Referential Ambiguity in the Calculus of Brazilian Racial IdentityAuthor(s): Marvin HarrisSource: Southwestern Journal of Anthropology, Vol. 26, No. 1 (Spring, 1970), pp. 1-14Published by: University of New MexicoStable URL: http://www.jstor.org/stable/3629265Accessed: 19/08/2009 17:02

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available athttp://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unlessyou have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and youmay use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained athttp://www.jstor.org/action/showPublisher?publisherCode=unm.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printedpage of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with thescholarly community to preserve their work and the materials they rely upon, and to build a common research platform thatpromotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.

University of New Mexico is collaborating with JSTOR to digitize, preserve and extend access to SouthwesternJournal of Anthropology.

http://www.jstor.org

SOUTHWESTERN JOURNAL OF ANTHROPOLOGY

VOLUME 26 * NUMBER 1 * SPRING * 1970

Referential Ambiguity in the Calculus of Brazilian Racial Identity'

MARVIN HARRIS

Categorizations elicited from 100 Brazilian informants through the use of a standard- ized deck of facial drawings suggests that the cognitive domain of racial identity in Brazil is characterized by a high degree of referential ambiguity. The Brazilian cal- culus of racial identity departs from the model of other cognitive domains in which a finite shared code, complementary distribution, and intersubjectivity are assumed. Structurally adaptive consequences adhere to the maximization of noise and ambiguity as well as to the maximization of shared cognitive order.

T HE COMPARATIVE STUDY OF RACE RELATIONS in Brazil and the United States has brought to light important differences in culturally con-

trolled systems of "racial" identity. Many observers have pointed out the partial subordination of "racial" to class identity in Brazil exemplified in the tendency for individuals of approximately equal socio-economic rank to be categorized by similar "racial" terms regardless of phenotypic contrasts, and by the adage, "money whitens" (Pierson 1942, 1955; Wagley 1952; Harris 1956; Azevedo 1955). Other aspects of the Brazilian calculus of "racial" identity lead to categoriza- tions that are inconceivable in the cognitive frame of the descent rule which underlies the bifurcation of the United States population into "whites" and "negroes" (now, more politely, "blacks"). Experimental evidence indicates that

phenotypically heterogeneous full siblings are identified by heterogeneous "ra- cial" terms. Children of racially heterogeneous Brazilian marriages are not sub- ject to the effects of hypodescent; where the phenotypes are sharply contrastive, full siblings may be assigned to contrastive categories (Harris and Kottak 1963). It has also long been observed that the inventory of terms which defines the Brazilian domain of "racial" types exceeds the number of terms in the analogous domain used by whites in the United States (and probably by blacks as well).

The suggestion has been made that the most distinctive attribute of the Brazilian "racial" calculus is its uncertain, indeterminate, and ambiguous out-

put. Subordination of race to class, absence of descent rule, and terminological efflorescence all contribute to this result (Harris 1964a, 1964b). Several different

1 I wish to acknowledge the indispensable assistance rendered by Ruth Martinez, Douglas White, Roger Sanjek, David Epstein, Sam Gorenstein, William McGuire, Sanders Kirsch, Paul Byers, and George Morren's mother.

1

SOUTHWESTERN JOURNAL OF ANTHROPOLOGY

indications of the absence of a common shared calculus should be noted: ego lacks a single socio-centric racial identity; the repertory of racial terms varies widely from one person to another (holding region and community constant); the referential meaning of a given term varies widely (i.e., the occasions in which one term rather than another will be used); and the abstract meaning of a given term (i.e., its elicited contrasts with respect to other terms) also varies over a broad range even within a single community.

Clarification of the nature of the ambiguity in the Brazilian "racial" cal- culus awaits the development of cross-culturally valid methods of cognitive analysis. In this essay I report on a preliminary attempt to employ a test in- strument to elicit the Brazilian lexicon of "racial" categories and to provide a measure of referential ambiguity and consensus with respect to the elicited terms.

The instrument employed consisted of a deck of 72 full face drawings con- structed out of the combination of three skin tones, three hair forms, two lip, two nose, and two sex types. All other features were held constant. The draw- ings were presented in a standardized random order. Each respondent was per- mitted to glance at the whole deck before being asked to identify the first draw- ing. Initiation of response was achieved without using terms presumed to be part of the domain. In general, it was found that a question involving a request for the qualidade or typo or rafa of the person depicted in the drawing was

adequate to prime a response process. The word "cor" was used as a last resort. The deck was shown to 100 native born adult Brazilians, 39 women and 61

men, at sites in five different states: Bahia 28; Alagoas 30; Pernambuco 12; Ceara 7; Brasilia 8; Sao Paulo 15. From occupational data and place of birth, the respondents can be classified into the following socio-economic strata: urban

upper class 18; urban middle class 31; urban lower class 13; rural upper class 12; rural lower class 26. On the basis of objective criteria (photo portraits were taken of most respondents), their phenotypes can be classified as predom- inantly caucasoid 42; marked caucasoid/negroid mix 32; predominantly negroid 16; predominantly caucasoid/negroid/indian mix 6; other types 4. The size of the sample and its nonrandom nature obviously render all conclusions highly tentative.

The lexical productivity of the graphic stimuli exceeded expectations. In a strictly lexical count, the sample responded with 492 different categorizations. Twenty-five percent of the sample responded with 15 or more categorizations, with the range extending from 2 to 70 categorizations per respondent, and the median at 9 per respondent.

This diversity of response is the salient characteristic of the cognitive do- main under study. Although under other circumstances, as for example, scien- tific chemical or botanical terminologies, large numbers of terms correlate with precise usage, this is not the case with Brazilian racial categories. Disagreement appears to be a fundamental characteristic of the domain defined by these terms. The most frequently employed terms were, in fact, applied to almost all

2

AMBIGUITY IN BRAZILIAN RACIAL IDENTITY

of the drawings, and each of the drawings was identified by at least 20 different lexical combinations.

It would be incorrect to conclude that this is a totally orderless domain. On the contrary, cognitive models can be constructed which maximize the order inherent in the raw data. The problem, however, is to provide a model which maximizes the order without ignoring the pronounced tendency towards am- biguity, which is also a real aspect of the raw data.

With this objective in mind, analysis of the drawings in relation to response patterns was undertaken, utilizing only the 12 most commonly employed terms, each of which occurred more than 100 times. These terms are: moreno, branco, mulato, preto, negro, alvo, moreno claro, cabo verde, claro, sarari, escurinho, escuro.

a b

c d /c~pll~d

FIG. 1. Examples of male drawings, with their most frequent identifying terms: a) alvo, b) claro, c) cabo verde, d) mulato.

Figures 1 and 2 present the male and female drawings which were most frequently identified by some of the most popular terms.

It is apparent that each of the more popular terms has associated with it a particular combination of skin color, hair form, and nose and lip width. Were

3

"I

SOUTHWESTERN JOURNAL OF ANTHROPOLOGY

this the only information available, these components could be used to define cognitive categories which would appear to be in complementary distribution in the manner of componential schemes now enjoying wide interest among anthro- pologists (see below). Thus a branco could be defined as an individual of male sex, light skin color, thin lips, straight to wavy hair, thin nose; a preto, as an individual of male sex, dark skin tone, thick lips, kinky hair, narrow or wide nose, etc.

a b

c d

FIG. 2. Examples of female drawings, with their most frequent identifying terms: a) preta, b) branca, c) sarard. d) morena.

Such definitions, however, would neglect over 50 percent of the occasions which in fact furnished the stimuli for eliciting the terms branco, preto, etc. Moreover, such a scheme would leave no room for the fact that these terms for many Brazilians are not in complementary distribution. Indeed, almost every "key" term was found to be modifiable by another "key" term in the sample.

Among the more spectacular lexical combinations turned up by the presen- tation of the drawings, special note should be taken of:

4

AMBIGUITY IN BRAZILIAN RACIAL IDENTITY

branco africano branco amarelo branco indio branco mestico branco nago branco serara caboclo preto claro branco indio preto indio moreno moreno caboclo moreno cabo verde moreno escuro claro moreno mestico moreno preto moreno sararA

(white African) (yellow white) (Indian white) (mestizo white) ( * ) ( * ) (

* )

( * ) (black Indian)

( *

) (

* )

( * )

( *

) ( * ) (

* )

( * )

mulato branco mulato caboclo mulato indio mulato mestico mulato sararA

negro branco

negro mulato escuro

negro preto preto amarelo

preto claro

preto louro

preto mestiso

preto moreno

preto negro preto sarara

(white mulatto) ( * ) (Indian mulatto) (mestizo mulatto) ( * ) (white Negro) (dark mulatto Negro) (black Negro) (yellow black) (light black) (blond black) (mestizo black) ( * ) (Negro black)

( * )

Attention should be directed toward the problem of translating these re- sponses. Since the cognitive domain under discussion is radically unconform- able with the experience of non-Brazilians, any gloss into English is misleading. To achieve operationally valid definitions of these terms, each should be paired with the drawings which elicited them. Nonetheless, since this would involve extravagant typographical displays on behalf of being precise about what is ex hypotheosi ambiguous, I have yielded to the expedient of glossing some of the conventionally more acceptable examples. The phrases marked with an asterisk contain terms for which any conventional glosses would obstruct the basic task of explicating the cognitive specialties of the Brazilian system.

The complexities involved in attempting any conventional gloss can be grasped by comparing curves of the frequencies with which particular pictures were paired with the most popular terms.

In Figure 3 the vertical axis was constructed by arranging the 36 male draw- ings in the descending order of frequency (from top to bottom) with which the term branco was applied to each. The horizontal axis indicates with what frequency each of the drawings was identified as either branco, preto, or both. The resulting curves show how certain drawings are called either branco or preto and seldom or never both branco and preto; while other drawings are termed both branco and preto with moderate to low frequency. The latter re- sponse is characteristic of almost half of the drawings. It should also be pointed out that the drawing which was most frequently termed branco was also the one drawing which was most frequently identified by any single term among all the drawings. Yet, the number of respondents calling that drawing by some term other than branco exceeded the number who identified it as branco.

In Figure 4 a similar procedure is employed to show the relationship be- tween the frequency with which each of the female drawings was called either branca or preta, or both branca and preta. Although essentially similar to the

5

BRANCO AXIS Drawing Number

27

7

Is

34

14 26

10

3 .0

33 0o.

22 0

S 0.

23 0: 30 O0 36

240

I ? O 12 0

26 0 / 35 * 0 /

17 6 / 32 *. /

9 ,/ o

2 /. 0 13 / '* 0 16 / 0 -.

/ 0

31 0

29 / 0

20

21 0

6 5

4

/ '9..

(L~~~~~~~~~~~~~~~~0,.0 C?~~~~~~~~~~~Opdo i~~~~~~~~~~~~0 0

4116~~~~~~~oo00.0 0

1)~~~~~~~~~~~~~OO (C~~~~~~~~~00 w

r)~~~~~~~~~~00

I~~~~~~~~~~~~~mo.-

-

C) c,3

r41

0- H

Pj

z

0

(11-

Pd

0

0

1-3

It1

t'll

C) 0--

---- BRANCO

.......... PRETO 00

oo~~~~~~~~~~o 0*'***..,

'**'",.,. ?

2 3 4 5 6 7 8 9 10 if 12 13 14 15 16 17 18 19 20 21 22 23 24 26 26 27 26 29 30 31 32 33 34 35 36 37 38 39404 41 43 44 45 46 47 48 48 60 NUMBER OF RESPONSES

FIG. 3. Branco-preto comparison, male drawings.

AMBIGUITY IN BRAZILIAN RACIAL IDENTITY

BRANCA AXIS Drowing Number

41 ?

45: -

e. *e //

S9 * * /

45 - . .

6 **39 :

*6 . B

61 0 - 4 0 **** - . 4 s

70 * * . .

7 -

s6 0 / 5 ?

so . /... PRETA

63 ./ 0

?5 / . 0 46 / 0

52 /

7 1

44 4

47

72 2*

1 2 3 4 6 e 7 6 9 10 11 12 3 14 15 16 17 18 19 20 21 22 23 24 25 26 27 26 29 30 31 32 33 34 35

NUMBER OF RESPONSES

FIG. 4. Branca-preta comparison, female drawings.

branco preto pattern, there appears to be somewhat greater overlap and

scattering. Figure 5 plots female drawings against male drawings for frequency of

branco and branca identifications. In this graph, each female drawing is

equated with the male drawing which it resembles in all diagnostics except sex. The parallel nature of the two curves indicates that an essentially similar calculus is at work in the case of both the male and female drawings, at least with respect to the branco/preto distinction.

In Figure 6 the frequency of alvo and claro identifications are plotted along the branco/preto axes. These terms show maximum contrast with preto and considerable similarity to branco. Nonetheless, we must guard against the con- clusion that branco/alvo/claro are "synonyms" in some absolute sense. There are many operations which might be utilized to arrive at different measures of and, hence, different definitions of synonymy. One might insist, for example, that two terms were referentially equivalent only when they were applied to both similar and different stimuli with precisely the same frequencies. If branco

7

00

BRANCO AXIS BRANCA EQUIVALENCE Drawing Number Drawing Number

27 .. 41

180 .. 39

34 0 . 45 14 .. 0 68

25 0 40

10 .. - 5 0 3.6ci 61 CI

56 H 22 0 ..* 37 7

? ...d 70 23 .. 0 48 00 30 0 . 4 36 .. /0 56 2 0 .' 0 64 Z is3 / 53 12 O/ 69 O

3 ' 0o 60

17 0 9 > 32 0 BRANCO 38 r

a ............ BRANCA oo 62 0 19 0 54

s / 063 > 2 ?0/ 9

4o 13 0 /57 0 571

31 5, 0 29 62 I

I" / ;~~~42 : 20 ?4 4s C) 2 * 0 es

1 44 4 47 5 72 6 67

1 2 3 4 5 5 7 8 9 10 11 12 13 14 15 1617 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 3 39 40 41 42 43 44 45 46 47 48 49 50 NUMBER OF RESPONSES

FIG. 5. Branco-branca equivalence, male-female comparison.

BRANCO AXIS Drawing Number

27 + 0 0

7 + 0 5 + 0 O

34,+0 34 + 0

26 4 0

o0 + 0 " 23

Ac 0/ 0

33: + * 00

2 2 + 0 / M

0 + PRETO

2 ' 0

23: /

30 '

0

2 4 ? 0

2Is +|o. CLAR

24 +? "

0 ...........VO oRo

3, 2

>

32 /

s + z

4

I 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 19 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 4445 46 47 48 49 30 NUMBER OF RESPONSES

FI. 6. Alvo-claro compared with preto-branco, male drawings.

51 "

24 t o

II 2545 675111' 3415611',,123222222333234567190144444444 NU0E OF REPNE

FI.6+locaocmae wit 'rt-brno,ml drawngs

SOUTHWESTERN JOURNAL OF ANTHROPOLOGY

and alvo mean precisely the same thing, why are they not elicited with precisely the same frequency? "Noise" introduced by the observational operations ac- counts for some, but not all, of the difference.

Figure 7 compares the terms cabo verde, alvo, sarard. Cabo verde and sarard tend to occur in complementary distribution with respect to each other, yet the relatively low overall frequency with which they occur limits our ability to con- firm the trend. Alvo, however, overlaps markedly and systematically with both cabo verde and sarard. Since it has already been shown that branco, preto, alvo,

CABO VERDE-SARARA AXIS Drawing Number

19

29

31

13

9

16

I 7

30

8

24

36

2S

4

25

10

7

6

16

34

3

14

27

26

22

33

20

2

2 1

32

1 8

2e

12

35

. 0

0

N.

N N.

+~~~~~ p Q

,

+ 0) 0

+

0~

0 4

i 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1 I19 20

NUMBER OF RESPONSES

FIG. 7. Cabo verde-alvo-sarard comparison, male drawings.

and claro are systematically related along the same set of dimensions, we now can confirm the existence of a calculus by which the probability that a given drawing will be identified as either alvo, branco, claro, preto, cabo verde, or sarard can be stated.

With the generous assistance of Dr. Douglas White, I have attempted to discover the patterned relationship between these six terms and moreno, moreno claro, moreno escuro, and mulato. The relationship between mulato frequencies

- CABO VERDE

--- ALVO

.......... SARARA

0 0

00

++

10

\o

0 .' o<

COMBINED MORENO- MULATO AXIS

Drawing Number 22 + .

8 23

30

36

26 + 10

16

24

29

33

9

31

25

19

14

7 + I +

13 / 6 /

34 /+ 27 +\

4 0 \

5 0 17

12 o

I a

28 0

35

32

20

21

2

+

+ + \

\ +

/ + o + ,/ o

+ 0

+ /

+ /l + , + 7/

0

0

0 + 0

0

0

0

-.- . MORENO-MORENO CLARO + +

MULATO 0

O o

+

t-I

z 0

rT

tri

z fZq H- H- +-

O + %% 0+ o.' ^

1 2 3 4 5 6 7 6 9 10 11 12 13 14 15 16 17 15 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 30 36 39 40 41 42 43 44 45 46 47 48 49

NUMBER OF RESPOSES

FIG. 8. Moreno-mulato comparison, male drawings.

SOUTHWESTERN JOURNAL OF ANTHROPOLOGY

and the sum of moreno and moreno claro frequencies was found to be that de- picted in Figure 8. All of our attempts to relate these curves to the branco/preto axes have been unsuccessful. If there is an orderly principle by which morenos or mulatos are distiguished from brancos, pretos, serards, alvos, claros, and cabo verdes, it is an extremely complex one. At the moment it seems as if Brazilians will call almost any combination of facial features by the terms moreno or mulato with a high but unpatterned frequency.

It should be noted, in conclusion, that this analysis of the calculus of "racial" identity in Brazil has been carried out in conformity with an epistemological and methodological model which challenges certain assumptions about cog- nitive domains now popular among anthropologists. According to the strategists of the New Ethnography (Sturtevant 1964; Frake 1964; Hammel 1965; Colby 1966) culture is the manifestation of a finite shared code, the code being a set of rules for the socially appropriate construction and interpretation of mes- sages. This school has published numerous descriptions of what are alleged to be shared finite sets of rules which account for the terminological distinctions in several different domains. The principal domains studied have been: kinship terminologies (e.g., Hammel 1965; Lounsbury 1965; Goodenough 1965a; Rom- ney and D'Andrade 1964; and Wallace and Atkins 1960); deference patterns (Goodenough 1965b); disease categories (Frake 1961); color categories (Conklin 1955); firewood categories (Metzger and Williams 1966); and the spacing of house sites (Frake 1962). The ambiguous output of the Brazilian racial calculus casts doubt on the assumption that the codes or rules associated with the ab- stract distinctions and actual identification of many classes of phenomena con- stitute intersubjectively uniform sets. Equally plausible is the assumption that actual classificatory performance is the expression of indeterminate and variable "competence" (cf. Chomsky 1966:4). This assumption is especially attractive if the prime social function of the rules is not the maintenance of orderly dis- tinctions but the maintenance or even maximization of noise and ambiguity. Brazilian racial categories appear to constitute such a domain. Lacking caste distinctions based on racial identity, Brazilian social structure contains no evi- dent practical requirements for achieving high levels of intersubjective "com- petence" with respect td raciological taxonomies. On the other hand, given the highly stratified nature of Brazilian society, there may be a positive, conserv- ative structural reason for maintaining and maximizing the noise and ambi- guity in the calculus under discussion. Objectively, there is a correspondence between class and race in Brazil (cf. Fernandes 1964); the more negroid the phenotype, the lower the class. Prevention of the development of racial ide- ology may very well be a reflex of the conditions which control the develop- ment of class confrontations. In the United States, racism and racial caste divi- sions have split and fragmented the lower class. In Brazil, racism and caste formation would unite the lower class. "Black power" in the United States lacks the revolutionary potential of the preponderant mass; "black power" in Brazil contains this potential. The ambiguity built into the Brazilian calculus of racial identity is thus, speculatively at least, as intelligible as the relative precision with which blacks and whites identify each other in the United States.

12

AMBIGUITY IN BRAZILIAN RACIAL IDENTITY

BIBLIOGRAPHY AZEVEDO, THALES DE

1955 As Elites de C6r; un Estudo de Asencao Social. Sao Paulo: Companhia Editora Nacional.

CHOMSKY, NOAM 1966 Cartesian Linguistics. New York: Harper and Row.

COLBY, B. N. 1966 Ethnographic Semantics: a Preliminary Survey. Current Anthropology

7:3-32. CONKLIN, HAROLD C.

1955 Hanun6o Color Categories. Southwestern Journal of Anthropology 11: 339- 344.

FERNANDES, FLORESTAN 1964 A Integracdo do Negro a Sociedade de Classes. Sao Paulo: Faculdade de

Filosofia, Ciencias e Letras da Universidade de Sao Paulo. FRAKE, CHARLES

1961 The Diagnosis of Disease among the Subanun of Mindanao. American Anthropologist 63:113-132.

1962 Cultural Ecology and Ethnography. American Anthropologist 64:53-59. 1964 "Notes on Queries in Ethnography," in Transcultural Studies in Cognition

(ed. by A. K. Romney and R. G. D'Andrade), pp. 132-145, American An- thropologist, Special Publication, vol. 66, no. 3, part 2.

GOODENOUGH, WARD 1965a "Yankee Kinship Terminology: a Problem in Componential Analysis," in

Formal Semantic Analysis (ed. by E. A. Hammel), pp. 259-287. American Anthropologist, Special Publication, vol. 67, no. 5, part 2.

1965b "Rethinking 'Status' and 'Role': Toward a General Model of the Cultural Organization of Social Relationships," in The Relevance of Models for Social Anthropology (ed. by Max Gluckman and Fred Eggan), pp. 1-22. New York: Praeger.

HAMMEL, EUGENE

1965 "A Transformational Analysis of Comanche Kinship Terminology," in Formal Semantic Analysis (ed. by E. A. Hammel), pp. 65-105. American Anthropologist, Special Publication, vol. 67, no. 5, part 2.

HARRIS, MARVIN

1956 Town and Country in Brazil. New York: Columbia University Press. 1964a Racial Identity in Brazil. Luso-Brazilian Review 1:21-28. 1964b Patterns of Race in the Americas. New York: Walker and Co.

HARRIS, MARVIN, AND CONRAD KOTTAK 1963 The Structural Significance of Brazilian Racial Categories. Sociologia 25:203-

209.

KOrTAK, CONRAD 1966 The Structure of Equality in a Brazilian Fishing Community. Unpublished

Ph.D. dissertation, Columbia University, New York, New York. 1967 Race Relations in a Bahian Fishing Village. Luso-Brazilian Review 4:35-53.

LOUNSBURY, FLOYD 1965 "Another View of the Trobriand Kinship Categories," in Formal Semantic

Analysis (ed. by E. A. Hammel), pp. 142-185. American Anthropologist, Special Publication, vol. 67, no. 5, part 2.

13

SOUTHWESTERN JOURNAL OF ANTHROPOLOGY

METZGER, DUANE, AND GERALD WILLIAMS 1966 Some Procedures and Results in the Study of Native Categories: Tzeltal

"Firewood." American Anthropologist 68:389407. PIERSON, DONALD

1942 Negroes in Brazil. Chicago: University of Chicago Press. 1955 "Race Relations in Portuguese America," in Race Relations in World

Perspective (ed. by Andrew W. Lind). Honolulu: University of Hawaii Press.

ROMNEY, A. K., AND R. G. D'ANDRADE 1964 "Cognitive Aspects of English Kin Terms," in Transcultural Studies in

Cognition (ed. by A. K. Romney and R. G. D'Andrade), pp. 146-170. American Anthropologist, Special Publication, vol. 66, no. 3, part 2.

STURTEVANT, WILLIAM 1964 "Studies in Ethnoscience," in Transcultural Studies in Cognition (ed. by

A. K. Romney and R. G. D'Andrade), pp. 99-131. American Anthropologist, Special Publication, vol. 66, no. 3, part 2.

WAGLEY, CHARLES (ED.) 1952 Race and Class in Rural Brazil. Paris: UNESCO.

WALLACE, A. F. C., AND J. ATKINS 1960 The Meaning of Kinship Terms. American Anthropologist 62:58-80.

COLUMBIA UNIVERSITY NEW YORK, NEW YORK

14