12
SOUTHWESTERN JOURNAL OF ANTHROPOLOGY V O L U M B 2 0 NUMBER 3 AUTUMN 1964 DESCENT, CATEGORY, AND ALLIANCE IN SIRION6 SOCIETY RODNEY NEEDHAM T HE SIRION6 are a most interesting society, and have attracted repeated attention in rccent anthropological writings, including two articles in this Journal (Needham 1954; 1961). Their interest has now been enhanced by a brief note by Eyde and Postal (1963) which raises useful and resoluble ques- tions concerning the interpretation of ethnographic facts and the aptness of cur- rent terms and notions in the analysis of descent systems. In the present article I wish to examine certain empirical issues in the Siriono ethnography, and then dis- cuss their theoretical significance in the particular context of prescriptive alliance. I The background to Eyde and Postal's note is as follows. In 1961 they pub- lished an ambitious and stimulating article on the development of unilateral cross-cousin marriage and Crow-Omaha kinship systems, testing their theoretical statements against a corrected version of Murdock's "World Ethnographic Sample" (1957). In discussing their use of this source, they pointed out certain changes which they had made, one of which was: "The Siriono are listed [by Murdock] as having bilateral descent, but we consider them to be matrilineal" (Eyde and Postal 1961:769, n. 3). They referred to four of the societies in their sample as having "bilateral descent" (p. 749), and listed the Siriono in a foot- note as one of these societies, but they went on to give their reasons for consider- ing the Siriono as not cognatic but matrilineal: The Siriono are the only socicty in a sample of 204 which violate the rule that preferential matrilatcral cross-cousin marriage does not occur in societies with bi- lateral descent. Because of this uniqueness and because there arc strong theoretical grounds for believing that preferential unilateral cross-cousin marriage can only arise 229

SOUTHWESTERN JOURNAL OF ANTHROPOLOGY · southwestern journal of anthropology volumb 20 • numbe 3 • autum •r 196n4 descent, category an, alliancd in sirione societ6 y rodney

Embed Size (px)

Citation preview

SOUTHWESTERN JOURNAL OF ANTHROPOLOGY

V O L U M B 20 • N U M B E R 3 • A U T U M N • 1964

DESCENT, CATEGORY, A N D ALLIANCE IN S I R I O N 6 S O C I E T Y

RODNEY NEEDHAM

TH E S I R I O N 6 are a most interesting society, and have attracted repeated attention in rccent anthropological writings, including two articles in this

Journal (Needham 1954; 1961). Their interest has now been enhanced by a brief note by Eyde and Postal (1963) which raises useful and resoluble ques-tions concerning the interpretation of ethnographic facts and the aptness of cur-rent terms and notions in the analysis of descent systems. In the present article I wish to examine certain empirical issues in the Siriono ethnography, and then dis-cuss their theoretical significance in the particular context of prescriptive alliance.

I

The background to Eyde and Postal's note is as follows. In 1961 they pub-lished an ambitious and stimulating article on the development of unilateral cross-cousin marriage and Crow-Omaha kinship systems, testing their theoretical statements against a corrected version of Murdock's "Wor ld Ethnographic Sample" (1957). In discussing their use of this source, they pointed out certain changes which they had made, one of which was: "The Siriono are listed [by Murdock] as having bilateral descent, but we consider them to be matrilineal" (Eyde and Postal 1961:769, n. 3 ) . They referred to four of the societies in their sample as having "bilateral descent" (p. 749) , and listed the Siriono in a foot-note as one of these societies, but they went on to give their reasons for consider-ing the Siriono as not cognatic but matrilineal:

The Siriono are the only socicty in a sample of 204 which violate the rule that preferential matrilatcral cross-cousin marriage does not occur in societies with bi-lateral descent. Because of this uniqueness and because there arc strong theoretical grounds for believing that preferential unilateral cross-cousin marriage can only arise

229

2i6 SOUTHWESTERN JOURNAL' OF ANTHROPOLOGY

in unilineal societies, we feel that the Siriono must be matrilineal, since their mar-riage rule appears indubitable. The presence of a Crow terminology supports this conclusion. We therefore take the always dangerous step of disregarding the pub-lished ethnography and consider the Siriono as matrilineal in this paper (1961:769, n. 6; authors' italics).

In a further footnote they again expressly pointed out that they had interpreted

the Kaska and Siriono cases "somewhat differently from the original ethnogra-

phies" (p. 770, n. 15) .

In an article on comparative method and prescriptive alliance I took up this

matter briefly, by way of illustration. Holmberg, I maintained, had clearly re-

ported that the Siriono were matrilineal, to the extent even of having what would

conventionally be called a "Crow" terminology; and I therefore reached the con-

clusion, which I now learn was incorrect, that Eyde and Postal had given them-

selves unnecessary trouble by neglecting to read the original ethnography at al l

and by relying instead on Murdock's characterization. T o this the rider was

added that:

W e all make mistakes, of course, and it is not my intention to point out that these [including another example] are mistakes. The point is that they derive di-rectly from a particular method of sociological investigation, and from the relative neglect of conventional scholarly considerations that it entails (Needham 1962a: 168-169).

This drew from Eyde and Postal the short note of expostulation which has

initiated the present discussion. They write:

Let us be clear that the question under discussion is not whether the Siriono are matrilineal, but whether the author of . . . 'Nomads of the Long Bow' (Holmberg, 1950) says that they are matrilineal (1963:285; authors' italics).

T o settle this question they adduce, not the ethnographic sources that I had cited

(Holmberg 1948; 1950; cf. Needham 1961:242) , but a private letter which they

received from Holmberg, in which the author writes that according to his evi-

dence there are no such unilineal descent groups as matril ineages in Siriono

society. Eyde and Postal go on to suggest that the difference of opinion between

themselves and me is traceable to differing usages of the term "matri l ineal ," and

they maintain that I have failed to make the distinction between matril iny and

matrilocality:

Under Needham's extraordinarily idiosyncratic usage, a society can apparently be

SIRION6 DESCENT, CATEGORY, AND ALLIANCE 231

called 'matrilineal' in the absence of any unilincal descent groups whatever. His criti-cism of our article rests upon the absurd assumption that we share his definition.1

They imply, though they do not state explicitly, that contrary to my inference

they had indeed read Holmberg's monograph, and one supposes them to claim

that they had af ter all "systematically examined the facts" (Needham 1962a:

173) on the Siriono. If this is really their meaning, I can only hasten to admit

that I must have been mistaken in inferring that they had not careful ly read and

made a full-scale analysis (cf. Needham 1962a: 166) of the monograph, and the

discussion which follows will be based on the assumption that they did so. I can

only observe, by way of apologetic explanation, that I simply could not believe

that my colleagues should have read the evidence which we shall consider below

yet still maintain that they were "disregarding the published ethnography" in

taking Siriono society to be matril ineal.

According to my reading of the evidence, Siriono society is "matri l ineal ,"

whereas according to Eyde and Postal, who recognize merely "matrilocal ex-

tended famil ies," the ethnography says no such thing. Th i s is really a most in-

triguing situation, and well worth sorting out. It poses in a clear fashion, and

with regard to a single social feature, a typical problem in the interpretation of

evidence; and an examination of the issues involved may, it is hoped, be of some

methodological interest.

I I

The very first question, obviously, is whether, as Eyde and Postal ask, Holm-

berg says the Siriono are matril ineal. Here are relevant passages from his two

accounts of this society:

The Siriono are organized into bands, each made up of matrilineal extended fam-ilies. . . . (Holmberg 1948: 458; italics supplied).

1 Since the language of their note seems rather inappropriate, perhaps I may observe that I am puzzled to see them put such a polemical construction on my entirely ironic examination of their a rgument . I should explain that I had merely mainta ined that they were factual ly mis-taken in their characterization of three societies out of the hundreds in their sample, and that one of their 25 theoretical conclusions was false, and I inferred that these slips were caused by relying on Professor Murdock ' s l istings. I wrote expressly, however, that I did not at al l wish "to disparage the analyt ical content ( the ideas, as distinct from the method) of [ t h e i r ] inter-esting paper" (1962a : 180, n . 1 7 ) ; and one can only regret that, in response to the suggestion that they might actually be mistaken in some particulars, they should resort to a l leging that I charge them with being "gui l ty of a mult i tude of . . . sins against the canons of anthropo-logical research (as defined by N e e d h a m ) " . These rather vehement words, I should much like it to be appreciated, bear no resemblance to my own, and they are surely not the language of rational discourse.

2i6 SOUTHWESTERN JOURNAL' OF ANTHROPOLOGY

Besides being a member of a nuclear family, every Siriono also belongs to a larger kin group, the matrilineal extended family. Such unilineal kin groupings as moieties, clans, and sibs are not found among the Siriono. Because of matrilocal resi-dence, groups of matrilineal relatives tend to cluster together in the house and to form extended families. An extended family is made up of all females in a direct line of descent, plus their spouses and unmarried children (1950:50; italics supplied).

The ethnographer thus plainly says l iterally that the Siriono are "matri l ineal ,"

and that the largest kin group is composed primarily of women in a direct line of

descent, which here can only mean matri l ineally. It is rather perplexing that

Eyde and Postal should seem to maintain (they do not actual ly do so, in fac t )

that Holmberg does not say that the Siriono are matri l ineal ; and it would have

been helpful, it may be suggested, if they had cited these express and literal

evidences in their rejoinder, so that the grounds for their position should be

clearer.2

In confirmation of this account of the situation, let us see whether or not

other anthropologists have taken Holmberg's report to assert that the Siriono are

matri l ineal :

Descent is matrilineal . . . (Murdock 1951:44, s.v. Siriono).3

Examples of matrilineal bands are found among . . . the Siriono. . . . Each band is matrilocal and matrilineal, and it consists of a number of nuclear families related through the female line (Steward and Faron 1959:385, 428) .

. . . The matrilocal and matrilineal extended family of the Siriono (Spencer 1963:136).

There thus seems to be no doubt in the minds of these authors, including two

leading authorities on South American ethnography, that the Siriono are properly

to be described as matril ineal, and there is no sign that they think themselves to

be departing from the original ethnography in saying so.

Nevertheless, we might all be wrong in believing that the ethnography shows

2 There is also a slight cause for doubt in Professor Holmberg ' s letter, as quoted by Eyde and Postal. H e writes: " I d id not go beyond the statement that the Siriono have matri-local extended fami l ies" (italics suppl ied) , whereas what he actual ly wrote in each of his pub-lished accounts, as we have seen, was "matr i l ineal extended famil ies" . It may be conjectured, in Eyde and Postal 's defense, that they were led by this changed phrasing to lay undue stress on residence groups among the Siriono and may thus have been distracted from lineal considerations.

3 Professor Holmberg carried out his field research in association with Professor Murdoch's Cross-Cultural Survey at Ya le Univers i ty (Holmberg 1 9 5 0 : 1 ) ; Professor Murdock had access to his manuscript before it was published (cf. Murdock 1949 :371 ) , and presumably was in a position to discuss such matters directly with the ethnographer.

SIRION6 DESCENT, CATEGORY, AND ALLIANCE 233

the Siriono to be matrilineal in some sense, and the next thing to do, therefore,

is to see whether it reports any of the conventional signs of matril iny. First of

all , the terminology is that of a lineal descent system, and not of a cognatic

socicty. Moreover, it makes certain so-called "Crow" equations (Holmberg

1950:54; cf . Needham 1961:249-250), and, as we have seen, Eyde and Postal

themselves think that the "Crow terminology" supports their conjectural con-

clusion that the Siriono are matrilineal. Secondly, there are other indications of

the probable importance of matrilineal relationship: an orphan may be raised by

the mother's mother, the mother's sister, or the mother's classificatory sister

(Holmberg 1950:49) , all of them matril ineally-related women; and in one case

a man inherited a wife from his mother's brother (1950 :82) . 4 Third ly , matril iny

is indicated by the inter-generational terminological equations. This may best be

appreciated by comparison with diagrams of patril ineal systems of asymmetric

alliance. When the relationship terms of the Kachin or the Purum, for example,

are represented in identically-constructed diagrams, with these equations arranged

lineally (and otherwise they arc totally incomprehensible), they show a pro-

nounced skew from the top right of the d iagram down to the bottom left

(Leach 1954:305; Needham 1962b:76) ; this is characteristic of such systems,

and is found also in the terminologies of the Batak , Lamet, Mamboru, Thado,

Vaiphei, etc. But if a similar diagram is constructed for the Siriono, by exactly

the same rules except for the matril ineal descent, it shows by contrast a definite

skew in precisely the opposite direction (Needham 1961:244, Fig. 2 ) , and the

identical form of the Mnong Gar terminology (Condominas 1960:19, 21-23),

which also has Crow equations, makes it appear that this formal feature is

characteristic of a matrilineal system of asymmetric alliance. Quite apart from

Holmberg's explicit use of the term "matri l ineal", therefore, his ethnography

shows that Siriono society is not cognatic but is ordered by a matrilineal rule of

descent. In the face of these cumulative indications, it is not easy to see why

Eyde and Postal thought they were "disregarding" the ethnography, as well as

departing from Murdock's "bi lateral" listing, in considering the Siriono as

matrilineal.

Whether the Siriono have matrilineal descent groups, now, is a distinct ques-

tion, and, in spite of Holmberg's report that such "unilineal kin groupings as

4 I overlooked this case in my analysis of Siriono society; it should have been dealt with together with the single reported marr iage with a F Z D D (Needham 1961 :245-246) . It may be significant that both these cases of marr iage outside Ego's genealogical level, which is categori-cally forbidden, were secondary marriages.

2i6 SOUTHWESTERN JOURNAL' OF ANTHROPOLOGY

moieties, clans, and sibs are not found" (1950 :50) , one that is capable to some

degree of empirical investigation. The definition of the Siriono extended family

as "all females in a direct line of descent, plus their spouses and their unmarried

children" (Holmberg 1950:50) gives no idea of the number of generations in-

volved, or of the degree of matrilineal relationship that is reckoned in the com-

position of this group; but taken as it stands it covers such a possible range of

matrilineal relatives as to make it a question whether to call this corporate group

a "matrilineal extended fami ly ," as Holmberg does, or a "matri l ineage." There

is indeed one indication in the ethnography that the latter may perhaps be a

more apt term, viz. the report that one orphan was adopted by "his mother's

paral lel cousin (a classificatory sister) who also happened to be his father's

second wife" (Holmberg 1950:49) . If these two women were co-members of the

same "extended family ," which their common marriage to one matrilocally-

resident man leads one to infer, then their degree of relationship argues a range

of matrilineal connection within a corporate group to which the designation

"matri l ineage" might be more appropriate.

Now to use this term does not imply any corporate activities other than those

which Holmberg ascribes to the extended family , or that it is a named group, or

that it has any particular degree of persistence through the generations, or that

it is part of a hierarchically-ordered segmentary system. It is simply a conven-

tional term, convenient for analytical and comparative purposes, referring to the

ethnographic fact that Siriono society is composed of corporate groups of matri-

lineal relatives. In the band of Aciba-eoko, for instance, which comprised 94

individuals, there were five "extended famil ies" (Holmberg 1950:51) , giving an

average of about 19 individuals each. This is a fair ly small group, but no smaller

than others in ethnographic literature which have also been called lineage groups.

Similar definitional considerations apply to the four extended families, divided

into 14 nuclear families, composing the band of Eantandu, with a total of 58

members. In each band there are local matrilineal descent groups, each compris-

ing between three and four matrilocal nuclear families, and it is in accord with

these ethnographic facts to conclude that the Siriono have a matrilineal descent

system in this respect also.

T o these considerations, moreover, we may also add the arguments adduced

in my analysis of Siriono society as a system of asymmetric alliance (1961:247-

250) , viz. those concerning the significance of prestations to affines and of the

lineal terminological equations. It should be specially noted, however, that we

SIRION6 DESCENT, CATEGORY, AND ALLIANCE 235

still have no evidential warrant to argue that Siriono matril ineal descent groups

are corporate in the sense that they are alliance groups (cf. Needham 1961 :249) ;

and even if they were they would still not have the same importance, in this

small-scale society composed of mainly endogamous bands, as that which all iance

groups typically possess in other asymmetric systems.

Nonetheless, in the light of the above interpretation of the facts, it is signif-

icant indeed that Murdock, with his special access to the original ethnographic

material ,6 should describe Siriono society as composed of "exogamous matri-

l ineages" (1951 :44 ) . He also writes that in listing Siriono institutions in the

"Ethnographic At las" of the journal Ethnology it is assumed that "the matri-

lineal extended families are organized around matril ineage cores" (Murdock

1962 :133) ; and under column 22 ("Matr i l ineal kin groups and exogamy") he

writes "L ," meaning that the Siriono possess "Lineages of modest size, i.e., matri-

lineal kin groups whose core membership is normally confined to a single com-

munity or a part thereof" (1962 :123) . Even more than this, finally, we have a

published admission by Holmberg, reported by Murdock, that he "may con-

ceivably have overlooked" the existence of lineages among the Siriono (Murdock

1962:133) .

In sum, then, as far as the ethnographic literature on the Siriono is con-

cerned, the position is this. In the most literal sense, the ethnographer actual ly

says that the Siriono are matri l ineal ; other anthropologists, reading his reports,

have accordingly taken them to be matri l ineal ; the ethnographic facts give cogent

and independent reason to conclude that the Siriono are in fact matri l ineal ; and

there are matril ineal descent groups which may be conventionally recognized as

matrilineages.®

W e must readily accept that Eyde and Postal had read the published facts

careful ly , thereby taking Siriono society to be cognatic ("b i la tera l" ) , and that

they then believed themselves to be departing dangerously from the original

ethnography in deciding that it was not cognatic af ter all but matri l ineal ; yet it

remains a matter for some perplexity that, in the face of the facts adduced

5 C f . footnote 3 . 6 Since this characterization of the facts does not wholly accord with Professor Holraberg 's

own, I part icularly desire it to be realized that it is not arrived at by any high-handed contro-version of the ethnographer's reports. M y own experience of the difficulties and dangers of research among forest nomads inspires me, to the contrary, with an appreciation of Professor Holmberg ' s ethnography which very few other anthropologists are in any position to share or challenge. W e are considerably in Professor Holmerg ' s debt for having endured what he did to bring us back such information on this fascinating and unusual society.

2i6 SOUTHWESTERN JOURNAL' OF ANTHROPOLOGY

above, they should ever have analyzed Siriono society as cognatic in the first

place.7

I l l

From such specific empirical matters I now turn to more general and conse-

quential aspects of the issue. Eyde and Postal 's note turns on the question

whether a society can be called matril ineal in the absence of any unilineal de-

scent groups. Now, whether or not there are matril ineal descent groups in Siriono

society, are lineal descent groups essential to prescriptive al l iance? In other and

more particular words, if it could clearly be shown that there was nothing in

Siriono society which could reasonably be called a descent group, would it there-

fore follow that it could not after all , in spite of my arguments (1961) , practice

asymmetric al l iance?

T h e focus of studies in prescriptive alliance is the notion of category, and my

own analyses of individual prescriptive systems have all had as their prime aim

the understanding of categories of social classification (cf. Needham 1963:272) .

T h e categories under examination in Siriono society are the relationship terms in

Holmberg's monograph, admirably recorded with 121 genealogical specifications

and other denotations (1950:52-56) . These terms, when analyzed, cohere into a

consistent terminology of asymmetric all iance (Needham 1961:243-245), and it is

by this feature itself that we determine that we are in the presence of such a

system. In the case of the Siriono there is no reasonable doubt about the struc-

ture of the terminology: we have to accept that this, as a matter of plain eth-

nographic fact, is the form of Siriono social classification. W e may be inclined

to infer that there are indeed matrilineal alliance groups in Siriono society, or,

if there arc no such alliance groups, we may well be puzzled to conjecture any

social advantage which the Siriono might derive from a system which is more

readily understandable in other societies which do possess lineal alliance groups;

but neither of these considerations has the slightest effect on the conclusion that

the Siriono employ, through their terminology of social classification, a system

of asymmetric prescriptive all iance.8

7 M y own bewilderment is intensified, indeed, by their note: for otherwise, on the basis of the further examination conducted here, I should only feel the more confirmed in my pre-vious conclusion that they had merely relied on Professor Murdoch 's listing ( 1957 :685 ) , and this I could have understood.

8 Th i s may remove any suspicion that my argument is at al l intended to jus t i fy my dif-ference with Eyde and Postal on whether the Siriono are to be defined as matri l ineal . I should dearly like to know what the decisive facts are, of course, but for my general theoretical con-cerns it is totally immaterial whether the Siriono have corporate descent groups or not.

SIRION6 DESCENT, CATEGORY, AND ALLIANCE 237

The categories order relations, but what the terms denote is indeterminate

and cannot be inferred from the classification. In the Siriono case, they could

apply to persons or to extended families or to shallow matrilineagcs; but the

same terminology could equally well be applied to far larger corporate groups in

a wide-range situation involving a population five hundred times that of the

Siriono. In other words, it is not scale that is important, but only the structure

of the system as expressed by the categories.

More than this, even the rule of descent is structurally unimportant, so long

as the system is harmonic (Levi-Strauss 1949 :270) ; for rule of descent is secon-

dary to rule of prescriptive alliance (Needham 1961:23-24), and not only docs

asymmetric alliance not require descent groups of any kind but it does not even

require a particular rule of lineal descent. The demonstration of this proposition

was the chief aim of my analysis, and this lesson constitutes the first theoretical

interest of Siriono society. Far from being concerned, therefore, to argue the

existence of a system of corporate matrilineal descent groups in Siriono society,

which is what apparently Eyde and Postal require if they are to recognize a

society as matril ineal, I am interested to discern the principles of Siriono social

classification, without presupposing the existence of any kinds of groups to which

this may be applied.

The relative inconsequence of descent groups and of the distinction between

patril iny and matriliny in the comparative study of prescriptive all iance is fur-

ther demonstrated by consideration of two-section systems. Not only do they not

require descent groups or a particular rule of lineal descent, in which they are

l ike asymmetric systems, but it may not be feasible even to say that they are

either patrilineal or matril ineal. Pul Eliya, for example, has a simple two-section

(Dravidian) terminology of symmetric alliance (Leach 1960:124) , but it has no

corporate and enduring unilineal descent groups, and property is heritable by

either men or women. Although it has a "patril ineal ideology", it so far lacks the

conventional features of a unilineal descent system that Leach can go so far as

to say that it is neither unilineal nor non-unilineal (1961:10-11) . This is consis-

tent with the fact that it is impossible to tell from the terminology alone, unlike

asymmetric systems, whether it is that of a patrilineal or a matrilineal society;

but the terminology itself is certainly not cognatic but clearly lineal, and it is

ordered by symmetric alliance, which is wholly inconsistent with a cognatic ter-

minology or society."

9 A South American example is that of the Tr io , in Guiana . They employ a two-section terminology (Riviere 1963) , but M r . Riviere, who is among the Tr io , reports that there is no definite rule of descent (personal communication, February 1964) .

2i6 SOUTHWESTERN JOURNAL' OF ANTHROPOLOGY

In all this the essential distinction to be made is that between the categories

of a descent terminology and the descent groups which they may or may not

variously denote, and this is precisely what is at issue in the case of the Siriono.

The Siriono terminology is that of asymmetric alliance, and this is the crucial

principle of social classification; only secondarily is it to be defined as matri-

lineal, and even then there is no necessity that any kind of descent group shall

correspond to the descent lines.

IV

A more general methodological point, finally, has to do with the possibility of reliance on the listings in Murdock's world ethnographic survey for the pur-poses of a theoretical investigation. It is not necessary to repeat arguments which have already been set out elsewhere (Needham 1962a) , but the more important conclusions can apparently bear re-emphasis. Eyde and Postal may not have been misled by Murdock's listing of the Siriono as "bilateral ," but they must have relied upon his survey for a large proportion of the societies in their sample; and the trouble with the survey is that even if the facts have been careful ly estab-lished the summary characterizations under which they are listed place them be-yond all reach of critical attention. Such a tabular survey is immensely conven-ient for preliminary exploration of an issue, as an ethnographic guide,1 0 but it cannot do our work for us once we have been directed to the societies which we have to understand.

Since no typology, moreover, can represent ethnographic facts in such a fashion that a survey-table shall be able to cope with any conceivable question, it is evident that in order to make a comparative study in any scholarly sense we have first to make an analysis, in particular terms proper to the problem at issue, of each of the societies to be compared. And even when this has been done, so that we believe the facts to be reliably determined in each case, there is still little confidence to be gained from the use of such sophisticated formulas as Yule 's coefficient of association and Yates ' correction in computing chi-square (cf. Eyde and Postal 1961:769, n. 3 ) , for "The method of statistical correlation can only pose questions, it cannot give us the answers to them" (Evans-Pritchard 1963:14) .

The Siriono case is of direct and instructive relevance to these considerations,

10 Professor Murdock 's " W o r l d Ethnographic Samp le" has a permanent place on my own desk; and, far from wishing to decry its usefulness, I only wish there were a vaster thing of the k ind, an electronic device, for example, into which ethnographic information could be fed and which would rapidly retrieve it in any desired combination of features.

SIRION6 DESCENT, CATEGORY, AND ALLIANCE 11

for not only does it raise them in an empirical form, but it shows the lengths to

which it may be necessary to go in order to establish the ethnographic facts and

to find out what they mean. It does not seem idiosyncratic or extreme to maintain

that this has to be done before their significance in relation to a theoretical issue

can be assessed.

BIBLIOGRAPHY

C O N D O M I N A S , G E O R G E S

1960 "The Mnong Gar of Central Vietnam," in Social Structure in Southeast Asia (ed. by George Peter Murdock), pp. 15-23. New York: Viking Fund Publications in Anthropology, no. 29.

E V A N S - P R I T C H A R D , E . E .

1963 The Comparative Method in Social Anthropology. L. T . Hobhouse Memorial Trust Lecture, no. 33. London: The Athlone Press.

E Y D E , D A V I D B . , A N D P A U L M . P O S T A L

1961 Avunculocality and Incest: the Development of Unilateral Cross-Cousin Marriage and Crow-Omaha Kinship Systems. American Anthropologist 63:747-771.

1963 Matrilineality and Matrilocality among the Siriono: a Reply to Need-ham. Bijdragen tot de Taal-, Land- en Volkenkunde 119:284-285.

H O L M B E R G , A L L E N R .

1948 "The Siriono," in Handbook of South American Indians (ed. by Julian H . Steward), Bulletin 143, Bureau of American Ethnology, vol. 3, pp. 455-463.

1950 Nomads of the Long Bow: the Siriono of Eastern Bolivia. Smithsonian Institution, Institute of Social Anthropology, public, no. 10.

L E A C H , E . R .

1954 Political Systems of Highland Burma: a Study of Kachin Social Struc-ture. London: G. Bell and Sons.

1960 "The Sinhalese of the Dry Zone of Northern Ceylon," in Social Structure in Southeast Asia (ed. by George Peter Murdock) , pp. 116-126. New York: Viking Fund Publications in Anthropology, no. 29.

1961 Pul Eliya, a Village in Ceylon: a Study of Land Tenure and Kinship. Cambridge: Cambridge University Press.

L E V I - S T R A U S S , C L A U D E

1949 Les Structures £lementaires de la Parente. Paris: Presses Universitaires de France.

M U R D O C K , G E O R G E P E T E R

1949 Social Structure. New York: Macmillan. 1951 Outline of South American Cultures. New Haven: Human Relations

Area Files. 1957 World Ethnographic Sample. American Anthropologist 59:664-687. 1962 Ethnographic Atlas. Ethnology 1:113-134.

2i6 SOUTHWESTERN JOURNAL' OF ANTHROPOLOGY

N E E D H A M , R O D N E Y

1954 Siriono ?nd Penan: a Test of Some Hypotheses. Southwestern Journal of Anthropology 10:228-232.

1961 An Analytical Note on the Structure of Siriono Society. Southwestern Journdl of An tlnopology 17:239-255.

1962a Notes on Comparative Method and Prescriptive Alliance. Bijdrdgen tot de Taal-, land- en Volkenkunde 118:160-182.

1962b Structure and Sentiment. Chicago: University of Chicago Press. 1963 Symmetry and Asymmetry in Prescriptive Alliance. Bijdrdgen tot de

Tddl-, Land- en Volkenkunde 119:267-283.

R I V I E R E , P . G .

1963 An Ethnographic Survey of the Indians on the Divide of the Guidnese dnd Amazonidn River Systems. Unpublished B.Litt. thesis, University of Oxford.

S P E N C E R , R O B E R T F .

1963 "Fate and Freedom in Primitive Religion." in The Concept of Freedom in Anthropology (ed. by David Bidney), pp. 130-151. The Hague: Mouton.

S T E W A R D , J U L I A N H . , A N D L O U I S C . F A R O N

1959 Native Peoples of South Americd. New York: McGraw-Hill.

UNIVERSITY OF OXFORD

OXFORD, ENGLAND