e a s t i n the sense that one of the men to whom each i s a sexual t r u s t i s the same person. In such a s i t u a t i o n we a n t i c i p a t e the tendency f o r a woman to be sexually oriented i n the same d i r e c t i o n s as her father's s i s t e r ; that i s to behave as sexually continuous with the FaSi as actor and as object. This means that we a n t i c i p a t e she w i l l adopt at le a s t the sexual p r o s c r i p t i o n s proper to her father's s i s t e r s . Any circvmstance which attenuates the sharing by a man of h i s father's trusteeship roles w i l l attenuate t h i s FaSi/BroDa sex-role continuity. Before I move on, l e t me remind you that we have been dealing expressly with a s i t u a t i o n showing predominance of male-continuous l i n e s . We may also r e f e r , here, to a sexual equivalence of brothers. This i s d i r e c t l y implied by t h e i r mutual continuity from the same s i r e . However, what about a s i t u a t i o n i n which there - 98 -i s predominance of a female-continuous line? Remember, we do not imply p o l i t i c a l or status dominance of women; only that Mo/Da sex-role continuity obtains where Fa/So continuity i s submerged. Predominance of a female-continuous line This situation would parallel the one of male-continuity in the obverse sense: (1) A woman is sexually continuous with her mother, sharing the mother's sexual taboos and perhaps her general sexual p r i v i -leges. (2) Brothers are sexually equated under the common trustee-ship of their mother. They, as i t were, derive from the same womb. For obvious reasons, they cannot, like their sisters, be equivalent sexually as continuations or duplications of the maternal womb. (3) The measure of transmission of the trusteeship role over boys from their mother to their sister i s a contingency upon which depends the measure of sex-role continuity that obtains between a man and his mother's brother. We expect that one factor w i l l complicate this situation in a greater measure than the case of male-continuous lines. Most often, jural authority over young men is exercised by the mother's brother even more conspicuously than by his sister, i f not moreso than his mother. Yet, I contend that this jural authority derives from the boy's mother. The MoBro is what Radcliffe-Brown terms a "male mother". - 99 -In these cases the mother and to some degree the sister have trusteeship in the sexual disposition of a young man. This, together with the more generalized features of jural authority (authority in training in male s k i l l s etc.) are in part delegated to the nearest male of the boys 1st ascending generation over whom the mother has a measure of influence. This is the woman's own brother and the boy's maternal uncle. This does not mean that a woman has generalized jural authority over her brother. She does have at least a shared sexual trusteeship in him. Perhaps some il l u s t r a t i o n of our general notions w i l l be helpful at this point. Throughout the remainder of this essay I shall continue to exploit Murngin data, from Warner, for much of my i l l u s t r a t i o n . In the f i r s t place, the Murngin appear themselves to have a f a i r l y explicit attitude of generalized Fa/So role continuity. Warner observes that the daath of a son is a great social loss to the father, since there i s no one to take his p l a c e . ^ Likewise the position of the father as sexual trustee over both his sons and daughters seems clear in the following remarks: As soon as a boy is circumcised and old enough to under^ stand and remember what i s told him, his galle ^loBroDa/_ i s pointed out. At a similar age a galle mielk MoBroDa/ has her due diramo /FaSiSo/ shown to i l e r . This is done by the male parent of both children. and, Usually an older male due" and male galle have an under-standing that their sons and daughters shall marry.70 - 100 -This last remark means that as between a man and his MoBroSo, who may be brothers-in-law (Murngin have a rule of matrilateral c.c. marriage), there is an agreement the son of the former shall marry the daughter of the latter. This would of course be a matrilateral c.c. marriage. My,major point in these citations i s that i t i s the fathers who make the arrangements and who form the agreements — at least formally. These comments may also help to point up that there i s some difference between sexual trusteeship and jural authority. Homans and Schneider have expressed uncertainty concerning whether the locus of jural authority rest, among the Murngin; with the 71 father, the MoBro or the MoMoBro. They include the Murngin as one of two exceptions (Yir-Yiront are the other) to their expec-tation that patrilineal descent w i l l mean patripotestality. Yet they do not feel i t to be a " f u l l exception" (p. 42). On a l l of 72 this, Needham is in agreement. Thus, even in a situation, such as among the Murngin, where the disposition of generalized jural authority i s unclear, i t i s quite possible that the formal locus of specific sexual trus-teeship i s perfectly unambiguous. There can be l i t t l e doubt that the Murngin sexual trustee i s Ego's father. The discussion to this point allows us to posit four major combinations of the general dispositions so far considered: - 101 (1) Fa/So sex-role continuity; paternal sexual trusteeship which i s not considerably shared by the g i r l ' s brother; equivalence (sexual) of brothers and of sisters, but not of the BroDa/FaSi dyad. (2) The same as the above except that there i s a strong tendency for a man to assume his father's trusteeship role vis-a-vis the sister of the man. This would result, ex hypothesi, in a greater tendency to a sex-role continuity between a g i r l and her FaSi. ( 3 ) Mo/Da sex-role continuity; maternal sexual trusteeship over sons which is not significantly shared by the boy's sisters; sexual equivalence of brothers and sisters, but a weak ten-dency to continuity in the SiSo/MoBro dyad. ( 4 ) The same as ( 3 ) except that there i s a strong tendency for a gi r l ' s assumption of the trusteeship role over her brother. This would result in a greater tendency to a sex-role contin-uity between a man and his MoBro. - 102 -CHAPTER 6  Exogamy and Descent In the following I shall attempt to apply some of the principles already introduced to rules of exogamy imposed upon the membership of groups recruited on the basis of a rule of uni-lineal descent. A concept of an exogamous patrilineal descent group has the following features: (1) A patrilineal descent group i s a set of men who recog-nize a common ancestor who i s linked to each member through a tem-porally receding line of males. This line may be viewed as com-posed of dyads of fathers and sons. (2) The women of these groups are the daughters of each of these men (and, of course, their sisters). Collectively, the women w i l l be viewed as sexual wards of the men; each of her father. (3) The rule of exogamy i s viewed as the proscription among the men of a descent group to their mutual sexual wards — i.e., their daughters. This last characteristic in particular needs further clar-i f i c a t i o n . In the f i r s t place, i t does not imply that among a l l men of a descent group or even between a l l members of adjoining generations there i s a more generalized sex-role continuity with sharing, by each, of a l l of the sexual taboos proper to every other. Clearly this would extend the sexual proscriptions of each - 103 -person vastly beyond the membership of his own descent group which is my immediate concern. Thus, the proscription characteristic of an exogamous descent group i s particularly that to the sexual wards of a l l the men composing the group. The kinds of sexual proscriptions which derive from Fa/So sex-role continuity are considerably more extensive, where they occur, than the abstention from sexual relationships (of a man and his son) with their respective wards. As I have already noted this proscription (to trusts) applies doubly in a father/son dyad in which the son also assumes a trusteeship posture vis-a-vis his sis-ter. However, the concept of Fa/So continuity in sexual orientation implies a wider range of sexual regulation than obtains as a func-tion of mutual abstention from sexual x^ards. The latter i s charac-t e r i s t i c throughout the membership of the descent group. The more generalized sharing of taboos implied in Fa/So sex-role con-tinuity w i l l seldom extend throughout the membership of an entire descent group. This category implies Ego's proscription to women whose sexual taboos are proper to his father. This cateogy of women can extend considerably beyond those women who are the father's own sexual wards. This w i l l be exemplified later. How-ever, I should point out that such exhaustive sex-role continuity among a l l members of a descent group is a requirement of the ex-change theory models of matrilateral connubium. Only under this condition w i l l a l l men of a descent group behave identically in their marital relations with other whole descent groups, and only then can the entire group exchange models be applied. - 104 -The foregoing remarks are a way of viewing exogamous social groupings. Behind them i s a theoretical viewpoint con-cerning the basic structural features of enduring associational groups. In particular, i t i s my own argument that the phenomenon of sexual trusteeship i s f a i r l y much a social universal. In ad-dition, i t i s my view that there i s a widespread and forceful disposition among persons who compose a corporate grouping with a substantial measure of endurance to refrain from sexual rela-tions and marriage with their respective sexual wards. The key words here are corporate and enduring. To put i t in simple words, I would argue that any condition which w i l l cause a size reduction or size expansion of a group whose members undertake any close corporate interactions w i l l tend to create a corresponding change in the size of the exogamous grouping. Such corporate interactions may center upon any of a variety of c i r -cumstances most noteably economic and/or politico-military. Thus, for example, there seems l i t t l e doubt that, among the Purum, the sib was at one time a s t r i c t l y exogamous social unit but that i t had undergone change in this respect. Das observes of a number of cases of sib endogamy that in spite of their non-observance of the exogamous rule in connection with the sib they were neither ex-communicated nor punished in any other way for this social offence. They continued to remain members of Purum society and lived in a Purum village taking part in a l l socio-religious r i t e s and ceremonies performed by the village community for the welfare of i t s consti-tuent members. In short this was not regarded as an offence at a l l . But this does not seem to have been the case at the time when Shakespeare wrote on this - 105 -tribe. According to him "Among the Anal, Purum and Lamgang marriages must be made within the clan but not within the family."21 Shakespeare's 'clan' refers to our 'tribe 1 and his 'family' is identical with our 'sib'. Even nowadays whenever a Purum i^ s asked whether a man can marry within his sagei /sib/ he at once denies i t and this i s the general feeling of the community. But the instances under reference show that the intensity of this feeling has diminished to a very appreciable extent and the sib is no longer the most important exogamous social unit. The idea of incest which gener-a l l y accompanies the sib-concept has now been trans-ferred in Purum society from the sib to the subsib.73 What 1 am suggesting i s that, had we the information, we would be able to demonstrate a parallel reduction in the measure of sib corporateness. *^his might well be reflected in i f not caused by reduction of economic and p o l i t i c a l cooperation of sib members and perhaps by a growing measure of sib dispersion consequential to the apparent t e r r i t o r i a l expansion which the Purum were undertaking. Conversely, this viewpoint argues that any circumstance which inspires sexual union and marriage among the members of an ideally exogamous unit w i l l tend to have a divisive effect upon the unit. Thus, even should a named unit continue, a while, to be so named and recogniaed, the corporate and cooperative interactions of i t s membership w i l l be substantially reduced, or what may amount to the same thing, the size of the groups whose members do cooperate closely w i l l be reduced to sets of people smaller than those which compose the ideal exogamous unit. Any condition which, for ex-ample, reduces the availability of women over the entire population might result in fragmentation of the earlier exogamous corporate units into smaller corporate groupings. - 106 -Since a l l of this is bare-faced speculation I had just as well carry the argument to i t s extreme. Why should one anti-cipate what amounts to the incompatibility of enduring corporate-ness and a persistent endogamous pattern among the group members? Anthropologists have long been inclined to view marriage and sex-ual relations of man and woman as events which have profound ef-fects upon the social relations of man with man. Specifically, the occurrence of sexual relations between a man and woman always demands that there be some social adjustments made between a g i r l ' s sex — cum marriage — partner and other men associated with her; most especially the men whcml have designated a g i r l ' s sexual trustee. I suggest that whenever two men have claims on the loy-alty of a given woman, they are in a condition of potential vul-nerability vis-a-vis one another. Perhaps even the most casual sexual encounters carry this implication, but i t is certainly an aspect of prolonged male-female intimacy that each of the partners may expose themselves considerably to the other. This tendency may be as readily precipitated in circumstances of sexual intimacy as under any social encounters general to human relations. This i s sheer speculation and i s , possibly, very ethnocentric in i t s source. I f something of this nature i s so, however, i t may prove very useful i n understanding social phenomena of many kinds. Where a woman owes loyalty to her sexual trustee as well as to her sex partner, the sex partner could, in a sense, - 107 -place himself i n a condition of vulnerability before the g i r l ' s trustee. She can become the bearer of information of perhaps the most intimate kinds about her sex partner. I f , in cultures besides our own, men are prone to punctuating sexual intimacy and the sense of conquest that often accompanies i t — with a general ver-bal "beating of the chest", women are placed in a position of great potential power. Such information could often be of a damaging nature should i t become known to other men — such as the g i r l ' s sexual trustees. This i s reminiscent of many male-female encounters recorded both i n mythology as in reliable history. Our own Western histories are replete with cases of treacherous women who employ their feminine wiles toward the destruction of their lovers; Lucre-zia Borgia, Katherine the Great are, purportably, classic examples. More recently, the comical events surrounding the "Profumo Scandal" and the "Munsinger case" are an excellent example of the common assumption that a man is seldom so vulnerable before othajf man as when he shares a sexual intimacy with a woman whose loyalty may be in doubt. A man who i s the trustee of a g i r l i s also vulnerable where, even though sexual intimacy i s taboo between them, the intimacy of the trustee/ward relation may arm the g i r l with infor-mation which i s potentially damaging to the trustee. Thus, a woman may be in a position to transmit intimate and otherwise guarded information between her trustee and sex-partner in either direction. This places each in a position of some vulnerability vis-a-vis the other. - 108 To my mii4 the widely recognized state of tension found so often in af f i n a l relationships in many societies may well de-rive in some measure from this implicit interpretation placed, by men, upon their intimate relations with women. A factor of this na-ture might well become manifest in institutionalized exogamy where the exogamic group comprises an association of males who, for whatever reason, are oriented to a high degree of cooperative in-teraction. Conversely, regular breach of the exogamic disposition would result i n a condition of tension among the men of a group. At least i t might induce sufficient uncertainty to compromise an enduring close cooperativeness among them. I do not wish to imply that cooperativenss cannot be propitiated among affines; only, rather, that i t i s cooperation from a distance such as i s described by the attitude of recipro-city. Affines w i l l exchange produce but are less able to co-produce. I frankly admit that at this point I am unable to draw sharp lines between the aff i n a l and kin cooperativeness. I only say that there i s a difference in their degree and that this may be in part engendered automatically as i t were by their reduced ab i l i t y to control information about themselves. This implies, paradoxically, that a condition of high cooperativeness requires a measure of personal secretiveness among i t s participants. Per-haps this i s what i s meant by our own folk wisdom that "familiarity breeds contempt." - 109 -Perhaps I should make one point clearer. The informa-tion to which the g i r l , has access need not be of a specifically sexual nature. Often, a g i r l i s not only not allowed to have sexual intimacy with her trustee, they mjjst not even speak of sexual matters in one another's presence. On the other hand, the g i r l , has, as a rule, been raised in the camp of her trustee. Many peoples appear to guard important information from their women by means such as outlawing their membership in secret societies. At the same time she may be able to learn, simply through propinquity, the important s k i l l s and even military stratagems of her trustee. In her relation with her sex-partner, the act of intercourse alone may evoke revelations from the man, not to mention the in-formation she might gain simply through liv i n g in the camp of her mate. The other side of this coin i s that, where a sexual trus-tee approves a given man as sex-partner to his ward, he implies some measure of trust in the man. Indeed, such trust i s very com-monly symbolized through so-called "marriages of state"; these frequently punctuate the formal termination of ho s t i l i t y . The foregoing discussion i s not an attempt to explain rules of descent as such. Rather I am primarily concerned with the phenomenon of exogamy which i s so often superimposed upon groups constituted through a rule of urilineal descent. There are, of course, some well-known examples of societies which practice a rule of unilineal descent without ever creating on i t s basis any enduring - 110 -(or stable) corporate groupings. In such situations there simply are no groups to be exogamous. The Arab Bedouin provide an inter-esting example. The Bedouin practice a rule of patrilineal descent to be sure; yet, as Murphy and Kasdan have described them, In this system, i t i s almost impossible to isolate a solidary ingroup, and groupings are continually being activated or redefined through struggles that may even pit members of the nuclear family against each other. Such f l u i d i t y is congruent with the extreme weakness of internally exercised authority in agnatic sections, what-ever the level of segmentation in question.74 They il l u s t r a t e this analysis with the Arab proverb: Myself against my brother; my brother and I against my cousin; my cousin, my bro-75 ther and I against the outsider. In view of the transitory nature of corporate groupings, i t i s d i f f i c u l t to refer to the Bedouin preference for marriage with the patrilateral parallel cousin as an example of endogamy. The marriage preference does mean, however, that the patrilineal segment composed of two brothers and their children is indeed en-dogamous. It is most t e l l i n g that such endogamy should occur in a social structure characterized clearly by potential h o s t i l i t y at nearly any point — including between l i t e r a l brothers who squabble interminably T over inheritance. I shall emphasize this point with one fruther remark from Murphy and Kasdan: Thus, in a system where every male sibling i s a poten-t i a l point of segmentation, and therefore a significant p o l i t i c a l role player, even the interests of brothers or of sons and fathers are not necessarily convergent. . . Cohesive relations between and within sections do not have an enduring, continuing quality, but are situational and opportunistic.76 - I l l -The authors of this fascinating a r t i c l e carry their argument a step further: It i s , then, our hypothesis that the peculiar nature of agnatic sections among the Arabs i s closely related to the practice of parallel cousin marriage. It might also be hypothesized that lack of internal solidarity and homogeneity in Arab kin groups is promoted by the com-bining of affinal and consanguineal ties.77 In other words, they too seem to argue that there i s some source of incompatibility between the creation of enduring corporate groupings and a practice of endogamous marriage. Their explanation of this emphasizes factors which I have not, however. They accept Homans and Schneider's general postulate that people w i l l marry away from jural authority, and, since corporate groups reveal such an authority structure, people marry out of their corporate groups. Thus, since the Arab Bedouin have only a very weak tendency to cor-porateness, and a weak jural authority structure, there i s no jural authority, within even narrow agnatic segments, to marry away from. My own view i s somewhat different. A f f i n a l relationships are inherently tense for the reasons given whereas corporate rela-tions are inherently closer—more cooperative. Nothing i s directly implied here concerning the affectional or sentimental dimension of these relationships. It is even possible, to my way of thinking, to feel great affection for a person whom one may have reason to fear (such as one's affine ) while, at the same time, one may feel keenly i r r i t a b l e towards a person in whom he i s ready to place a great deal of trust (one's consanguineal kin). At the same time we expect sentiment to generally parallel measures of corporate closeness. - 112 -In terms of the concept of trusteeship, then, while the Bedouin undoubtedly reveal dispositions to a relationship of sex-ual trusteeship between a man and his daughter, and while a man's sons very probably assume a measure of this role as well, there i s manifestly no sexual avoidance of the mutual wards of the men who are agnatic kin. A man, as I have already noted, i s s t r i c t l y pro-scribed from having sexual relations with a g i r l who i s his own ward. Thus, brother/sister as well as father/daughter, mother/son sexual taboos are s t r i c t l y maintained. However, the trusteeship role does not appear to extend beyond these members of the immedi-ate nuclear family. In addition, there i s no father/son sex-role continuity as such. A boy does not assume the taboos proper to his father. Thus, the only group within which a rule of exo-gamy consistently is enforced among the Bedouin i s the nuclear family. Conditions of desert l i f e are such that near association among men i s always tenuous and conducted against a backdrop of perpetual competition. These circumstances counter the formation of enduring corporate groups. The relationships among consanguin-eally near agnatic kin are a tissue of mutual distrust to begin with. Thus, not only i s there no orientation to avoiding the es-tablishment of affinal tension among agnatic kin through endogamous marriage, but such marriages are themselves expressive of a pre-existing disposition of only very hesitant mutual trust. Indeed, such marriages, in my view, also have the effect of propagating - 113 -the tension among agnatic kin and of rendering the formation of enduring corporate groups an impossibility short of prolonged and intensive confrontation with mutual enemies — outsiders. In sum, then, a rule of unilineal descent may exist in-dependently of the formation of enduring corporate social groups and without a rule of exogamy that extends beyond the members of the nuclear family. The persistence of intra-familial incest taboos would indicate, in terms of my argument, that a sharing of trusteeship roles and a respecting of their mutual sexual wards between a man and his sons is a part of the general disposition of father/son identification which will scarcely yield under even the most competetive of social circumstances. The converse to this would be that circumstances under which breaches of the incest taboos do occur might be ones involving a substantial breach in the identification of father and son. Let me summarize some of the major points in this dis-cussion of systems which tend to create agnatic clusters: (1) A father is the principal individual having the social role of sexual trustee over his daughter. (2) The relationship of trustee to ward is incompatible with the sexual union of its members. (3) A man's own sons always share (ideologically) their father's role of trustee over their sister. This adoption of father's trus-teeship role may or may not be extended further to include partak-ing of the father's trusteeship over the father's sister. Trusteeship - 114 -sharing may be expected where the closeness of the male association is strong. Thus, a man may or may not share a trusteeship proper to his own brother (i.e., over BroDa). In general as social close-ness is intensified among the men who recognize agnatic ties, so is the measure of trusteeship sharing. (4) There is a firm disposition among corporately near men to avoid sexual relations among their respective wards even where they do not assume trusteeship positions vis-a-vis these women. Thus, a man may assume no responsibility in the sexual disposition of FaSi, but will usually refrain from intimacy with her since she is a ward of his father. Thus, the limits of the effectively corporate unit tend to approach the limits of the exogamous unit; ie., the unit whose members abstain from sexual relations among their respective fe-male wards. We may expect the limits of the assumption of an ef-fective trusteeship role to f a l l short of the bounds of exogamy. Major decisions concerning the sexual disposition of a girl will s t i l l reside primarily with her own father although he may occa-sionally have to call upon the other men of his group to support his decisions in some crisis circumstance such as the retrieving of a daughter who has eloped or been stolen by a man of whom he disapproves. Such a situation could threaten hostilities in which he might require military backing. Before I move to a brief consideration of matrilineal groupings I would like to add one final admission of uncertainty. - 115 -It concerns the vague word corporate. Clearly we are dealing with a very ill-defined notion. What we have been concerned to relate the notion to has been i t s relationship to the phenomenon of exo-gamy. I f , as I have argued, corporateness of a stable kind requires exogamy, i t was because in-marriage would create a kind of relation-ship among men which tends to be incompatible in i t s qualitative character with the character of relationships in a solidary or cor-porate group. The distinctive quality of such kin-relationships resides in the mutual expectancy that, vis-a-vis the rest of the world, a man can ultimately trust in the support and protection of the other members of his group. There are probably no other kinds of relationship which as successfully inculcate this attitude. It i s more d i f f i c u l t to maintain such an attitude among men who avail themselves of the favors of women who are their respective wards. Apart from the possibility of jealousy (even of a sexual nature) on the part of the trustee, the factor of control of highly personal information may part way acconnt for this. It i s probable that a l l human beings guard items of information which concern them-selves against exposure to even the most trusted of their fellows. There is no reason to believe that men of simple tr i b a l social sys-tems are any less vulnerable in their own view than are men of our own social system. Perhaps a l l close associations require that we accept what i s somewhat illusory of others and foster, in their minds, illusions about ourselves. In terms of the concepts I have provided exogamic group-ings based upon a rule of matrilineal descent w i l l have the following characteristics: - 116 -(1) Maternal trusteeship: boys are the sexual wards of their mothers. Such a rple may be shared in varying degrees by the boy's sister; most probably an older sister. This means that the sex-ual disposition of a boy i s held to be primarily the concern of his mother, even though her brother w i l l assist in supporting and sanctioning her decisions. (2) Thus, a matrilineal descent group i s a set of women who recognize a common ancestor (or ancestress) who i s linked to each member through a temporaly receding line of women. This line may be viewed as composed of dyads of mothers and their daughters. In this case, i t i s the women who fundamentally compose the group, whereas the men who are members of i t are viewed as the sexual wards of particular women of the group. As I have noted, the trusteeship role i s primarily that of the mother, but may extend outward from her to include other women of the group — conceivably a l l . (3) As a result of postulates (1) and (2) the rule of exo-gamy is here viewed as at least resulting from the prescription of sexual relations between any women of the group and the particular sexual ward of any of the other women of the group. Where the trus-teeship role i s widely shared among these women, the exogamic rule also obtains by reason of the firm disposition against sexual re-lations with one's own sexual ward. Thus, sons are primarily the property of their mothers until their marriage. Needham observes of the matrilineal BELU, - 117 -for example, that the children "belong" to the wife and "It i s the wife and her brother who together possess jural authority over the children, as they possess rights over land, house, hereditary 78 property, and a l l other things of value". As a rule, the train-ing of a young man in his male roles i s undertaken by a male. In these cases, the job i s undertaken by the nearest male kin of the boy's mother, the mother's brother. Yet the rights of sexual dis-position over a son are, ex hypothesi, ultimately those of the mother in consultation with her brother. It seems that the tendency of most societies to relegate major decision making into the hands of men w i l l result i n a lesser tendency for a boy's sister to assume the mother's trustee role than in the parallel patrilineal situation where a boy assumes father's trusteeship over sister. Thus, we anticipate that we shall seldom find that a woman is a major sexual trustee over her own brother. This w i l l minimize any tendency to sexual continuity between a man and his own mother's brother. That i s , to the extent that they are discretely the sexual property of their own respective mother's and not of a trustee which they mutually share, the pro-cess of their identification may be somewhat intercepted. In gen-eral, I believe there is always a tendency to male dominance i n decision making. This tendency results in reduction of the degree in which men belong to a line of women which includes women of their own generation. As a result, the men of matrilineal descent groups tend to be sexually identified through the process of shared - 118 -trusteeship less than do women of patrilineal groups. Thus we anticipate a greater tendency toward sex-role continuity of a g i r l and her FaSi in patrilineal systems than of a boy and his HoBro in matrilineal ones. This absence of sex-role continuity between a boy and his MoBro does net imply any absence of avun-cular jural authority which, as here understood, i s an authority delegated from the mother. We are now prepared to move onto the subject of models of our various cousin marriage systems. In these models the f o l -lowing intra-familial incest taboos are assumed: (1) Mother/Son. (2) Father/Daughter. (3) Brother/Sister. - 119 -CHAPTER 7  Matrilateral Cross-Cousin Marriage These systems may be nnderstood in terms of the existence of dispositions toward both of father/son and mother/daughter sex-role continuity operating together in a single context. As a function of the former, a man assumes the sexual taboos proper to his father. At the same time, a g i r l assumes the sexual taboos proper to her mother. The primary role of trusteeship over children (sons and daughers) may belong to either the mother or the father although i t i s most often i n the hands of the latter. Where the father i s sexual trustee over his children, his son adopts a supportive part in this function vis-a-vis his sister. In such a system, then, a g i r l i s the sexual ward of her father, subject to his decisions in her sexual career, while she i s at the same time sexually continu-ous with her mother, sharing most notably the taboos proper to the mother. Clearly, the father's decisions are his to the extent that they do not violate the dispositions of Mo/Da continuity. These remarks summarize the basic features of a system of matrilateral c.c. marriage with paternal trusteeship. In the following I shall offer an analysis of the sexual dispositions of Ego vis-a-vis each of his four types of cousin: (1) The Father's Sister's Daughter. According to the principle of Mo/Da continuity, a woman and her mother are sexually equated by a l l men except the woman's - 120 -father (her mother's husband). If mother i s not allowed to have sexual relations with a given man on the ground that i t would be incestuous, then neither may her daughter. Thus, whereas a man is proscribed from sexual relatiofts with his own sister, so is he proscribed from the daughter of his sister. From his viewpoint, his sister and her daughter are identified continuous -- as sexual objects. Conversely, a woman assumes the proscriptions proper to her own mother. Thus, she takes upon herself the taboo to sexual relations with her MoBro. From this perspective, a woman and her daughter are viewed as sexual actors. But, according to the principle of Fa/So sex-role conti-nuity, a boy assumes the incest taboos proper to his father. Thus, he also i s proscribed to sexual relations with the daughter of father's sister. Conversely, a woman identifies her brother and her brother's son as sexual objects; she is thus proscribed to both. Thus, her daughter, assuming the incest taboos proper to her mother is likewise proscribed to her MoBroSo. (2) The Father's Brother's Daughter. A man and his own brother share mutual taboos. This in-cludes both taboos proper to themselves (e.g. against mother and sister) and those which they adopt from their mutual sire. In addition, they are proscribed from sexual relations with the sexual wards of one another -- their daughters. From the viewpoint of each of these g i r l s , as to a l l other people, a man and his brother are sexually equivalent. The son of each of these brothers adopts the taboo to the daughter of the other. - 121 -(3) The Mother's Sister's Daughter. According to the principle of Mo/Da sex-role continuity sisters are equivalent as actors and objects. Thus, the Mo/So in-cest taboo applies equally between Ego and MoSi and, therefore, between Ego and MoSiDa. (4) The Mother's Brother's Daughter. There are only two situations in which I can envision this combination of identity principles resulting in the proscrip-tion of sexual relations with the MoBroDa. These are: (a) sex-role continuity of FaSi/BroDa. Such a situation means that a g i r l assumes the sexual proscriptions of her FaSi including that to FaSiSo, and conversely where a g i r l i s sexually equated with her FaSi by men. The prin-ciple of Mo/Da continuity does not achieve such an identification since a woman and her BroDa are of different lines of women. The circumstance under which there may be a disposition toward this identification is one in which these women are equated through being the common sexual property of a common major trustee. This iden-t i f i c a t i o n , then, could develop where a man's sister i s treated equally as his own sexual ward as is his own daughter. This would mean that the g i r l ' s FaSiSo i s proscribed to her as he i s to his own mother (the g i r l ' s FaSi). In my view this kind of continuity is of a much less d i -rect nature than Mo/Da continuity. A mother and her daughter are more readily sexually identified as though they were the same kind of sexual and maternal principle. - 122 -The union of two women under common trusteeship need not develop into this kind of sexual identification. The p o l i t i c a l interests of the trustee are a more explicit concern in the situa-tion. The union of two women under his dispositional aegis need not mean their complete sexual identification with the consequent sharing of taboos. It is more probable that their identification i s of the kind proposed by the exchange theorists: they are viewed as the women belonging, for purposes of sexual disposition, to a line of men, the line consisting at least of a man and his sons. It is this sharing of trusteeship among the members of a line of men which equates the women. To an outsider, including the son of one of these women, others of the women are viewed as,col-lectively, potential x^ives to the men of his line which includes his father and brothers. The size of such bride/giving and bride-taking lines i s highly variable as was noted earlier. (b) sex-role continuity of MoBro/SiSo. This analysis parallels the one above except that the continuity would develop i n a situation of strong maternal trustee-ship. The identification of a man and his MoBro as sexually con-tinuous may occur where they are clearly the common wards of a solidary Mo/Da shared trusteeship. Again, however, i t i s more probable that this identification w i l l involve regarding a man and his SiSo, common trusts of a single female line, as equivalent as potential spouses. A g i r l may be proscribed to particular ones of these men by intra-familial incest taboos; e.g. the woman is - 123 -specifically taboo to her own father as the spouse of another wo-man of her line (her mother's), but has access potentially to other men belonging to the same line of women as does her father. This would include her FaSiSo. In general, then, I would argue that i t is only seldom the case that the sexual-continuity w i l l develop in full-blown form. A man, therefore, does not view his own mother and mother's brother's daughter as sexually equivalent as sister and mother are. Rather, Mo and MoBroDa are distinctive sexual principles who are, nevertheless, identified as equally available to ego's line as the wards of a bride-giving line. The Generalization of Shared Taboos. In i t s f i r s t formulation, I defined Father-Son sex-role continuity as the assumption, by a boy, of the sexual taboos pro-per to his own father. This was presented, for expediency's sake, in an unqualified manner. Yet, in the f i r s t section of this essay I have argued at some length that a science of human beings must accept as i t s most effective explanatory postulates the predication to human beings of behavioural dispositions. These must be under-stood only as tendencies which never operate in an unqualified manner. In keeping with this approach, I shall now expand some-what, and qualify somewhat, our notion of Fa/So continuity as i t obtains in systems of matrilateral cross-cousin marriage. I shall be concerned f i r s t with such systems in societies also characterized by predominant paternal trusteeship. - 124 -The qualified minimal definition of Fa/So sex-role conti-nuity i s that a boy assumes at least two of the sexual proscrip-tions proper to his own father. These are: (1) proscription to a l l women who are his father's own sex-ual wards. (2) proscription to the sexually continuous line of women In the following diagram, I have indicated a number of Mo/Da lines, each being the "mother's line " of a member of a line of men. Each of the Mo/Da lines is distinguished by an encircled numeral. The men of a receding Fa/So line are designated by Cap-i t a l A and each man distinguished by a subscript numeral. initiated in Father's mother. This line includes FaMo, FaSi, FaSiDa, FaSiDaDa, etc. EGO CD Figure 10 Continuous lines of women. - 125 -The "mathematical" implications of the notion of Fa/So sex-role continuity are that Ego-Aj- i s proscribed not only to the women of line(3i above, but since Ego's father - (A4) - i s proscribed to his father's (A3's) mother's line, i.e., line (<2j Ego should be proscribed to this line (line © ) as well since i t is one of the proscriptions proper to his father. Thus, by extend-ing the logic of Fa/So continuity, Ego's proscriptions would "pile up" through time, since the same can be argued of line(T]jetc. Thus, we would ask, i s Ego invariably proscribed to women who are his genealogical FaFaSiDaDa, or again, his FaFaFaSiDaDaDa? For purposes of definition, then, I shall qualify what I mean by"proper to"to indicate those women to whom a man i s sex-ually proscribed independently of the continuity process which calls for his sharing the taboos of others. Thus, again, two kinds of proscription are proper to every Ego: (1) sexual pros-cription to his own mother and her line, which includes one's sister. (In figure 10, A5 1s proscription to the women designated b y © ) . (2) Sexual proscription to one's own sexual ward, i.e., one's daughter — and her line (DaDa, etc.). (In figure 10, A5's proscription to women designated by (5), with the exception of his own wife). Thus, minimally, a man has two general sets of proscrip-tions: (1) Those proper to himself; mother's line and daughter's line. (2) Those proper to his father and adopted by himself; sister's line and father's mother's line (line (3)in figure 10, and again l i n e @ ) . - 126 -Collectively, these are lines(5) , © , and@of figure 10. When combined \7ith the phenomenon of classificatory termin-ology, any tendency to inherit proscriptions to further lines of women (such as(2)and(l)) could raise d i f f i c u l t i e s . There i s a tendency -- and I must emphasize the word tendency — to apply to the classificatory sisters of a proscribed woman, the same pro-scription. The sisters of such a woman may be the women of her descent group or of some section of her descent group. Since the women of a line are, where there is a rule of patrilineal descent, of different descent groups from those immediately above and be-neath them in generation, i t can be seen that Ego could come to confront so many proscriptions as to have no women available to him-self. Thus, I have suggested that Ego is proscribed not only to FaSiDa but to FaSiDaDa. The latter is of a different descent group from the former. Is Ego proscribed to a l l of the sisters (many of whom could easily be of Ego's own age) of FaSiDaDa? A problem of this kind would of course multiply greatly as Ego i n -cluded lines(2)and(I)as well. At this moment I shall not attempt to be more specific on this issue except to argue that there are several general ways that this problem may be managed. The f i r s t answer i s that very distant sisters of a directly proscribed woman may simply not be treated as her classificatory sister. The second i s that there may be a tendency to minimize the accumulation of lines. This w i l l be a function of the d i f f i c u l t y of maintaining a clear historic - 127 -record of lines which originated in the marriage of Ego's male ancestors in distant past generations. The third point offers an additional perspective on the more positive features of a prescrip-tive system. The matter i s far too complex for the present analy-sis, but a simple i l l u s t r a t i o n may serve to put the general idea across. The following diagram, figure 11, illustrates the idea that a tendency to limit the number of descent groups from which the persons of one sex of a given descent group may take a spouse w i l l have the effect of holding down the development of further descent group proscriptions. C A O A- 'O line 1. line 2. line 3. » r<5^ A- t5^A = O * ro. -A= Figure 11  Fig. 11. Ego i s darkened triangle. In figure 11, the broken lines encircle the women of a Mo/Da con-tinuum. Line 1. i s the line initiated by Ego's FaMo, line 2. i s initiated with his own mother, and line 3. with his daughters. Locate Ego's FaSi. There might be several l i t e r a l FaSi's, not to - 1 2 8 -mention other close classificatory ones — e.g., the actual sisters of father's f i r s t patrilateral parallel cousins. If each of these women i s allowed to marry arbi t r a r i l y into any descent group she chooses other, of course, than her own, then Ego so long as he i s under any pressure to recognize a l l these women as FaSis (which means to recognize their brothers as fathers) -- w i l l be obliged to recognize FaSiDas in as many descent groups as those into which the classificatory and several l i t e r a l FaSis have married. Given, further, that he generalizes the label FaSiDa to the descent group sisters of these g i r l s , i t could easily come to pass that Ego w i l l have no women whatever whom he i s allowed to marry without break-ing the proscription against the FaSiDa category. The same general d i f f i c u l t y w i l l arise i f I am correct in saying that, in these systems, Ego w i l l be proscribed to sis-ter's daughter. If Ego has actual and/or close classificatory sisters who are allowed to marry arbi t r a r i l y into any descent groups they choose, Ego may in theory come to be proscribed to marry further women who are descent group sisters of his sister's daughters. Where Ego, thEn, i s a young man with a substantial number of older father's sisters and sisters a l l of whom have married into different descent groups, Ego could be confronted with a situation where every woman he meets i s a taboo woman. Paradoxically, any tendency for a line of men to limit the descent groups into which they allow their daughters to marry w i l l hold down the range of descent groups from which they and - 129 -their sons are sexually and therefore maritally proscribed. Ac-cording to this analysis Ego i s not allowed to marry women of FaSiHu descent group (FaSiDa proscribed) or of sister's husbands descent group (note that the Purum exp l i c i t l y prohibit this; vide  supra, p.68 ) because each contains a g i r l of a specifically pro-scribed continuum. If the range of descent groups into which sister and father's sister (at least socially close and important ones) are allowed to marry is limited, then the men of the line w i l l not find themselves without spouses. I submit, further, that the factors which determine the descent groups into which Ego and his sisters may or may not marry w i l l be many. In general they w i l l relate a socially near group of men to other socially close units (local descent groups?). Factors such as spatial proximity, economic and p o l i t i c a l inter-action w i l l certainly loom large. We may anticipate then that corporate clusters w i l l usually favour marriage in one direction with other corporate clusters and w i l l define each such cluster as bride-giving or bride-taking vis-a-vis i t s e l f by defining the en-tire descent group to which i t belongs as one or the other. Other corporate clusters which belong to the same descent group as the first-mentioned one may have quite different definitions of these other descent groups. Societies like the Purum appear to have ex p l i c i t l y b u i l t - i n mechanisms which help to avoid such d i f f i c u l t y . Their "whole-descent group" -- SIB — regulations w i l l limit the l i k e l i -- 130 hood that future generations find themselves without potential sex-partners and wives. I do not argue that such regulations reflect a conscious awareness of the aforementioned structural d i f f i c u l t i e s on the part of the culture participants themselves. Indeed, I should be surprised i f such a d i f f i c u l t y did not periodically arise. Where i t does, I would anticipate that i t could become the source of a quasi-fission of the relationships in an otherwise close social unit. Thus, Ego might simply ignore the marriages of certain sis-ters and father's sisters, in order to avoid defining too many gi r l s as maritally and sexually proscribed. Such a situation could even lead to heated dispute and the beginning of group f i s -sion. Such a picture makes i t by no means unthinkable that the culture participants in such systems are fu l l y aware that these d i f f i c u l t i e s w i l l arise unless strictures are placed (in terms of descent groups) upon the marriages of their daughters and sisters. It w i l l be obvious that such strictures can have uniformity and succesful implementation only within a group having sufficient corporateness to achieve communication and agreement among i t s members. The scale of such a group i s clearly subject to a vast number of variables. In sum, awareness of such a mechanism i s by no means out of the question, and i f we have never been told of i t perhaps i t i s because we have never (to the best of my knowledge) asked. - 131 -Finally, I shall explicate a l i t t l e more clearly an issue which has already been raised. An exogamous descent group i s one in which the men are proscribed to the sexual wards of one another. This represents the generalization <fa disposition which appears in a l l individual Fa/So dyads. The further tendency of a boy to assume the proscription, proper to father, of the line of father's mother w i l l tend not to generalize as far; i.e., to fewer of his c l a s s i f i -catory fathers. I have already argued that this tendency w i l l vary according to the amount of social closeness between Ego and his various fathers. This, in turn, i s a variable subject to many conditions, p o l i t i c a l , economic, etc. — even personality factors. (See the citation from Warner given above, pp. 65-6). The following diagram attempts to summarize my argument. It shows the minimal sexual proscriptions of Ego in a situation of matrilateral cross-cousin marriage based upon our principles of continuity and trusteeship. A^O © A;6 A ? O A^O B (A) © b ® Bi a © A4(l) a ® A2(i; (3) OA a (£) A3(2) A=0 ATO A=<5A 5~E6 ArO 07\ 6"A c © C5(2) Figure 12 - 132 Legend: (1) Ego i s the darkened triangle. (2) Capital letters indicate a line of Fa/So continuity. (3) Small letters designate that a woman i s the sexual ward of some man (or men) bearing the same letter capitalized. Such women are taboo sexually to a l l men bearing the same capital letter either as the particu-lar ward of the man or as a function of exogamy. (4) Encircled numbers indicate a line of Mo/Da continuity. (5) Non-encircled numbers designate men pro-scribed to a line of women who bear the same number encircled. This applies to Ego's proscription to his own mother and her line; i.e., based upon the given Mo/So incest taboo and principle of Mo/Da conti-nuity. (6) Bracketed numbers indicate proscription to the line of father's mother. An important consequence of the postulated limitation of any tendency to generalize Fa/So continuity throughout the member-ship of a named exogamous descent group, such as the Purum sib or even lineage, i s that we may expect to find classificatory sister-exchange, i.e., two-way movement of women between "whole descent groups", a not uncommon occurrence. Furthermore, such marriages need by no means be regarded as breaches of either the kin-term regulations or of the descent group regulationsproper to the people involved. It i s equally possible that some persons >f a descent group may regard the marriage of others as wrong because the l a t -ter have ignored the regulations proper to the former, and have, as a result, ignored them. The occurrence of such a reaction and measures which might be taken as a result must be regarded as an open question. The following diagram, Figure 13, illustrates this point: - 133 -Al o A H »--o EGO, = 6 (°) A A l l EGO. Figure 13 Legend: (1) The capital A represents people of a single whole descent group within this descent group two lines of men are shown and distinguished by one and two attached strokes respectively. (2) The 'X' and c i r c l e , '0', each designate membership in different descent groups. (3) The broken horizontal line indicates dis-tant classificatory brothers or sisters. In this i l l u s t r a t i o n , Ego 1 is proscribed to FaMo line of women which includes g i r l #1 and, for Ego, her classificatory sisters ( g i r l s of her descent group). The diagram describes a situation in which Ego 1 marries a g i r l of the very descent group to which the proscribed actual FaSiDa of Ego 2 belongs. Ego 1, ex hypothesi, does not recognize the actual father of Ego 2 as a significant father to himself and therefore does not assume the proscription to that person's mother's line (which includes g i r l #2 and her descent group sisters). The upshot of this i s two-way bride ex-change between "whole descent groups". Ego 1 has married a woman - 134 -of the same descent group as the one to which Ego 2s FaSiHu belongs. Yet, as far as Ego 1 is concerned, he has not married incorrectly. In general, then, beyond women traced through close kins-men the kin status of a given woman i s an ambiguous issue. I have already pointed out that purely on the logic of classificatory re-lationships i t i s impossible to determine the terminological status of f u l l y half of the sixteen second cousins (see Harrison Whyte, pp. 42-47). I shall pause briefly to il l u s t r a t e this point. On page 110 of "Structure and Sentiment", Needham remarks that "in a matrilateral system . . . the MBD is herself equivalent to MMBDD." I cannot take this to mean less than that a genealogical McMoBroDaDa is treated as of the same terminological category as a genealogical MoBroDa. How can we be satisfied of this? It is cer-tainly true in any circumstance in which the matrilateral c.c. mar-riage rule i s modelled as involving invariable marriage with the l i t e r a l MoBroDa (see f i g . 1, p. 10). However, as I have argued throughout this essay, such a model f a i l s abysmally to represent what the marriage rule means to the culture participants, not to mention what they actually do. What, then, can the arithmetic of classificatory terminology determine concerning a genealogical McMoBroDaDa? A-T6 cb AO A § ~k A 0 6 0 6 6 © EGO 1. 2. Figure 14 - 135 -Classificatory logic leads to certain conclusions con-cerning coincidence of genealogical relationships and terminologi-cal ones. We expect, for example, that, in figure 14, Ego w i l l apply the term for his own MoBroDa to his genealogical MoMoSiSoDa ( g i r l #1). To Ego's mother the father of #1 i s a matrilateral par-a l l e l cousin. Thus, Ego's mother w i l l refer to him as brother. To Ego he i s a mother's brother and, #1, therefore, i s a MoBroDa. But can we construct an equivalent argument for the MoMo-BroDaDa (#2)? The mother of #2, is a MoBroDa to Ego's mother. What does Ego's mother term the daughter of MoBroDa? What does Ego term the daughter of his mother's MoBroDa? I can discover nothing in the logic of classificatory terminology to answer the question. The only remaining phenomenon which can assist Ego in his decision i s the descent group membership of #2. But this, I have already argued, can be variable from one Ego to another with-in a whole descent group. In general, the variation of descent group regulation diminishes with an increase in the social closeness (which, incidentally, w i l l generally parallel genealogical close-ness) of a subset of men who are members of a larger set of men --the whole descent group. Matrilateral Cross-Cousin Marriage With Maternal Trusteeship. A matrilateral c.c. marriage rule can also be envisioned in a situation of predominant maternal trusteeship. This i s in keep-ing with the fact that the rule sometimes occurs in a society with - 136 -matrilineal descent; we expect the trusteeship to correlate with the direction of descent much as Homans and Schneider have argued that the locus of general jural authority correlates with descent rule. The following diagram illustrates the system from the view-point of female Ego: A=0 AO ® B l O O A2()() A = 6 A = o A = < 5 ~ A = 6 A = o B()K) b ® Dioo b Q O A 0 A A= B()51 b(|) D()X) d(l) a © A12() ( ) ® A32() a © C2()() A412 afi) A()32 a(5) C ( K ) c(|) A l) 2) a ® Figure 15  Female Ego is the darkened c i r c l e . Capital letters attach to women only and identify lines of Mo/Da sex-role continuity. Thus, in any given Mo/Da dyad of such a line, the daughter assumes the sexual proscriptions proper to her own mother. The proscriptions proper to a person include the particular sexual ward (son) of each, and the line of their own father. Thus female Ego i s proscribed to her own son's line, her father's line, and the line of her mother's father. The lines to which a given female Ego is proscribed are indica-ted by the - 137 -3) Non-Encircled numerals which attach to the women only. These numerals appear in the order of son's line, father's line and mother's father's line. The lines of men to which these num-erals refer are indicated by the corresponding : 4) Encircled numbers which attach to each man. That i s , they in-dicate Fa/So sex-role continuity. Where a woman i s proscribed to one member of a line she is proscribed to a l l . This i s the case for a l l women except a wife who i s proscribed to a son and his line exclusive of her own husband. Thus, the f i r s t in the sequence of non-encircled numbers attached to a woman indicates proscription to her son and his line, but while i t corresponds to the encircled number of her own husband, i t does not mean that she is proscribed to him as well — obviously. Thus female Ego is proscribed to lines(5)(son's),(l) (father 1s) and (2) (mother's father's) and has the sequence of numbers 412 attached to the right of her capital letter, A. It w i l l be noted that female Ego's FaBroSo belongs to line(1)while her MoBroSo belongs to line(2). Both are, therefore, sexually taboo to her. The latter one i s , of course, proscription of the FaSiDa frmm male Ego's viewpoint. 5) Small letters indicate men who are the sexual wards (sons) of the individual women of a Mo/Da line who bear the corres-ponding capital letter. A woman of a given line must abstain from sexual relations with any man who i s the sexual ward of another woman of her line. Thus, sexual relations are taboo - 138 -between women designated by a given capital letter and men bearing the corresponding small letter. This, i t w i l l be observed, includes the MoSiSo of female Ego. 6) Blank Brackets indicate letters or numerals not identified in this diagram. It w i l l be seen that female Ego's FaSiSo i s of his father's line, which is not one of the lines to which female Ego is proscribed; i.e., not her father's or mother's father's line. - 139 -CHAPTER 8  Murngin Kinship: An Illustration I stated, much earlier, in Section I, that a verification procedure involves the making of predictions, from the Explanatory postulates, which concern phenomena specifically not ones which the postulates are designed to explain. It seems to me that the model given above, of matrilateral c.c. marriage, makes a number of pre-dictions quite distinctive from i t s explanation of the f i r s t cousin regulations themselves. To i l l u s t r a t e this procedure, I have selected Warner's study of the Murngin, a tribe of Australian aboriginals, since i t is relatively complete in i t s description of sexual taboos. My major concern w i l l be to show that the Murngin social system reveals a strong tendency to proscribe, sexually and maritally, women whose proscription i s predicted by my model. These w i l l include women other than Ego's own FaSiDa, FaBroDa, MoSiDa, and the classificatory — descent group — sisters of each of these women. Needham describes the Murngin as, Not prescriptive in the sense required, but a very d i f -f i c u l t case. There i s indeed a categorical distinction between MBD and FZD, and the latter may not be married. But they also have an eight section system (though in certain respects, as Warner shows, i t resembles a four-section system), and marriage i s thus necessarily with a genealogically bilateral cross-cousin. 79 I am puzzled as to the meaning of Hgenealogical bilateral cross-cousin". The "categorical distinction between MBD and FZD" and the proscrip-- 140 -tion against marriage with the latter more than satisfies what my own universe of discourse w i l l include as a case of matrilateral cross-cousin marriage. I require, s t r i c t l y speaking, only one further stipulation: the FaSiDa marriage proscription must be paralleled by a sexual taboo. Warner does not specify such a taboo. However, he pro-vides some discussion of "extra-legal sexual relations." He makes the following observations: A man ordinarily would choose_ a galle /MoBroDja/ for such aff a i r s , and a woman a due /FaSiSo/; but they do occur between momo and kaminyer, l i t t l e arndi and waku, gawel and waku, and one case has been reported of momelker and dumungur. This last i s very rare and has a strong taboo against i t . A l l relationships except the above are com-pletely taboo, and no one would any more think of such an extra-legal union than a normal man in our society would consider a liason with his own mother.80 The taboo against " a l l relationships except the above" w i l l clearly include taboo against sexual relations of a man and his female DUE ("DUE' MIELK"),FaSiDa — as of a woman and her male GALLE ("GALLE DIRAMO") — MoBroSo. Having established the existence of a sexual taboo to FaSiDa (as well as a marital proscription) and the non-proscription of MoBroDa, we may proceed to examine the remaining sexual regula-tions. Figure 16 w i l l assist the reader in the following discussion. It w i l l be noted that most of the sexual taboos to which Murngin men are subject involve women of their own exogamous moiety. The Murngin moieties are named DUA and YIRITJA. On page 113 of his - 141 -book (1964) Warner provides a table showing the distribution of kin-terms according to moiety and subsection membership where a man i s in a given moiety. The relationships listed in the foregoing quotation a l l obtain between men and women of opposite moiety from each other. "Extra-legal" sexual relations take place only between persons of opposite moiety. Thus, i t might be argued, sexual taboos against " a l l relationships except the above" are simply a function of belonging to Ego's exogamous moiety. However this might be, i t i s interesting that Ego i s not allowed to have "extra-legal" sexual relations with any women of the opposite moiety. Nor, on the other hand, i s he limited, in these, to a GALLE MIELK (MoBroDa). The most conspicuous case of an opposite-moiety yet sexually taboo woman i s of course DUE MIELK herself ( i . e . , FaSiDa). But there are several other rela-tionships of interest to us. In the f i r s t place, i t is clear from Warner's statement that even though there has been a liason between a man and his MOM-ELKER — clearly a classificatory one -- the liason was regarded as wrong; normatively taboo. I t w i l l be seen in figure 16, that MOMELKER has the genealogical statuses of MoMoMoBroDa and McMoMo-BroSoSo. By elimination, we can also locate others besides MOMELKER and DUE. Thus, for example, Warner does not include on his l i s t of "extra-legal" sexual relations the liason of a man with his female - 142 -DUMUNGUR. This would be listed as a NAIJIWALKER-DUMUNGUR liason. Nor does he l i s t any liasons of NATI (MoFa) and KAMINYER (DaDa). But immediately following the above quotation Warner adds a qualifying remark: much indignation usually i s caused by a gawel-waku l i a -son /MoBro/SiDa/ i f discovered, though several were ob-served. This sentiment would not be that of the injured wife or husband alone, but of the tribe generally. Men having such a relationship with waku or momelker would be called dogs and considered evil-doers even by their own people, who might be defending them at the time from the attacks of the g i r l ' s husband and relatives. The woman would almost certainly receive a severe beating and occasionally be killed.®1 One f i n a l qualifying remark of importance: A l l the above affairs would be with distant relatives and not the actual ones.82 In sum, then, we may include the following categories (terminolo-gical) of women of the opposite moiety from Ego as sexually taboo to him at least in terms of Murngin sentiment; DUE, WAKU, KAMINYER, DUMUNGUR, MOMELKER, and one more not yet mentioned, ARNDI (Mo, Mo-MoMo). These categories are located in figure 16, where the at-tached subscript #2 indicates that they are women who are sexually taboo to Ego. The subscript #1 attaches to these women of opposite moiety whose affairs with Ego are not pictured specifically as i n -cestuously taboo even where they are "extra-legal". No numerals are attached to the women of Ego's own moiety since they are a l l sexually proscribed to him. People of Ego's own moiety are indi-cated as the darkened figures while those of the opposite moiety are not darkened. 1. Momo Nati (B) (B) GALLE (B) (B) A L i t t l e Gawel Arndi (C) (C) 2(Q (G Arndi Gawel MARI ( A ) (A) Momelker (B) Natjiwalker (B) u 3 M •i-l - 144 -Figure 16 also shows the distribution of personnel ac-cording to section membership (capital letters). The Murngin have eight named subsections. A child belongs to the specific subsec-tion of i t s father or of i t s mother. Yet i t s subsection membership is determined by i t s mother. A Murngin woman is prescribed in marriage to either of two subsections (as i s each man). Thus, for example, a woman of subsection Aj^ must marry a man of either of subsections B-^  or ^  • Whichever the subsection of her husband, however, her children w i l l be of subsection C . The following dia-gram (from Warner, 1964, p. 109) shows the marriage and descent relations among the eight subsections. Figure 17 Murngin Subsections This diagram shows that the subsections are arranged into pairs designated by the same capital letter, but distinguished by numerical subscript. However, the subsections as such do not reg-ulate Murngin marriages. Thus, a man of either of the A subsec-tions may marry a woman of either of the B subsections, and their children w i l l be of one of the D subsections (depending upon the B subsection of their mother). - 145 -Thus, for purposes of marriage regulation, Murngin soci-ety is dichotomized only twice, i.e., into sections. The f i r s t dichotomy is produced by the patrimoieties. Each moiety contains four subsections. These four are gathered into two subsection pairs in the sense that the marriage regulations of a l l members of a subsection pair are identical. Each member of the subsection pair Aj^-A2 may marry into either of the subsections B^-B , w^^-c^ together form a subsection pair — section — of the opposite moiety. The children of such a marriage w i l l be of the section D^ -D2 which i s the other section of their father's moiety. Thus, Warner's diagram (Fig. 17) may be simplified to represent marriage regulation as follows (Fig. 18): Dl: Moiety 1. Moiety 2. ' Figure 18 Murngin Sections; these are represented by the capital letters. In passing, we might observe that diagrams 17 and 18 may contain the source of Needham's decision that the Murngin marry a "genealogical bilateral cross-cousin". It w i l l be noticed in Fig. 18 that, as between sections, women are exchanged in two directions (e.g. A=B). This point i s illustrated in the following diagram (figure 19): - 146 -Figure 19 In this diagram each moiety is represented as a c i r c l e . The d i v i -sion of each moiety into two sections i s represented by the half-circles above and beneath the horizontal diameter. Each moiety i s further divided into two clans by the vertical diameter. The clans are indicated by the encircled numbers (l)- (5) . Thus, clan (D i s the l e f t hand semicircle of moiety 1. and is divided into B and C sections by the horizontal radius. In a l l , there are four clans,which i s the minimal required for such a structure since there must be an even number of clans, more than two to avoid d i -rect exchange between them. Solid arrows represent the direction of bride movement between clan-section and clan-section. - 147 -The diagram shows several noteworthy consequences: (1) two-way movement of women between moieties -- obviously. (2) two-way movement of women between sections of moieties. (3) one-way movement of women among clans forming a completed c i r c l e ; ® »(3) ± © _ » ® ^ (£> . Thus, one-way movement of women between "whole-descent groups" — on the clan level i s structurally possible in Murngin society. But this does not allay the fact that marriage could, according to the logic of classificatDry terminology, s t i l l be with the classificatory (note; NOT genealogical a la Needham) bilateral cross-cousin. Figure 20 shows this consequence: Clan 4 6 Clan 2 O A A T Figure 20 Clan 3 Here, Ego is the darkened triangle. Solid lines repre-sent direct consanguineal relationships, and broken lines represent classificatory, or better, potential classificatory relationships. Ego i s shown to marry, in this hypothetical situation, a woman of - 148 -the same clan as his own mother; i.e., a MoBroDa. However, she is the actual daughter of a woman of clan 3 which belongs to the same moiety, and perhaps the same phratry (which "group together cer-83 tain clans within one moiety" ). The mother of this woman could be regarded as a classificatory FaSi and thus Ego's wife as a classificatory FaSiDa as well as a classificatory MoBroDa. But the Murngin do not marry anyone they term FaSiDa. Thus, the possible patrilateral relationship of Ego and his wife (and surely many potential wives) i s simply disregarded. Having said this, I must add that while a model of mat-r i l a t e r a l connubium is not inconsistent with what we know of Murn-gin social structure,there i s on the other hand, no reason to be-lieve that marriages actually or even ideologically operate in this way. At the same time, there is no question that the Murngin prescribe marriage to the category MoBroDa, and proscribe the category FaSiDa. Further, the two w i l l not be allowed, in their view, to overlap in a single person. Let us now pick up once more Ego's sexual regulations vis-a-vis the various female personnel who surround him. Figure 16 indicates the section membership of these personnel by the bracketed capital letters. In this diagram, Ego i s a member of section A and i s therefore prescribed to marriage with women of section B. The distribution of the numbers 1 and 2 among the women of opposite moiety from Ego indicate that some of them are accessible for "extra-- 149 -legal" sexual relations while others are taboo. Some of those who are taboo are of the very section to which Ego i s maritally pre-scribed. These include DUE', KAMINYER, DUMUNGUR, MOMELKER. At the same time he has sexual and marital access to GALLE, also of section B. Conversely, he i s allowed sexual access to some (LITTLE ARNDI) but not others (WAKU) of section C, to which he is maritally proscribed. In general, these regulations force the conclusion that, as such, sectional membership does not regulate sexual relations even where i t does regulate marriage. I do think, however, that my earlier model of matrilateral cross-cousin marriage with pater-nal trusteeship does help considerably to shed light upon these features of Murngin society. A basic feature of the model was that women are viewed as collected into Mo/Da lines, and Ego tends to be proscribed to a l l of the women of a line where he i s proscribed to one of them; he treats them as sexually continuous. Thus, for example, Ego's Mo, MaMo, S i , SiDa and SiDaDa compose such a line. In diagram 16 , these are the women termed respectively, ARNDI, MARI, YEPPA, WAKU, and KUTARA who compose just such a line, and are a l l sexually taboo to Ego. The taboos to ARNDI (mother) and YEPPA (sister) are giv-en as intra-familial incest taboos, but what of the remainder? MARI (MoMo) and KUTARA (SiDaDa) are of Ego's own moiety. This leaves WAKU who is not of Ego's moiety. The fact that she i s of the section - C - to which Ego is not maritally prescribed seems - 150 -to have no direct bearing on sexual access since Ego does have "extra-legal" affairs with LITTLE ARNDI who i s also of C section. Yet, where liaisons with WAKU are concerned the Murngin, as des-cribed by Warner, view i t as "a normal man in our society would consider a liaison with his own mother." In addition, our model called for the adoption, by Ego, of his father's proscription to the line of women initiated in the mother of the latter. This line is composed of MOMO (FaMo), MOKUL BAPA (FaSi), DUE (FaSiDa), GURRONG (FaSiDaDa), and DUMUNGUR (FaSi-DaDaDa). Here, MOKUL BAPA and GURRONG are of Ego's moiety, but DUE and DUMUNGUR are of opposite moiety and the prescribed sec-tion -- B. Yet he i s sexually, and maritally, taboo to both. In MOMO we encounter what appears to be our single ex-ception to the notion of a continuous Mo/Da line. Warner observes that MOMO: is frequently a sweetheart of his. This i s of course not his father's mother, or any of her sisters of clan members, but a woman who stands in this t r i b a l relation to him. 8 4 The f i r s t point to note, an obvious one, is that the "sweetheart" relationship involves a very distant MOMO. Our principles, through-out, have stressed that the continuity concepts develop, for Ego, around close relationships from whence they may radiate outward to other persons classed, on the basis of descent group and sectional membership with the persons closely related. The radiation effect can be expected to weaken with social and genealogical distance. - 151 -The exception of MOMO raises one further interesting possibility. It w i l l be noticed that the vertical line of female personnel to whom men of Ego's male line are married present a striking contrast to a l l other vertical lines. A l l except ARNDI are sexually accessible to Ego. Warner observes that "a waku often has an affair with a distant arndi. This is considered 85 wrong, but i t occurs". I t i s d i f f i c u l t to estimate the strength of such a taboo from such a remark. If my following speculation has any credence, I should not be surprised to discover considerable ambivalence concerning affairs with a distant ARNDI. Much earlier I introduced the thought that a possible aspect of Fa/So sex-role continuity might involve the sharing of sexual privileges as well as proscriptions. The d i f f i c u l t y in describing such a phenomenon derives from the fact that the im-mediate and direct sexual privileges of a boy's father are s t r i c t -ly proscribed to Ego. Thus, Ego's mother and, as a rule, any of father's wives w i l l be absolutely taboo to Ego. This, of course, i s in keeping with our earlier argument that women may operate as a wedge in the relations between men which demand trust and loyalty. Thus, among the Murngin, Warner observes: Except under rare circumstances there would be no pos-sible chance of a son's marrying a possible mate for father, just as the opposite would not be allowed.86 Any woman whom Ego's father would regard as a "possible mate" to himself would be termed GALLE by the father and ARNDI by Ego. Thus, while Ego does not marry an ARNDI, he does periodically have an - 152 -affair with a distant one. While Warner does t e l l us that this i s "considered wrong" there i s a lack of emphasis in the statement that allows me to wonder just how serious such a transgression i s taken to be. Again, I would be inclined to predict an ambivalent reaction from Ego's father at worst. The cases of Ego's sexual privileges with MOMO and LITTLE ARNDI are clearer. Again, these invoke distant relatives. My present hypothesis, then, i s that Ego is virtually expressing the continuity of his male line in his affairs with women of the same terminological category as the women with whom FaFa, Fa, him-self, and his own son have the greatest sexual privileges. Or, to put i t another way, the continuity i s expressed in the low degree of negative sanction to such liaisons. A category of sexually accessible women i s determined (or defined) somewhat differently from a category of sexually taboo women. Proscription tends to spread in two ways: (1) Sexual proscription tends to spread, mainly downward but sometimes generationally upward along a Mo/Da continuum from i t s point of origin in a specifically proscribed woman (e.g. own mother, own sister, SiDa, SiDaDa, etc; FaBroWi (point of origin), FaBroDa, FaBroDaDa, etc.). As this w i l l introduce multiple generations, i t may, as i t does with the Murngin, introduce people (women) of several terminological categories. (2) Sexual proscriptions tend to radiate outward from each of the women in a particular proscribed Mo/Da line. Thus, the - 153 -classificatory sisters of each of these women may be incorporated into the proscription. This may involve a l l of the descent group sisters of the separate women in such a line. In this kind of proscription radiation, each of the women in the line becomes a sep-arate point of origin for the proscription of her sisters. Where this involves a variety of terms, a variety of terminological cat-egories are introduced. Radiation of this kind tends to weaken as social (and genealogical?) distance increases away from i t s point of origin. Sexual privileges (as opposed to marital prescriptions) tend to be located in the very terminological categories to which belong the women in whom proscribed lines originate ("originate" in the sense of proposition 1, above). Thus, they w i l l be, from Ego's view, the very distant descent group sisters -- probably not even of the same clan but of a distantly related clan — of the women with whom the men of Ego's male line (including FaFa, Fa,Ego himself, and his son) have sexual and marital privileges. These women also span a series of generations and as such may involve a number of term categories. The relations among distance, strength of dis-position and point of origin are in this case the reverse of pro-position 2. That i s , sexual privileges weaken as the personnel of the category approach the point of origin. Among the Murngin, at least, sexual privileges vis-a-vis such a terminological category dis-appear entirely at the boundary of the actual clan of the point of origin of the category. Thus, for example, Ego could never have a sexual relationship with a woman termed MOMO who i s also of his own FaMo clan. - 154 -To return to Murngin regulations as such I might add two points for completeness. KAMINYER (figure 16) is a sexually taboo category though she is of the prescribed section. She i s of Ego's daughter's (GATU MIELK) line. The case of MOMELKER, Ego's MoMcMoBroDa (figure 1<3) i s also very interesting. Warner's remarks seem to contain the same type of viewpoint as the postulate of Mo/Da continuity: The momelker relationships i s built out of the mokul _ rumera /mother-in-law/ behaviour. . . . She /momelker/ is the mother of mokul rumera, the matives say; and just as one avoids a mother of one's wife, so one avoids her mother.87 Thus, the same general avoidance relationship which applies to one's mother-in-law i s applied to MOMELKER because "She i s the mother of mokul rumera". This amounts to an explicit statement of an attitude of Mo/Da continuity except, as I had anticipated ear-l i e r , the identification has other forms than merely sexual ones. While we are on the subject of more generalized role-identification i t is interesting, though hardly surprising, that mother's sisters and mother's mother are regarded as the most ap-propriate maternal surrogates. When a mother goes away with the father for a trip to distant kins people, the older children are l e f t with the mother's sisters, the mother's mother, or the mother's brother.88 It i s important to appreciate that this phenomenon cannot be ac-counted for simply in terms of the interest of a clan i n these children since MARI MIELK (MoMo) i s neither of Ego's own or of - 155 -his mother's clan. MARI MIELK is very akin to Ego's own mother in many respects. She i s quite singled out in her relationship to Ego by the following statement: Mari mielk also occupies a position of considerable importance. She looks out for kutara /daughter's child/ when he i s small. Instances of a man's having a greater affection for his mari mielk than for his own arndi were recorded several times. Mari mielk corrects the children when they are mischievous or_ bad mannered^ Chastisement i s hardly ever practiced, /nor does Arndi/ but ridicule i s effectively used. 8^ Several further observations should also convince the reader that Fa/So role-identification in a more general sense is of great importance in Murngin l i f e . Consider, for example, the following: A man prefers a son, however, because of his value in the constant tr i b a l feuds, not only through his own assistance and the aid he brings to his relations (as, of course, the daughter does through her marriage), but also because almost the entire ceremonial l i f e of father's clan i s centered around males. A son inherits the right to perform certain dances through his father and never through his mother. By a son's i n i t i a t i o n into the various ceremonies the father's social p a r t i c i -pation i s further increased as i t would not be i f the child were a daughter.^0 Father i s of primary importance in other ceremonial events in his son's l i f e . When a boy is circumcised (at from six to nine years of age), i t i s the father who decides the type of i n i t i a -tion ceremony. The other older men of the tribe confer with him as to time, place, and other arrangements.^ Also, The father i s one of those who teach his son how to hunt, fight, and conduct himself in the best possible way in the practical affairs of l i f e . If the father i s a cere-- 156 -monial leader he instructs his son in the routine of songs, dances, and words that make up the great cere-monies. 92 One f i n a l point before we leave our analysis of Murngin kinship. I have expressed skepticism throughout that ideological prescrip-tions of one-way bride exchange among whole descent groups are the best way to understand the rule of matrilateral cross-cousin mar-riage. I argued this point at some length in connexion with Purum social structure. My skepticism i s more f u l l y supported in the case of the Murngin, Warner states: Theoretically an individual can marry into any one of the clans in the opposite moiety.93 Thus, i t would appear that even where a model calling for whole-descent group — clan — prescriptions is consistent with the other features of Murngin social structure (vide supra, pp. 145-147), such i s simply NOT THE CASE. The remainder of this essay w i l l offer a series of further models based upon our principles of trusteeship and contin-uity. No effort w i l l be made, however, to illustrate these as I have just done for matrilateral cross-cousin marriage with paternal trusteeship. - 157 -CHAPTER 9  Patrilateral Cross-Cousin Marriage Analysis in terms of the general postulates would predict a l l around attenuation of the continuity principles as they relate specifically to the sharing by Fa and So and by Mo and Da of sex-ual taboos. It would predict that these situations are dominated by an over-riding principle of maternal trusteeship. Thus, a woman is the major sexual trustee of her son. Her own daughter may share in such a role to some extent, but, in view of the tenden-cy for boys to be generally trained under the aegis of other men, much of the authority over a young man, including a role in his sexual disposition, i s delegated from the mother to the mother's brother. Metaphorically speaking, a woman and her own daughter may be viewed as a single kind of "womb", alike in the sense that they produce like offspring. A boy and his MoBro derive, as i t were, from the same "womb". Thus, a boy is sexually identified most with his mother's brother and i s , like him, a sexual ward of the same Mo/Da dyad. Thus, male Ego shares the sexual proscriptions of his MoBro. There i s , at the same time, an attenuated recognition of Fa/So identification. They are alike as sexual objects but less so as sexual actors. Thus, ego i s proscribed in the eyes of any woman to whom father i s proscribed, but i s equated with MoBro as a sexual actor as well as object. These principles of Fa/So vs. MoBro/SiSo continuity are, i n a l l probability, somewhat psycho-logically competitive. - 158 -From female Ego's viewpoint, sexual prohibitions proper to her include those to her own son and his line (identification as objects) and to her father and his line. Thus, a woman's taboo to her own brother has a double source i n the point of view of this analysis; he tends to be identified as an object with his Father, and he i s the sexual ward of the g i r l ' s mother. The wo-men of a line abstain from sexual relations with their mutual wards (sons). In spite of the tendency for women of a line to be equa-ted as sexual objects, especially in the sense that they are alike as life-giving wombs, there is no tendency for a g i r l to share the proscriptions proper to her mother. Thus, in sum we have an attenuation of Fa/So sex-role continuity and of the corresponding Mo/Da principle. In both cases, the principles tend to be confined to equivalence as sexual objects without the sharing of taboos proper to one another. The Mo/Da identification as sexual objects assumes the particular aspect of equivalence of life-giving principles. This results in the sex-role continuity of a boy and his MoBro. A man is directly pro-scribed to the lines initiated in his own mother and his own daugh-ter. The following diagram demonstrates the consequences of these dispositions: - 159 -A=0 © AfO AfO A O A 6^ A=6 Byl. b(I) Dng b © A22 w © 2^-2 a2 A Cm8. b © Dbl. d © Abl. a © Aw4. a © Ca 2 c © EGO Figure 21 Female Ego is the darkened c i r c l e . Capital letters attach to women only. They indicate lines of Mo/Da sexual identification. These women are equivalent in the double sense that sexual taboo to one means the same vi s -a-vis the rest (from male Ego's viewpoint), and in the sense that they produce sexually equivalent — continuous -- male children who are their respective sexual wards. These men have an attached Small letter which corresponds to the capital letter of their mothers and women of their mothers' line. Such men are sex-ually taboo to a l l women bearing the same capitalized letter. A boy i s sexually continuous with, i.e., shares the sexual taboos proper to his MoBro (who of course has the same small letter). On these grounds, female Ego i s proscribed sexually to MoSiSo who bears a small letter a. In this diagram, small letters attach also to women. Here they indicate a female Ego's proscription to her father and to - 160 -a l l men who are taboo to her by virtue of being of the same womb as father. That i s , men who are the co-wards of a line of Mo/Da identification. These men, of course, bear the same small letter as father and as female Ego. This letter i s f i r s t in the sequence characterizing a man, but second in the sequence characterizing a woman. On these grounds fe-male Ego is prohibited from having sexual intercourse with her FaSiSo who i s of the same line of women as her own father. Female Ego and FaSiSo, therefore, both bear the small letter b. 4) Encircled Numerals indicate the very attentuated Fa/So con-tinuity whereby they are merely identified as sexual objects by females (except, of course, the wife of father; i.e., mother of Ego). 5) Non-encircled Numerals attaching to women indicates a woman's prohibition against a given Fa/So line whose members bear the same number encircled. This line i s initiated in her incest proscription of her own father. It w i l l be seen that female Ego i s taboo to FaBroSo on this account. 6) Female Ego, f i n a l l y , shares no numbers or letters with her Mo-BroSo. The subject of patrilateral c.c. marriage was raised much earlier in this essay (pp.34 et.seci) At that time, i t was observed that this marriage rule i s cross-culturally far less com-mon than the matrilateral form. One of the factors in this pheno-- '161 -menon could well be that situations of pronounced maternal trustee-ship are relatively rare, especially where i t is sufficiently pro-nounced so that, along with the sexual equivalence of Mo/Da as maternal principles, i t causes a sex-role continuity of Ego and MoBro to the point of i t s competing with a Fa/So sex role con-tinuity. Where such a situation does arise, I would anticipate that the rule of descent w i l l parallel the direction of trustee-ship, i.e., matrilineal descent. The cases collected by Homans and Schneider do, indeed, prove to be mostly ones of matrilineal descent. Furthermore, I suggest that the latent tendency to Fa/So sex-role continuity would at the f i r s t opportunity become vitalized to the point where such systems would change to ones in which a l l near cousins are proscribed (Fa/So sex-role continuity superimposed upon the above situation would cause proscription of FaSiDa as well). I am not afraid to admit that, in this argument, I am assuming that there is something more "natural" about Fa/So con-tinuity than MoBro/SiSo continuity. - 162 -CONCLUSIONS Only the major themes of the essay w i l l be reviewed here. Section I considers Needham's view that unilateral c.c. marriage could only be adequately analysed in terms of the relationships among explicit whole descent groups. My concern was to show that, to the contrary, i t w i l l not offend common sense to discover the causes of norm systems in dispositional orientations of the col-lection of human individuals who compose a society. The theory developed in Section II i s essentially of this order. Some of the main issues were: (1) Needham's use of the concept of marital prescription (as opposed to preference) denotes an ideological social system i n which kin-terms are distributed among the persons surrounding Ego in s t r i c t accordance with the explicit descent groupings to which these persons belong vis-a-vis Ego's own descent group. This i s a necessary structural condition of idealized one-way bride ex-change between descent groups. The resulting system of unilateral connubium i s then explained in terms of the laws of reciprocity, the relative statuses of the various groups being one of the key balancing elements in the exchange system. Needham's distinction implies that prescriptive and pre-ferential rules of marriage are, in some ultimate sense, quite different phenomena demanding different explanations. This of course begs the question. When and i f a theory i s brought forth - 163 -which, unlike Needham's, i s able to accommodate both types of rule they w i l l no longer be regarded as inherently different phenomena. Homans and Schneider attempted such a theory. It is one thing to c r i t i c i z e the specific contents of their argument and quite another to pre-emptively assert as Needham has done that a theory such as theirs i s in error even before i t i s begun. As far as I am concerned, other things being equal, the greater cross-cultural scope of such a theory i s sufficient to make i t a more valuable one. The model proposed in this essay is an attempt to ac-commodate both types of rule; i t s potential cross-cultural range i s therefore substantially wider than that of the exchange model. (2) The internal logic of the exchange model was examined. In chapter 3 I proposed that an explanatory model i s simply a sequence of propositions arranged deductively so as to terminate in a sum-mary statement of the problem norm. As Needham has himself writ-ten the exchange model does not appear to provide an explanation of the phenomenon of matrilateral c.c. marriage. His argument suggests rather that given a prescriptive rule of matrilateral c.c. marriage, a one-way bride movement between descent groups is logically implied. The converse of this argument does not f o l -low as directly (see Salisbury, 1956). Thus, since Needham does not wish to argue that the marriage rule per se. i s to be understood as a teleological mechanism for perpetuating a system of unilateral connubium, I have concluded - 164 -that his argument employs the marriage rule in the position of a premise rather than as i t s conclusion. If this is correct, Need-ham's i s not a theory of matrilateral c.c. marriage. However, I have pointed out that a somewhat different interpretation of the Needham model is possible. Given an expressed ideology which amounts to in perpetuum descent group unilateral con-nubium, a male Ego w i l l find only matrilateral cross-cousins in the proscribed descent groups and only patrilateral ones in the prohibited groups. But do we find expressed ideologies of unila-teral connubium? (3) Certain Purum norms suggest a system of inter-sib unilateral connubium. At the same time, specific data concerning the idealized basis of terminological identification of members of various des-cent groups (sibs) by the members of other ones appear to contra-dict the model of inter-sib unilateral connubium. I have suggested that these contradictions disappear i f we understand the sib-regulations to apply to the members of de facto socially close groupings -- smaller in scale certainly than the idealized group-ings, sibs or lineages. From this point of view each Purum Ego belongs to a reference set of socially important kin through whom links to the various sibs are traced and the membership of each terminologically identified. This reference group i s smaller (among the Purum) than are the groups (sibs) with which i t exchanges women. The exchange model, to the contrary, requires that the rele-vant groupings be structurally equivalent (in size range, composition and especially as ideological units). - 165 -Thus, the interpretation of Purum data offered in this essay precludes Needham's mechanical model. I am not satisfied that the data allow us to accept even an implicit ideobgy of uni-lateral connubium among normative whole descent groups. At the same time, a tendency toward de facto one-way bride movement among sub-lineage reference groups must exist. The location of such reference groups must proceed without the aid of the Purum ideal descent group categories. Here, Leach's concept of local descent groups may prove helpful. More generally, however, the nature of the important reference groups must be treated as a cross-cultural variable as well as subject to flux through time within a culture. This ap-proach injects a dis t i n c t l y s t a t i s t i c a l element into our model building. It argues that we have not f u l l y comprehended the nor-mative structure of a society until we understand how the norms may be accommodated to a wide range of de facto, and changeable, social conditions. Purely mechanical models on the other hand cannot accommodate themselves to conditions of social change since they are confined to handling explicit ideologies. Ideologies are by their nature ambiguous. As such they can endure effectively in the face of de facto social change. Thus, for example, in the face of significant alteration in the composition (and functions) of the corporate structural units of a society, a man may well employ new c r i t e r i a in the distribution of kin term categories. Yet the rules of marriage, such as a - 166 -matrilateral c.c. prescription, can endure, nor need they become a mere vestige. Precisely this consideration allows us now to avoid ex-plaining away,as deviant cases, the many actual Purum marriages which f a i l to conform to Needham's model. They are indeed devia-tions from Needham's ideal, but there i s no evidence that the Purum themselves feel the same way. ( 4 ) The matrilateral c.c. marriage rule i s therefore treated as analytically independent of the idealized whole descent group structuring of the societies involved. In seeking to explain the marriage rule our unit became human individuals embedded within a set of basic role-relations with other people of great direct significance in their lives. The primacy of Ego's relations with his actual father, mother, siblings, and (to a lesser extent) his grandparents i s purported to be universal to the societies being studied. Given this assumption, my purpose has been to delineate certain cross-cultural regularities characteristic of this set of role relations. The explicit unilateral c.c. marriage rules are then deduced from the regularities, which I termed principles. These are concerned with a process of role-continuity obtaining between specific personnel of adjacent generations. The major focus of this process concerns the expression of personal social identity through the particular allocation of one's sexual p r i v i -leges and prohibitions. - 167 -(5) To be sure, this approach may invite an accusation of psycho-logical reductionism. My analysis i s not at a l l concerned with the dimensions of personality theory but the accusation would be correct in the sense that I am basically interested in the formu-lation of theoretical propositions which employ human beings as their subject element and which predicate to them generalized behavioural dispositions. I regard Marcel Mauss's concept of reciprocity as just such a theory. In other words, the exchange theories, being profoundly dependent upon the concept of reciprocity, are as much psychological as my own. It was observed that a dispositional analysis i s not tantamount to role analysis since the latter i s composed by and large of explicit cultural norms. Explicit norms are the anthro-pologist's problems, not his explanations. Explanation involves the activity, on the anthropologist's part, of creative inven-tion. He invents statements which his informants themselves do not make. In this sense, most explanatory models are unconscious models. (6) These views have certain methodological implications. Parti-cular explanatory models are a sequence of propositions which move deductively from their theoretical dispositional statements to the problem norm being considered. Yet, in my view, the primary ob-ject of the social sciences i s to discover and lend support to the theoretical statements themselves. The verification process has two logical phases. - 168 -F i r s t l y , the problem norm occurs in particular spatially and temporaly localized situations. A theory of such a norm must be believable in the context in which the norm appears. Thus, we attempt to show that features of the expressed culture other than the problem norm of our i n i t i a l model can likewise be deduced from our theoretical statements. This verification process i s i t s e l f sufficient to generate the ho l i s t i c orientation which has dominated Anthropology. I t also j u s t i f i e s , for me, the necessity of impres-sionistic ethnography as a stage prior to the development of a sharply defined problem approach. Secondly, the cross-cultural duplication of such intra-cultural verification strengthens our confidence that our theory has credence in each particular instance of i t s use. Finally, i t encourages us to seek out new kinds of situations in which such dispositions may be manifest. - 169 -REFERENCES CITED Ackerman, Charles 1964 "Structure and Statistics: The Purum Case." American  Anthropologist, Vol. 66, 53-66. 1965 "Structure and Process: The Purum Case." American  Anthropologist, Vol. 67, 83-90. Das, Tarak Chandra 1945 The Pururns: An Old Kuki Tribe of Manipur. Calcutta. Elkin, A.P. 1931 "Social Organization in the Kimberley Division, North-western Australia." Oceania, Vol. 2, 296-333. Fei, Hsiao-Tung 1939 Peasant Li f e in China: A Field Study of Country Li f e  in the Yangtze Valley. London. Homans, George C. and Schneider, David M. 1955 Marriage, Authority, and Final Causes: A Study of Uni-lateral Cross-Cousin Marriage. Glencoe, 111. Lane, Barbara S. and Lane, Robert B. 1962 "Implicit Double Descent in Southeastern Australia and the Northeastern New Hebrides," Ethnology, Vol. 1, 46-52. Leach, Edmund R. 1961 Rethinking Anthropology. London. Levi-Strauss, Claude 1962 "Social Structure", in Anthropology Today (Ed. Tax, Sol). Chicago. Lounsbury, F. 1962 Review of Structure and Sentiment: A Test Case in  Social Anthropology (by Rodney Needham). American  Anthropologist, Vol. 64, 1302-1310. - 170 -Mauss, Marcel 1954 The Gi f t . Glencoe, 111. Murphy, Robert F. and Kasdan, Leonard 1959 "The Structure of Parallel Cousin Marriage". American  Anthropologist, Vol. 61, 17-29. Needham, Rodney 1958 "A Structural Analysis of Purum Society". American  Anthropologist, Vol. 60, 75-101. I960 "Structure and Change in Asymmetric Alliance: Comments on Livingstone's Further Analysis of Purum Society". American Anthropologist, Vol. 62, 499-503. 1962 Structure and Sentiment: A Test Case in Social Anth- ropology. Chicago. Nutini, Hugo G. 1965 "Some Considerations on the Nature of Social Structure and Model Building: A Critique of Claude Levi-Strauss and Edmund Leach." American Anthropologist, Vol. 67, 707-731. Radcliffe-Brown, A.R. 1950 Introduction to African Systems of Kinship and Marriage (Ed. Radcliffe-Brown, A.R.). London. Salisbury, R.F. 1956 "Asymmetrical Marriage Systems." American Anthropologist, Vol. 58, 639-655. Warner, W. Lloyd 1964 A Black C i v i l i z a t i o n . New York. - 171 -SOME ADDITIONAL REFERENCES Cowgill, George L. 1964 "Statistics and Sense: More on the Purum Case". Ameri- can Anthropologist. Vol. 66, 1358-1365. Coult, Allan D. 1962 "An Analysis of Needham's Critique of the Homans and Schneider Theory". Southwestern Journal of Anthropology, Vol. 18, 317-335. Geoghegan, William H. and Kay, Paul 1964 "More Structure and Statistics: A Critique of C. Ackerman's Analysis of the Purum." American Anthro-pologist. Vol. 66, 1351-1358. Hsu, Francis 1945 "Observations on Cross-Cousin Marriage in China." Amer- ican Anthropologist, Vol. 47, 83-103. Josselin de Jong, J.P.D. de 1952 Levi-Strauss's theory on Kinship and Marriage. (Med-edelingen van het Rijks museum voor Volkenkunde, No. 10). Leiden. Lane, Barbara S. 1961 "Structural Contrasts Between Symmetric and Asymmetric Marriage Systems: A Fallacy". Southwestern Journal  of Anthropology, Vol. 17, 49-55. Leach, Edmund 1954 P o l i t i c a l Systems of Highland Burma. London. 1961 "Asymmetric Marriage Rules, Status Difference, and Direct Reciprocity: Comments on an Alleged Fallacy." Southwestern Journal of Anthropology. Vol. 17, 343-351. Livingstone, F.B. 1959 "A Further Analysis of Purum Social Structure", American  Anthropologist, Vol. 61, 1084-1087. - 172 -MUller, Ernst W. 1964 "Structure and Statistics: Some Remarks on the Purum Case." American Anthropologist, Vol. 66, 1371-1377. Needham, Rodney 1958 "The Formal Analysis of Prescriptive Patrilateral Cross-Cousin Marriage". Southwestern Journal of Anthropology, Vol. 14, 199-219. 1964 "Explanatory Notes on Prescriptive Alliance and the Purum." American Anthropologist, Vol. 66, 1377-1386. Radcliffe-Brown, A.R. 1952 Structure and Function in Primitive Society. London. Sahlins, Marshall D. 1961 "The Segmentary Lineage: An Organization of Predatory Expansion." American Anthropologist, Vol. 63, 322-345. Wilder, William 1964 "Confusion Versus Classification in the Study of Purum Society." American Anthropologist, Vol. 66, 1365-1371. - 173 -FOOTNOTES 1. Ackerman, 1964. 2. Needham, 1962. 3. Cited in Needham, 1962, p. 13. 4. Needham, 1962, p. 14. 5. IBID., p. 78. 6. Leach, 1961, p. 56. 7. Needham, 1962. 8. IBID. 9. Needham, 1958. 10. Das, 1945. 11. Needham, 1962, p. 15. 12. IBID. 13. Needham, 1958, p. 76. 14. Ackerman, 1964, 53-54. 15. Needham, 1962, p. 9. 16. IBID.,p. 73. 17. Needham, 1964, p. 1379. — note; U i s an abbreviation of KA-U. 18. Das, 1945, p. 125. 19. Needham, 1962, p. 55. 20. IBID.,p. 56 21. Elkin, 1931, p. 296. 22. Lane, 1962. 23. Needham, 1962, p. 8. 24. Needham, 1958, p. 75. - 174 -25. IBID. 26. IBID. 27. IBID. 28. Needham, 1960, p. 499. 29. Fei, H., 1939, pp. 50-1. 30. IBID. 31. Nutini, H., 1965, p. 716. 32. IBID., p. 723. 33. Fei, H., 1939, p. 51. 34. Homans and Schneider, 1955, p. 15. 35. IBID. 36. Leach, 1961, the t i t l e of his essay, p. 54. 37. Needham, 1958, p. 75. 38. Needham, 1962, p. 2. 39. Homans and Schneider, 1955, p. 13. 40. IBID. 41. Cited in Homans and Schneider, 1955, p. 12. 42. Whyte, Harrison C., 1963, pp. 42-47. 43. Homans and Schneider, 1955, p. 19. 44. Marcel Mauss, 1950, pp. 10-11. 45. Leach, 1961, 102. 46. IBID., p. 101. 47. IBID., p. 102. 48. IBID., pp. 56-58. 49. IBID., pp. 99-100. 50. Needham, 1962, p. 4. - 175 -51. Levi-Strauss, C., 1962, p. 324. 52. IBID. 53. IBID., p. 325. 54. IBID. 55. IBID., p. 324. 56. Lounsbury, Floyd, 1964, pp. 1302-1310. 57. Needham, 1960, p. 499. 58. Das, 1945, p. 123. 59. IBID. 60. IBID., p. 131. 61. XSID. 62. IBID. 63. IBID., p. 119. 64. Warner, Lloyd W., 1964, pp. 56-7. 65. Radcliffe-Brown, 1950, pp. 11-13. 66. Warner, 1964, p. 61. 67. IBID., p. 70. 68. IBID., p. 60. 69. IBID., p. 64. 70. IBID. 71. Homans and Schneider, 1955, pp. 37-40. 72. Needham, 1962, p. 59. 73. Das, 1945, 122-3. 74. Murphy, R. and Kasdan, L., 1959, p. 20. 75. IBID., p. 21. 76. IBID. - 176 -77. IBID., p. 26. 78. Needham, 1962, p. 67. 79. IBID., p. 56. 80. Warner, 1964, p. 72. 81. IBID. 82. IBID. 83. IBID., p. 9. 84. IBID., p. 88. 85. IBID., p. 85. 86. IBID. 87. IBID., p. 105, my emphasis on the words "the natives say' 88. IBID., p. 87. 89. IBID., p. 90. 90. IBID., p. 59. 91. IBID. 92. IBID. 93. IBID., p. 16. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items