You are on page 1of 98

r

Published by the Press Syndicate of the University of Cambridge


The Pitt Building, Trumpington Street, Cambridge CB2 1RP
40 West 20th Street, New York, NY 10011, USA
10 Stamford Road, Oakleigh, Melbourne 3166, Australia Contents
and
Editions de la Maison des Sciences de l'Homme
54 Boulevard Raspail, 75270 Paris, Cedex 06

@ Maison des Sciences de l'Homme and Cambridge University Press 1990

.@nry of Congress Cataloging-in-Publication Data


hanguage and the politics of emotion.
(Studies in emotion and social interaction) ,*
"Grew out of a session of the 1987 Annual Meeting
of the American Anthropological Association called List of itft*riUutors
*
page v.i
'Emotion and Discourse' "-Pref. Preface Vll
1. Sociolinguistics. 2. Emotions. 3. Discourse
analysis. L Lutz, Catherine A. II. Abu-Lughod, Lila. 1 Introduction: emotion, discourse, and the politics of
III. Series: everyday life 1
P40.L284 1990 306.4'4 90-7374 LILA ABU-LUGHOD AND CATHERINE A, LVTZ
ISBN 0-521-38204-1 hardback
ISBN 0-521-38868-6 paperback 2 Shifting politics in Bedouin love poetry 24
LILA ABU-LUGHOD
British Library Cataloguing in Publication Data
Language and the politics of emotion. 3 Moral discourse and the rhetoric of emotions 46

-(Studies
in emotion and social interaction) GEOFFREY M. WHITE
1. Man. Social interactions. Role of emotions
l.l.1ttz, Catherine A. II. Abu-Lughod, Lila IIL Series 4 Engendered emotion: Sender, power, and the rhetoric
302 of emotional control in American discourse 69
CATHERINE A. LUTZ
ISBN 0-521-38204-1 hardback
ISBN 0-521-38868-6 paperback 5 Topographies of the self: Praise and emotion in Hindu India 92
ISBN 2-7351-0367-6 hardback France only ARJUN APPADURAI
ISBN 2-7351-0368-4 paperback France only
6 Shared and solitary sentiments: the discourse of friendship,
play, and anger in Bhatgaon 113
DONALD BRENNEIS
& 7 Registering affect: heteroglossia in the linguistic
exPression of emotion 126
JUDITH T. IRVINE
8 Language in the discourse of the emotions 1,62
DANIEL V. ROSENBERG
9 Untouchability and the fear of death in a Tamil song L86
MARGARET TRAWICK

Index 207

,,,=ETEi]F,
Contributors Preface

&;1r
;!{,..-

h
Lila Abu-Lughod Catherine A. Lutz
Departments of Religion and out of affiJporogical fieldwork in India, Fiji, the United States, Ewpt,
Department of Anthropology
. Anthropology Senegal, and the Solomon Islands, and out of close critical readings of
State University of New yori earlier work in the anthropology of emotion, linguistics, and semiotics,
Princeton University Binghamton, Ny 13901
Princeton, NJ 08544 the authors in this volume have explored the interplay of emotion talk
and the politics of everyday social life. The contributors share a sense of
Daniel V. Rosenberg
the beauty and intricacies of conversation, as well as its involvement in
Department of Anthropology
Arjun Appadurai power - whether the power to move others or the power that shapes
Harvard University
Department of Anthropology discursive forms and the social relations in which they participate. They
Cambridge, MA 0213g
University of Pennsylvania show the many ways discourse- becomes emotional and emotion be-
Philadelphia, PA 19104 comes discourse, and they treat narratives, conversatiOn, performanCes,
Margaret Trawick
poetry, and song not as texts for cultural analysis but as social practices
Department of Anthropology
with serious effects. For this reason, we are huPPy to have the Cam-
Hobart and William Smith
Donald Brenneis bridge series on emotion and social interaction as the home for this book.
College
Department of Anthropology we would like to thank Paul Ekman and Klaus R. scherer, the series
Geneva, NY
Pitzer College 141456
cditors, for their enthusiastic suPport for a project that is so wholeheart-
Claremont, CA917l1 cdly cross-cultural and suspicious of psychology's essentialisms.
Geoffrey M. White
we would like to thank a number of other people who helped bring
Institute of Culture and
this proiect to fruition. The anthropological study of emotion and dis-
Judith T. Irvine Communication
Department of Anthrg>ology is an exciting and growing field; our work is enriched by being
East-West Center ".rrri"
in conversation with that of many others who do not have essays in this
Brandeis University 7777 East-West Road
Waltham, l|i4A0Z2S4 volume. That is a matter of history. This particular book grew out of a
Honolulu, }{I96848
scssion at the 1987 annual meeting of the American Anthropological As-
sociation called "Emotion and Discourse." we are grateful to Geof-
frcy M. white for helping to formulate the theme of the session and for
ruii.itir,g many of the papers in connection with it. ]ohn Sabini and Fitz
John Poiter Poole contributed generously to the lively dialogue
at the
c0nference and to our thinking about the issues. Ariun Appadurai, Tim-
othy Mitchell, Margaret Trawick, and Geoffrey M. white gave us helpful
critical rcadlnga of the introductory chapter. Steven Feld and Stephen
vt
vii
:

viii Preface

Foster gave us invaluable suggestions for improving each of the chapters


and the manuscript as a whole. Lila Abu-Lughod is grateful to the Insti-
tute for Advanced study in Princeton and to the National Endowment
L. Introduction: emotion, discourse,
for the Humanities for support to work on this project and for the ex- and the politics of everyday life
traordinary opportunify to participate in the Gender Seminar (1987-g),
two of whose members, Donna Haraway and Judith Butler, provided
clues for developing the arguments about discourse. For help with typ-
LILA ABU-LUGHOD AND CATHERINE A. LUTZ
;i:.JnB for the volume, we thank Lucille Allsen and Margaret Roe. We are
:,". grateful to Hugh M. Lewis for his careful and intelligent indefng. The
publication process has been made pleasant and productive by susan
&
Milmoe, Helen Wheeler, Laura Dobbins, and Sophia prybylski.
Our work on this book is for Jonathan and Lianna, and for Tim.
#.{# Emotions are one of those taken-for-granted obiects of both specialized
knowledge and everyday discourse now becoming part of the domain
of anthropological inquiry. Although still primarily the preserve of phi-
losophy and psychology within the academic disciplines, emotions are
also-ordinary concerns of a popular American cultural discourse whose
relationship to such professional discourses is complex and only par-
tially charted. Tied to tropes of interiority and granted ultimate facticity
by being located in the natural body, emotions stubbornly retain their
piu"", in all but the most recent anthropological discussions, as the
"rru.r or
aspect of human experience least subiect to control, least constructed
learned (hence most universal), least public, and therefore least amen'
able to sociocultural analysis. The essays in this collection seek
to dem-
onstrate, on the contrary, that the sociocultural analysis of emotion is
both feasible and important and to suggest new ways of going about it.
In this introductory chapter, we begin by setting out four strategies
that have been or could be used to develop the anthropology of emotion:
essentializin g, relattirzing, historicizing, and contextualizing emotion
discourse. we then consider the field of meanings and diverse deploy'
ments of the key term "discourse," without which, we argue' "emo'
&. tion,, cannot properly be understood. Paying special attention to the
theoretical terms "discourse" is meant to replace, we argue that the moet
productive analytical approach to the cross-cultural study of emotion ig
io examine discourses on emotion and emotional discourses as social
practices within diverse ethnographic contexts. Finally, we review the
.o**on themes and apecific arguments of the essays in this collection,
drawing out thelr contributions to a new aPProach to emotion, an aP'
proach-dietlngulshed by its focus on the constitution of emotion, and
lven the domiln of emotion itself, in discouree or eituated EPeech Prac'
ticee, by lts conrtilal of emotion as about social llfe rather than internal

i!@::&+,
j"
..

Lila Abu-Lughod and Catherine A. Lutz lntroduction

unlike much of the earlier ethnopsychological work on emotion, their culturally and the applications to social organization of emotional prac-
interpretive approach to emotions stresses not what culturally variable tice are variable, then any certainties about universals are undermined.
ideas about emotion can tell us about other "deeper" psychological pro-
cesses, but rather what implications these ideas have for social behavior A second strategy for those interested in emotions as sociocultural phe-
and social relations. These analysts helped place emotions squarely in nomena is to historicize them. That means subjecting discourses on
the realm of culture by pointing to the ways local cultural concepts of emotion, subjectivity, and the self to scrutiny over time, looking at them
emotion such as the Ilongot liget (anger) and the pintupi ngaltu (compas- in particular social locations and historical moments, and seeing whether
sion) borrow from broader cultural themes and reflect, in their ideolog- und ho* they have changed. Although a host of potential studies re-
ical shAib, the forms of indigenous social relationships. If these works main to be done, a few works have attempted this sort of investigation.
did n6talways or consistently deessentialize emotions (see Rosenberg, some have been concerned with the history of formal and informal the-
this volume), they certainly began the important process of suspending ories,-of emotions in the west, and others have examined the fate of
concern with the psychological paradigm. For both, furthermore, differ- partictt*bmotions (Cancian 1987; Gardiner, Metcalf, and Beebe-Center
ences observed in talk about emotion had to be traced to social structure igZO, Nfu.furlane t98Z; Stearns and Stearns L986). Norbert Elias (L978)
rather than to a pure realm of autonomous ideology. has argued, mostly from a reading of etiquette manuals, that vast trans-
while some of the work of relativizing has been done by examining formalons of affective life in Europe took place concomitant with the
specific concepts of emotion used in different cultures, many sfudies of development of the absolutist state. Among these he includes an expan-
emotion even show how fragile the category itself is. For example, How- sion ofthe contexts in which disgust occurs and a diminution of aggres-
ell (1981) argues that for the Chewong (Malaysia), what we call ,,affect,, sive affect or behavior. That he calls this the "cii:ilizing process" is
is seen as a minor phenomenon; talk about emotion is replaced by talk symptomatic of his uncritical interpretation of these changes as involv-
about normative rules that provide, she argues, ,,an idiom fo, Y. or- ing i refinement of a somehow preexisting affectivity, a position that
ganizing the individual's relationship to himself, to his fellow[s] . . . , many anthropologists would regard with skepticism. still, his work opens
and to nature and supemature" $a\. Obeyesekere (1935) shows that in up an argument about the kinds of changes that have taken place in one
sri Lanka emotion is likely to be taken as a sign of Buddhist religious geographical, historical setting.
prescription achieved or unachieved. For the Ifaluk (Micronesia), emo-
- otrrer scholars have examined these changes in terms of the disap-
tion is often construed as moral judgment and has a similar pragmatic pearance of or shift in the social locus of various emotions, as well as
force (Lutz 1988). the mar,ipulation of emotional discourses for state PurPoses. The prob-
In Riesman's work on the Fulani of west Africa, a subtre transition lem of ,ud.,"tt has received an impressive number of historical treat-
from the analysis of particular emotion concepts and their role in social ments. Jackson (1985), like Harr6 and Finlay-|ones (1986), takes on the
relations to the questioning of the very cultural meaning and social focused task of tracing the extinction of an emotion called "accidie" and
structural effects of emotionality itself illustrates the direction we think the significance of the obsolescence of "melancholy," both so important
the anthropology ofuemotion ought to take. In his earlier work, Riesman during medieval times, in the contemporary period. sontag (L977) ar-
(1977) was especially concerned to lay out the dimensions of Fulani no- gues that the nineteenth-century Romantic movement came to celebrate
tions of pulaaku (translated as 'Fulaniness' but something others might individuality in part by viewing sadness as a mark of refinement, as a
have called 'honor') and semteende, ot'shame,. In his later work (19g3), quality that made the person suffering from it "interesting." The rise of
he began to make a suggestive argument linking social hierarchy to emo- individualism brought with it the celebration of difference; one of the
tionality itself (see also Irvine, this volume), arguing that self-control or routes by which the new individuals could distinguish themselves was
relative lack of emotional expressiveness is simultaneously taken as a through a focus on feelings defined as aspects of unique personalities'
badge of, justification for, and realization of the social superiority of no- Radden (LgE7) takes these views further by noting that melancholy was
bles over their ex-slaves. If the meaning of emotionality differs cross- primarily a male complaint, one that was at least in part socially valor-

-;,"EiIGi*=#ir_-
6 Lila Abu-Lughod and Catherine A. Lutz lntroduction 7

ized. She argues that the related modern discourse on depression differs It is a strategy followed by the authors of the chapters in this volume,
in pinpointing women as its bearers and in portraying the syndrome as all of whom share a concern with emotion and begin with the assump-
more unequivocally deviant, deficient, and medical in nature.3 In a dif- tion that it is a sociocultural construct. They go on to explore, through
ferent vein and in a non-Western setting, Good and Good (1938) explore close attention to ethnographic cases, the many ways emotion gets its
the ways in which the Islamic Republic of Iran now organizes, to an meaning and force from its location and performance in the public realm
unprecedented degree, both public and private emotional discourses. It of discourse. They also ask how social life is affected by emotion dis-
has transformed the public discourse of sadness and grief, which before course. To assess the nature and value of this strategy first requires at-
the revolution was central to religious ritual, self-definition, and social tention to the term,at its center: the word "discourse'"
under$&nding, into a sign of political loyalty to the state. "Discourse" has become, in recent years, one of the most popular and
Wttdt-might be most productive, however, would be to begin by trac- least defined terms in the vocabulary of Anglo-American academics. It
ing the genealogy of "emotion" itself so that, in an enterprise analogous pervad.4s the humanities and now haunts many of the social sciences.
to Foucault's (1978) critical investigation of the production of "sexuali$r" iratheffih being alarmed by its spread, however, it might be better to
in the modern age, we might consider how emotions came to be consti- ask whfsci many have adopted it. The best way to pursue that question
tuted in their current form, as physiological forces, located within indi- is to consider what theoretical work they want the term to do'
viduals, that bolster our sense of uniqueness and are taken to provide As everyone readily admits, defining discourse precisely is impossible
access to some kind of inner truth about the self (Abu-Lughod, this vol- because of the wide variety of ways it is used. To get a sense of why
ume; Lutz 1986). One promising line of questioning might be to build people use it, and why we have found it useful in thinking about emo-
on Foucault's insights about the growing importance of confession (to iio", it might be helpful to consider what terms it replaces. What is dis-
which a discourse of emotion is often bound both inside and outside course ,,oi? To what is discourse counterposed? This varies by disci-
psychotherapy) as a locus of social control and discourse produ&on in pline, but we will be concerned only with anthropology because its
the eighteenth and nineteenth centuries. peculiar appropriation of the term from the French poststructuralist vo-
Foucault's description of his own project suggests more directly how cabulary is^inflected by the prior and concurrent usage of the term by
emotion discourse might represent a privileged site of the production of anthropological linguists.
the modern self. He writes, in the second volume of The History of Sex- First, particularly for those whose concerns are linguistic, the term
uality, that he wishes "to analyze the practices by which individuals were discourse marks an approach to language as spoken and used rather
led to focus their attention on themselves, to decipher, recognize, and than as a static code analyzable apart from social practice. In Saussure's
acknowledge themselves as subjects of desire, bringing into play be- languelparole distinction, discourse would fall on the side of parole' what
tween themselves and themselves a certain relationship that allows them thole who invoke discourse in this context might want to add, however,
to discover, in desire, the truth of their being" (Foucault 1985:5-6).4 He is that langue either does not exist (e.g', Hopper L987) or at least is al-
also notes that in each historical period it is "not always the same part *uy, in particular utterances by particular individuals. In
of ourselves, or of gur behavior, [that] is relevant for ethical judgment," "*6odied
privilegrng speech, those who use the term discourse generally also want
but in contemporary Western society, "the main field of morality, the io asse.t ine importance of pragmatics versus semantics' The "code,"
part of ourselves which is most relevant for morality, is our feelings" whether it be grammar, structure, model, or, in this case, some pur-
(1983:238). Feelings can play this role because they are currently consti- ported underlying presocial emotional matrix, is taken as emergent in a
tuted as the core of the self, the seat of our individuality.s social context, even if it is not analyzed as a peculiar western cultural
construct.
The third strategy is to focus on social discourse, building less on an- Although in some senses associated with speech, discourse is also
thropology's comparative bent or the broad historical framing of the .o**or,ly used instead to suggest a concern with verbal productions
problem than on a commitment to careful analysis of the richness of more formal, elaborate, or artistic than everyday conversation. Ex-
la-
specific social situations, whether here or there, as Geertz (1987) puts it. amples of claaElc forms of discourse in this sense are Poems, songs,
i

8 Lila Abu-Lughod and Catherine A. Lutz lntroduction 9

ments, prayers, myths, and verbal dueling forms such as sounding


Although only beginning to find its way into anthropological writing,
(Labov 1972). Discourse is also used by some who identify with post-
discourse in this much wider Foucaultian sense is being adopted to do
modemism in its literary incarnation to stress the spoken quality oi lu.,-
the theoretical work of refiguring two terms that it replaces: culture and
guage (Tedlock 1983, L984 and to evoke its dialogic aspect, allegedly
ideology. For many, the no less definable term "culture" has become
ignored by those of us who live in literate societies. yet others use the
problematic for several reasons. First, built into it is a distinction be-
term discourse as a way of including even the nonverbar, like music,
crying, or the "unsaid" of past utterances and present unarticulated
i*""r a realm of ideas, even if public rather than in people's heads, and
material realities and social practices, a distinction some users of dis-
imagination (Tyler 1978) in our consideration of the meanings humans
course would like to problematize. Second, the term seems to connote a
make;6"
certain coherence, uniformity, and timelessness in the meaning systems
slterzer's (198n recent article advocating a discourse-centered ap-
of a given group, and to operate rather like the earlier concept of "race"
proach to language and culture demonstrates the wide range of uses and
in iden{fying fundamentally different, essentialized, and homogeneous
resulting ambiguity of the term. Blending many of these senses of dis-
course together in his "purposely. vague" definition, he writes that dis-
socialffi (as when we speak about "a culture"). Because of these as-
sociafiofis, invoking culture tends to divert us from looking for contests
course is
for meaning and at rhetoric and power, contradictions, and multiple
a level or component of language use . . . [which] can be oral or discourses, or what some now refer to as "heteroglossia" (see Irvine,
written and can be approached in textual or sociocultural and social- this volume).6 It also falsely fixes the boundaries between groups in an
interactional terms. And it can be brief rike a greeting and thus absolute and artificial waY.7
smaller than a single sentence or lengthy like a novel or narration "Ideology" too has come to carry with it meanings that some social
of personal experience and thus larger than a sentence 4gd con- theorists want to shed. The Marxist alternative to culture, it has the vir-
structed out of sentences or sentence-like utterances. . . Dis- tues of seeming less unifying than culture. It can be pluralized even
course is an elusive area, an imprecise and constantly emerging within one society, and is always linked to historically specific social
and emergent interface between language and culture, created by groups assumed from the start to be engaged in struggles of domination
actual instances of language in use. (296) and resistance. However, it retains, perhaps even more strongly than
the notion of culture, the radical distinction between a realm of ideas
The unfortunate vagueness of this definition is the product of a failure
and a material or social reality because of its historical association with a
to grasp that terms are used to signal perspectives and to carve out aca-
distinction between base and superstructure.s And even more proble-
demic domains, not just to refer to definabre entities. The kinds of usages
matically, it sets up an implicit opposition between itself, denoting a
we have described thus far for the increasingly employed term discourse
mystifying or at least motivated and interested vision of the world, and
could be characterized as largely sociolinguistic or literary. All that is
some sort of uninterested, unmotivated, and obiective truth available
being keyed is an interest in language in context, texts, and the public
either to a class or, perhaps more commonly, to the critical social scien-
and social character of what we study. And for the most part, the term
tist. Foucault uses discourse to suggest his rejection of these dualisms
as used in this vol&me stays well within this range of meanings.
that are easily and sometimes unconsciously evoked by the notion of
Hovering around the edges of many of the chapters and informing
ideology.e
the project of the volume as a whole is another way of using discourse,
Although the chapters in this volume do not explore many of the im-
one with more ambitious theoretical goals and different disciplinary roots.
plications of Foucault's work, they do remain faithful to his premise that
Discourse in this other sense is a word that has been taken up by those l'dir.orrr., . . . [are] . . . practices that systematically form the obiects
who find the critique of social theory associated with French poststruc-
of which they speakf ' (Foucault L972:49). For the final work discourse is
turalists like Michel Foucault persuasive, or at least those who have be-
meant to do. as soCial theory/ is to suggest a concern not so much with
gun to borrow its vocabulary. with this move, the semantic field and
meaning as wlth a kind of large-scale pragmatics. Taking texts and talk
pragmatic deployment of the term have begun to shift.
and all eorte of other eocial Practices as productive of experience and
:,

10 Lila Abu-Lughod and Catherine A. Lutz


lntroduction 11

constitutive of the realities in which we live and the truths with which
that can be seen in social interaction, much of which is verbal. As the
we work, this approach also considers how power might produce dis-
sociolinguist Gumperz has also said of discourse studies, "mere talk to
courses as well.lo In suggesting that we attend to the efficacy of dis-
produce sentences . . . does not by itself constitute communication. Only
course, this newer and wider usage still resembles the more limited so-
when a move has elicited a resPonse can we say communication is tak-
ciolinguistic uses outlined earlier. yet it goes further by looking at more
ing place" (1982:1). Attention to discourse leads us therefore to study
than speech, by rccognizing the local, contradictory, and fragmented
new problems, such as how an audience's response to emotional perfor-
character of discourses, and by insisting that discourses be understood
mances can be unpredictable given the former's ability to attend to only
in relation not just to social life but to power.
some parts of the performance and to make idiosyncratic sense of those
Thuq.,although each of the authors in this volume uses discourse dif-
parts. Attention to discourse also leads us to a more comPlex view of
fererfldy, the term, resonating with its many current uses, stands as a
the multiple, shifting, and contested meanings possible in emotional ut-
token of our common wariness of mentarist models, our refusal to treat
terance{ and interchanges, and from there to a less monolithic concept
language as simply reflecting thought or experience, and our insistence
of emffi. The focus on discourse allows not only for insight into how
that all those productions in a cornmunity that could be considered cul-
emotioil like the discourse in which it participates, is informed by cul-
tural or ideological be analyzed as sociai practices, tied to relations of
tural themes and values, but also how it serves as an operator in a con-
power as well.as to sociability.
tentious field of social activity, how it affects a social field, and how it
can serve as an idiom for communicating, not even necessarily about
The chapters in this collection takes discourse, often as the situated so-
feelings but about such diverse matters as social conflict (white, this
cial practices of people speaking, singing, orating, or writing to and about
volume), gender roles (Lutz, this volume), or the nature of the ideal or
each other, as a point of entry for the study of emotion. They address
deviant person (Fajans 1985).
one or both of two issues: the discourse on emotions scientific oPevery-
- The study of emotion as discourse allows us to explore how speech
day, western or non-western - and emotional discourses, that is, dis-
provides the means by which local views of emotion have their effects
courses that seem to have some affective content or effect. Differing in
and take their significance. If earlier scholars who rejected the notion
the extent to which they bring the category of emotion itself into ques-
that emotion was sensation preferred the notion of emotion as judgment
tion and in the degree to which they speak as if emotions are internal
(Solomon 1976), their view has since been supplemented by the insight
things or not (and whether it even matters), the authors also differ in
that judgments might better be viewed as socially contested evaluations
the aspects and forms of language they explore. Nevertheless, they all
of the world phrased in an emotional idiom and evident in everyday
approach emotion through language and understand language as ines-
speech behavior. Rather than seeing them as expressive vehicles, we
capably and fundamentally social.
must understand emotional discourses as pragmatic acts and commu-
The turn here to discourse is a turn to detailed, empirical studies of
nicative performances. The more Seneral interest in the social sciences
conversation, poetics, rhetoric, and argument about and with emotional
in how language implements social reality coincides with the interest in
content. Building o^n the work of others who have explored facets of
how emotions are sociocultural facts. If emotions are social phenomena,
emotion in perfornffince and language (Basso 19g5; Brenn eis 19g7; Cra-
discourse is crucial to understanding how they are so constituted.
panzano 1989; Feld 1983; Good, Good, andFischer 19gg; Irvine 19g2; Ochs
The most important theme running through the chapters is that emo-
and schieffelinlgSg; sabini and silver 7987 and 19gg; B. schieffelin and
tion and discourse should not be treated as separate variables, the one
Ochs 1986; E. SchieffelinlgT6; Urban 1988; White and Kirkpatrick 19g5),
pertaining to the private world of individual consciousness and the other
we argue for a view of emotion as discursive practice. what advantages
does this have for our understanding of emotion? what can those inter-
to the public social world. Taking seriously Wittgenstein's (1966) in-
sights about the relationship between emotion and language, articulated
ested in emotion learn from considering its relation to discourse?
first in his description of what kind of "language-game" talk of ioy and
In contrast to other approaches, the emphasis on discourse in study-
anger is, we artue that emotion talk must be interpreted as in and about
ing emotion keeps us fixed on the fact that emotions are phenomena
soclal llfe rather than ae veridically referential to some internal state.

.gr*
i

12 Lila Abu-Lughod and Catherine A. Lutz


lntroduction L3

Emotion should not be viewed, as our quotidian perspective might


social relations (e.g., when ang4r glaring represents the imposition of
suggest, as a substance carried by the vehicle of discourse, exp.essedly
moral obligation), and (3) practices that reveal the effects of power (as in
means of discourse, or ',squeezed through,,, and. thereby perhaps
dis_ gestures of respect and shame in many cultures).ll
torted in, the shapes of language or speech. Rather, we shourd
view The move to ensure that emotions remain embodied, however, should
emotional discourse as a form of social action that creates effects in
the be seen as more than an attempt to position them in the human body.
world, effects that are read in a culturally informed way by the audience
Embodying the emotions also involves theoretically situating them in
for emotion talk. Emotion can be said to be *eated in, rathirthan
shaped the social body such that one can examine how emotional discourses are
by, speech in the sense that it is postulated as an entity in ranguage
formed by and in the shapes of the ecologies and political economies in
whgg#ts meaning to social actors is also elaborated. to say this is not
which they arise.
to fucuce the concept of emotion to the concept of speech, even
though Emotion can be studied as embodied discourse only after its social and
a discourse-centered approach might be construed is a rejection
or ob- culturalL its discursive - character has been fully accepted. To take lan-
scuring of the body.
grrg"$;ore than a transparent medium for the communication of
Although in this volume we focus on emotion as discourse, working
ir,n"r tnt"ghts or experience, and to view speech as something essen-
to pry emotion loose from psychobiology, that does not mean that we
tially bound up with local power relations that is capable of socially con-
do not recognize the possibility that emotions are also framed in most
structing and contesting realities, even subjectivity, is not to deny non-
contexts as experiences that involve the whole person, including
the linguistic "realities." It is simply to assert that things that are social,
body (see Appadurai, this volume). Here Bourdieu,s thoughts on,,Lody
political, historically contingent, emergent, or constructed are both real
hexis" are suggestive, providing ways of thinking ,boul the fact that
and can have force in the world.
emotion is embodied without being forced to concede that it must
be
"rtatrtral" and not shaped by social interaction. He defines bci8y hexis '[his volume goes a long way toward establishing the pragmatic force of
as a set of body techniques or postures that are learned habits
or deepry emotion discourse and the social character of emotion by showing how
ingrained dispositions that both reflect and reproduce the social rela-
centrally bound up discourses on emotion (local theories about emo-
tions thgt surround and constitute them. The child, for instance, learns
tions) and emotional discourses (situated deployments of emotional lin-
these habits by reading, via the body rather than the mind,s eye, guistic forms) tend to be with social issues. Because we think that it will
the
c,ltural texts of spaces and of other bodies (Bourdieu 1977:90). be more theoretically productive, we have made central, in organizing
Extending this definition to the emotions enables us to grasp how this volume, questions about the ways emotion discourse can be related
they, as culfural products, are reproduced in individuals in tire form to the social. We have not been particularly concerned with cross-cul-
of
embodied experience. To learn how, when, where, and by whom
emo_ tural differences or regional/cultural issues, although nearly all the chap-
tions ought to be enacted is to learn a set of body techniques including tc.rs make sensitive contributions to the ethnography of societies in In-
facial expressions, posfures, and gestures. For example, rather than clia, Fiji, the Solomon Islands, EWPI,Senegal, and the United States'
thinking or sp_eaki#g the respect (gabarog) that helps reproduce a gender Also, despite their intrinsic interest, we have not stressed the types of.
hierarchy on Ifahjk atoll in Micronesia, girls follow tire curve of their cliscourse subjected to analysis because of a link with the emotional. The
mothers' backs in embodying the bent-over posture of respect. simi- range, however, is impressive. The chapters analyze poetry, song, and
larly, emotions such as love or friendship that are thought to emanate gther aesthetic performances, narratives, actual conversations, inter-
from ineffable positive feelings between two people *igt t be cued, views, regulated modalities of verbal interaction, Iinguistic registers, and
Bourdieu notes (1972:82), by a sensed similarity tf utay rrexis produced scientific discourse.
by being reared under similar physicar and sociar i,ve might Two aspects of social relations emerge as crucially tied to emotion dis-
eventually develop an anarysis of the kinds of bodily"orditior,r.
discourse on emo- course: sociability and power relations. The links to sociability can be
tion that includes emotional postures that are simultaneously (1) phe- reen in the salience of emotion language in settings where solidarity is
nomenologically experienced, (2) vehicles for symbolizing ani affecting
being encouraged, challenged, or negotiated, or in the essentially inter-

i#a
..

1,4 Lila Abu-Lughod and Catherine A. Lrttz lntroduction 15

actional nature of discourse as it engages performers or speakers and emotion and emotional discourses can serve, in other instances, for the
audiences or interlocutors. Fajans (19g5) had earlier shown ihrt th" relatively powerless as loci of resistance and idioms of rebellion (Abu-
"or"
of Baining (New Britain) emotional discourse was concerned with threats Lughod, Ltrtz, Trawick), as means of establishing relationships and
to social cohesion; a central emotion term, translated as ,,hunger,,, was coercing gifts (Appadurai), or even as means of establishing complemen-
used to talk about the importance of ties to others and their mediation tarity with status superiors (Irvine). More broadly, the chapters tend to
through food exchange. In this volume, the chapters by white and Bren- concentrate on the politics of emotion discourse by looking either at the
neis describe ethnographic contexts in which a relatively formalized pragmatics of emotion talk, the social deployment of particular emo-
emotional discourse is used to promote social harmony. The A,ara prac- tional discourses, or the politics of ideologies of emotion.
tice aqgasi-therapeutic discussion to tark about and contain recent social Criticizing the referential models of language used by many anthro-
confliets that threaten a valued community or kin group sociability, and pologists and the assumptions about human nature that animate much
Indian Fijians enact perforrnances whose emotional gestures draw in work, {articularly that on cultural meaning and cognition, Abu-Lughod
their audiences rather than alienate them. argr{ffiAt we must ask not just what the cultural meanings of various
Recent work has begun to show that power seems even more thor- emotiofrs are and how emotional configurations might be related to so-
oughly bound up with such discourses. we rook particularly for the ways cial lir'e, but how emotional discourses are implicated in the play of Power
power relations determine what can, cannot, or must be said about self and the operation of a historically changing system of social hierarchy.la
and emotion, what is taken to be true or false about them, and what Building on earlier arguments about what she calls the "politics of sen-
only some individuals can say about them. The real innovation is in timent," she analyzes how Egyptian Bedouin love poetry, believed to
showing how emotion discourses establish, assert, challenge, or rein- have a certain force in the world, is now being deployed to challenge
force power or status differences. Discourses on fear have been sineled male elders. She also shows how this emotional discourse comes to have
out in a number of studies of colonial viorence as crucial urp"iP, or",n" new social meaning and a different social basis as the Bedouin political
discursive practices of dominant groups (stoler 19g5; Taussij tla4. rurc economy is being transformed. Taking as her central case the love poems
of fear of the dominated other in coronial contexts can be interpreted as (on cassette) of a young man whose marriage was thwarted by an uncle,
a meang by which powerful groups accomplish several purposes. she explores the ways in which the introduction of semicommercial cas-
They
justify their suppression of those their rhetoric of fear implicitly sette recordings combines with the erosion of the tribal ideology con-
painis
as powerful and threatening to erupt, as Taussig (1gg7) urgr", comitant with the economic transformations of Bedouin life to exclude
occurred
among rubber collectors in Columbia in the early part oitnir century. women from this discourse of defiance. That poetry, as an emotional
As stoler (1985) demonstrates in the case of Dutch pianters in Indonesia, discourse, is seen by the Awlad 'Ali Bedouins as having pragmatic force,
they also thereby bargain with other elites for the resources and support as suggested by the effects and intentions of playing these cassettes in
needed to face down the purported threat.l2 particular social contexts.
scheper-Hughes's fieldwork in a Brazirian underclass community traces Defending the importance of constructing models of indigenous con-
the relationrfi!_b-*-"un emotional discourse and political economy. In ceptual models in anthropological analyses of emotion, White argues
one analysis (1985IJ she shows how a purportedly universal mother love that the ethnography of emotion actually offers an opportunity to ex-
is replaced by an emotional rhetoric of detached waiting regarding young plore points of convergence between situated practices and interpretive
infants because of the high infant mortality rate. In rrrott*, (19gg), she models. In his chapter he tries to reconcile analytically, and show the
discusses how the syndrome neroios is part of discursive practice that interaction of, conceptual r.nodels and social institutions in a practice
transforms the symptoms of hunger into the less politically charged terms called'disentangling' found on one of the Solomon Islands. The core of
of emotional anxiety and "nerves" and of individuat pairrotofrr, whose his chapter is a fascinating analysis of a narrative from one disentangling
therapyis tranquilizers rather than a redistribution of ?ood, arrd meeting that shows how this discourse, whose overt purpose is to make
power.13
-!utth,
bad feeling public, worke through the narrative reconstruction of prob-
Authors in this volume have explored instead how discourses on lematic evenB to emphaalze reconciliation instead of rehibution by means

ffE
;

16 Lila Abu-Lughod and Catherine A. Lutz Introduction 17

of emotion language. The discursive practice of disentangling works to ,,inner" states of those involved, he shows how the
anything about the
create social harmony by re-creating and valorizing social relations. yet practices of praise nevertheless create sentimental bonds with social
power is not absent; through attention to both ideology and pragmatics, consequences.
white reveals that the same narrative performance using the rhetoric of f*e Wfrite and Appadurai, Brenneis begins with the argument that
reconciliation can simultaneously work to establish the moral advantage we have to take indigenous theories of emotion seriously because they
of the speaker over those with whom he or she is in conflict. This alerts inform emotional performances. Noting a tendenry in the new literature
us to the crucial importance of analyzing emotional discourses for mul- on language and emotion to focus on speakers, he warns us not to over-
tiple meanings, intentions, and effects. look the audiences who actively interpret and respond to emotional
In.eryoloring the many ways that ordinary Americans' discourses on communications and may become interlocutors. He suggests that local
emd,t{on are related to gender ideology, Lutz makes a strong case for the ,,social aesthetics" among a grouP of Hindi-speaking Indians in Fiji in-
argument that this emotion discourse is only apparently about internal form a{dience responses to communicative events. Here sociability and
state but in fact about social life, power relations in particular. she pre- emotiffiis.or6" appear to be closely associated in two ways. First,
sents examples of talk about emotions in a group of American women village* distinguish between social and individual emotions but only
and men, and argues that this discourse on affect is also a discourse on posif,vely value the former. Second, socially recognized emotions like
the nature of women, their subordination, and their potential for rebel- amity and friendship are the only emotions to be indexed and enacted
lion in Ameritan society. A "rhetoric of control" (of emotion) found. more in performances that are social and sociable. one implication of under-
frequently in women's than men's conversations is one of the primary standing how emotions are generated in particular types of events is
that we can begin to see how SrouPS excluded from participation in par-
ways people tell a narrative of women's weakness. she also traces the
deep resonance of this lay discourse on the reration between gender and
ticular events may thereby also be precluded from having certain emo-
emotion with scientific discourse on the same topic. But finally, hef%hapter
tional experiences. Brenneis hypothesizes that women in that commu-
presents a finding that demonstrates the importance of distinguishing nity may be prevented from realizing the socially valued emotions of
carefully between multiple levels of discourse. Her analysis of the orga--
,o.irfifiy because they do not join men in certain performances and
nization of emotion discourse on a syntactic level shows how it actually social settings.
fails to differentiate female and male speech. This suggests that ideas Turning t6 another dimension of the relationship between social per-
about links between emotion and the female, however pervasive at the
formance and emotion, Irvine explores one important way that linguistic
ideological and narrative levels, do not organize discourse at the more
structure is tied to emotion. she argues for the coPresence in many lan-
microscopic or unconscious levels. guages of registers, that is, situational variations in language use' many
Taking a different strategy for the exploration of the complex relation- iir *r,i"n have an affective dimension. Proposing the term "affective reg-
ship between discourse and emotion, Appadurai begins with a cultur- isters,, to suggest a culturally defined set of complementary representa-
ally inflected discursive form: praise in Hindu India. He then proceeds tions of linked to conceptions of the Person as well as the situ-
to show that understanding the meaning and pragmatic force of such a "-otior,
ntion, Irvine analyzes the differences between two wolof (senegal) styles
form in Indian life$equries attention to s=everal things: the multiple and
of speaking in these terms. she shows how a variety of features - Pros-
mutually relevant contexts (praise of the divine, of kings, of patrons, ody, phonology, morphotogy and syntax, lexicon, discourse manage-
and assessment of people and goods) that give it a particular meaning;
ment and interactional devices - distinguishes the speech of two social
the social uses to which this form is put in a variety of different social castes, nobles and griots (a.hereditary caste of bards) and how, further-
relationships (from flattery of politicians to coercion by beggars); and the
more, the contrasts in speech styles are rooted in images of the Person'
indigenous theories of emotion and the locar topography of the self,
Even more important than her argument that conventional linguistic
which render our own western judgments of its excesses and emotional displays of affectivity index social divisions is lrvine's suggestion that
inauthenticity inapplicable. Arguing that praise is a public, regulated wolof
the eesential complementarity of this heteroglossia also helps-the
discourse and an embodied strategy of interaction that does not assume difference neverthelese aociable'
define relatloneh$t of Power ae

, ffi,=
..

1-8 Lila Abu-Lughod and Catherine A. Lutz


Introduction t9
stepping back from the emergent field defined by this volume, Rosen-
berg offers a critical overview of the first wave of anthropological work the psyche or the natural body, they show clearly how discourses on
on emotion and personhood done by cultural anthropologists claiming emotion and emotional discourses are commentaries on the practices
an interest in language. with a linguist's eye, he reads some represen- essential to social relations. As part of the politics of everyday life, these
tative texts for the thorny problems raised when language is invoked as discourses are not, therefore, just the stuff of psychological anthropol-
a locus of meaning and as a methodological key. He argues that despite ogy but of sociocultural and linguistic theory as well. The chapters to
an avowed interest in situated discourse and rejection of ethnogrupt i. follow offer positive ways of developing both a nonindividualized and a
semantics, most earlier sfudies make a number of problematic theoreti- nonreductionist approach to emotion and a more dynamic socially and
cal mog-es. For example, they abstract individual words, mostly nouns, politically gtounded analysis of all discursive practice.
frorg 1fieir discursive contexts and then recontextualize them into a so-
cial matrix; they use these key terms as master metaphors for a culture;
they expand the references of nouns to incrude mental models or sche- Notes {
mas and misattribute to others the ideas we have about language and t. fo'rTffi"nt reviews, see Heelas (1986), Levy and Wellenkamp (1987), and
meaning; and through inadequate attention to the way actuai-conversa- Lutz and \a/hite (1986).
For a recent consideration of anthropology's (with the exception of feminist
tion proceeds, they mistake grammatical or indexical features of lan- anthropology's) role in developing cultural critiques/ see Marcus and Fischer
guage for nontrivial cultural facts. The problem with these moves, he (1e86).
argues, is that they smuggle back into studies of emotion and person- 3. This shift may have corresponded to the general process of medicalization
and normalization that, for Foucault (1978), characterize the modern age.
hood our ethnopsychology and our metalinguistic habits, making it dif-
4. See also Foucault (1983) for a clear discussion of his views on the relation-
ficult to distinguish methodologicar differences from cultural differ- ship between subjectivity and subjection or the creation of the individual
ences. He sorts out a number of distinct ways emotion and l#nguage through disciplinary power.
might be related and suggests that future work attend carefully to the Foucault's assertion, of course, calls for ethnographic evidence, the begin-
nings of which Lutz (1988:53-80) provides. It would be worth speculating
distinctions between semantics, reference, pragmatics, and ideoiogy. further whether the proliferation of emotion discourse in American life,
Bringing together issues of the language of power, the power of lan- combined with the construal of emotion as a private and subjective state,
guage, d'nd the entanglement of emotio., ana power, Trawick closes this might not both confirm a sense of self as separate (in giving the individual
"experiences" of his or her own - as Lutz (1986:299), Riesman (7983:123),
volume with a lyrical and itself quite moving meditation on caste pollu- and Foucault (1985:5) have argued in linking the construction of "experi-
tion, the fear of death, and a song sung into a tape recorder for her by ence" and the sense of individuality) - and provide an idiom for asserting
an untouchable Tamil woman from south India. Building on some of the existence of bonds between people in the face of the actual attenuation
of such bonds by mobility, distance, and the social fragmentation of class,
fulia Kristeva's thoughts on abjection and ranguage, she tries to answer gender, and race.
the question of "how it feels to be beyond the pale." Through a close 6. Heteroglossia is a term that seems to have filtered into anthropology, both
textual and stylistic analysis of this hymn, Trawick reveals the singer,s in the narrow linguistic sense of many languages and in the larger sense of
concern with the problems of inclusion and exclusion so crucial to caste many discourses, through Bakhtin (1981). For a critical discussion of the
absence of social theory in and the conservative implications of most of the
and with the issu& of separation and remainders so critical to a sense work on culture done under the rubric of cognitive anthropology, see Kees-
of personal wholeness. she argues that the singer's artistic technique, ing (1987). The notion of culture promoted by interpretive anthropology has
which involves deviating from the code of grammar as well as the social many critics, but Asad's (1983) consideration of the problems as related to
the study of religion is particularly intelligent.
code, is a strategy for challenging that which has cast her out. 7. See Appadurai (1988) for a persuasive argument that "natives," people from
In suggesting in their many ways that we consider not emotions but certain faraway places who belong to those places and are somehow incar-
the discourses of emotion, the chapters in this book do not deny the cerated in those places and especially in their "mode of thought," are "crea-
force of emotion and subjective experience. They do advocate a shift in tures of the anthropological imagination" - that is, produced by anthropo-
logical discourse. For a discussion of the similarity of the concepts of culture
focus that may be illuminating. Arguing that the reality of emotion is and race, see Mitchell (1988:105).
social, cultural, political, and historical, just as is its current location in See Willlama (1973, Pn) and Comaroff (1985) for attempts to mediate this
divide.

.dG
..

20 Lila Abu-Lughod and Catherine A. Lutz lntroduction 21,

9. For an elaboration of the problems with ideology, see Foucault (L980:1,17- Foucault, Michel. 1972. The Archaeology of Knowledge and the Discourse on Lan-
18). gzage. New York: Pantheon.
10. Foucault himself substituted the term 'apparatus' (dispositifl for discourse 1978. The History of Sexuality, Vol. 1. New York: Pantheon.
in some of his later work on sexuality in order to emphasize that he was 1980. PowerlKnowledge. Colin Gordon, ed. New York: Pantheon.
concerned with "a thoroughly heterogeneous ensemble" of nondiscursive 1983. On the Genealogy of Ethics: An Overview of Work in Progress. In
elements - statements, writings, architectural forms, rules, institutions, etc. H. Dreyfus and P. Rabinow, Beyond Structuralism and Hermeneutics,Znd, ed.
- that are related to one another in varying ways and have, as a formation, Chicago: University of Chicago Press, pp. 229-52.
"a dominant strategic function" (1980:19t1-5). t985. The Use of Pleasure. Vol. 2 olThe History of Sexuality. New York: Random
11. See Scheper-Hughes and Lock (1987) for a view of the "three bodies" that House.
can be applied to the three bodies of emotion just described. Gardiner, H. M., Ruth C. Metcalf, and ]ohn Beebe-Center. 1970 (1937). Feeling
12. Skilalso shows the power of the denial of fear by those same planters. and Emotion: A History of Theories. Westport, CT: Greenwood Press.
,Qeflial or negation both posits a fear and a threat and claims to have con- Geertz, Clifford. 1973. Person, Time, and Conduct in Bali. ln The Interpretation of
luared them (cf. Kress and Hodge 1978). Cultures. New York: Basic Books, pp. 36G411.
13. See also Hochschild (1983) on the relationship between power and the D87. lnorks and Liaes. Stanford: Stanford University Press.
emotional practices of service workers, such as stewardesses, in the United Cieert4ffi-dred. 1959. The Vocabulary of Emotion: A Study of Javanese Sociali-
States. ,uffiiro."r ses. P sychiatry 22:22*37 .

1,4. Other strong critiques of the referential view of language have been made Cood, Miry-|o Delvecchio, and By.o.r Good. 1982. Toward a Meaning-Centered
by Crapanzano (1981) and Good and Good (1982). Analysis of Popular Illness Categories: 'Fright Illness' and 'Heart Distress'
in Iran. In A. Marsella and G. White, eds., Cultural Conceptions of Mental
Health and Therapy. Dordrecht: D. Reidel, pp. 1,41-46.
1988. Ritual, the State, and the Transformation of Emotional Discourse in Ira-
References
nian Society. Culture, Medicine, and Psychiatry 12(l):4T63.
Abu-Lughod, Lila. 1985. Veiled Sentiments: Honor and Poetry in a Beilouin Society. Cood, Mary-]o Delvecchio, Byron Good, and Michael M. ]. Fischer. 1988. Emo-
Berkeley: University of California Press. tion, lllness and Healing in Middle Eastern Societies. (Special Issue). Culture,
Appl{qrai Arjun. 1988. Putting Hierarchy in Its Place. Cultural Antfo'opology Medicine and Psychintry 72(7).
3:3649. (iumperz, ]ohn. 1982. Discourse Strategies. Cambridge: Cambridge University Press.
Asad, Talal. 1983. Anthropological Conceptions of Religion: Reflections on Geertz. llarr6, Rom, and Robert Finlay-|ones. 1986. Emotion Talk Across Times. In Rom
Man (n.s.) 18:237-59. Harr6, ed., The Social Construction of Emotions. Oxford: Basil Blackwell, pp.
Bakhtin, MikhailM. 1981. TheDialogiclmagination. M. Holquist, ed. Austin: Uni- 22U33.
versity of Texas Press. lleelas, Paul. 1986. Emotion Talk Across Cultures. In Rom Harr6, ed,., The Social
Basso, Ellen. 1985. A Musical View of the Uniaerse. Philadelphia: University of Construction of Emotions. Oxford: Basil Blackwell, pp.234-66.
Pennsylvania Press. I liatt, L. R. 1984. Your Mother-in-Law Is Poison. Man 19:183-98.
Bourdieu, Pierre. 1977. Outline of a Theory of Prnctice. Cambridge: Cambridge I lill, Jane. 1987. Weeping and Coherence in Narrative and Selfhood: The Sor-
University Press. rows of Dona Maria. Paper presented at the annual meetings of the Ameri-
Brenneis, Donald. 1987. Performing Passions: Aesthetics and Politics in an Oc- can Anthropological Association, Chicago.
casionally Egalitarian Community. American Ethnologist 1,4:236-50. l"lochschild, Arlie. 1983. The Managed Heart. Berkeley: University of California
Biggs,lean. 1970. Neaer in Anger. Cambridge, MA: Harvard University Press. Press.
Cancian, Francesca. 1987. Looe in America: Gender and Self-Deoelopment. Cam- I lopper, Paul. 1987. Emergent Grammar. Proceedings of the 13th Annual Meeting
bridge: Cambridge University Press. ot' the Berkeley Linguistics Society. Berkeley, CA: Berkeley Linguistics Society,
Comaroff, fean. 198Q. Body of Power, Spirit of Resistance: The Culture and History of pp.139-57.
a South At'rican Ftople. Chicago: University of Chicago Press. I lowell, Signe. 1981. Rules Not Words. In P. Heelas and A. Lock, eds., lndige-
Crapanzano, Vincent. 1981. Text, Transference, and Indexicality. Ethos9:12248. nous Psychologies. London: Academic Press, pp. 133-43.
1989. Preliminary Notes on the Glossing of Emotions. Kroeber Anthropological lrvine, Judith. 1982. Language and Affect: Some Cross-Cultural Issues. In
Society P ap er s N os. 69-7 0 :7 U85. H, Bymes, ed., Contemporary Perceptions of Language: lnterdisciplinary Dimen-
Elias, Norbert. 1978 (1939). The History of Manners, trans. E. ]ephcott. New York: sions. Washington, DC: Georgetown University Press, pp. 31,-47.
Pantheon Books. Iackson, S. W. 1985. Acedia the Sin and Its Relationship to Sorrow and Melan-
Fajans, fane. 1985. The Person in Social Context: The Social Character of Baining cholia. In A. Kleinman and B. Good, eds., Culture and Depression: Studies in
"Psychology." In G. White and |. Kirkpatrick, eds., Person, Self , and Experi- the Anthropology and Cross-Cultural Psychiatry of Affect anil Disorder. Berkeley:
ence: Erploring Pacific Ethnopsychologies. Berkeley: University of Califomia Press, University of California Press, pp. 4162.
pp.367*97. Keesing, Roger. 1987. Models, "Folk" and "Cultural": Paradigms Regained? In
Feld, Steven. 1983. Sound and Sentiment: Birds, Weeping, Poetics anil Song in Kaluli D, Holland and N. Quinn, eds., Cultural Models in Language and Thought.
Expression. Philadelphia: University of Pennsylvania Press. Cambridger Cambridge University Press, pp. 369-93.
;,

26 Lila Abu-Lughod Shifting politics in Bedouin looe poetry 27

tracing the cultural systems in which these related concepts participate, human sciences and what he calls, following Nietzsche, our modern
and considering the situations to which these emotion constructs are Western "will to truth" are part.
tied.3 As Lutz and White (L986:420) point out, more attention is now begin-
Although I am sympathetic to the intent of this work, have con- ning to be paid to the relationship between emotions and social struc-
tributed to it myself (Abu-Lughod 1985, 1986), and believe it begins the ture.a In many cases, this turn to the social is found within the interpre-
important task of relativizing, there are at least three aspects of it that I tive paradigm. Myers (1986:107), for instance, who argues for a concern
find problematic. I will take up the first two here and the third later in with emotional constructs' "place within a larger cultural system of
my discussion of discourse. First, for many involved in this work of meanings," adds that "if the emotions are relational, the relationships
demddstrating the emotions to be cultural rather than natural, the West- they constitute are given meaning and value by the social process in
ernthought-feeling dualism has been seen as a key to our mistakes. As which they are embedded." Rosaldo too, in the latter part of her article,
a solution, however, the stress on the inherently cultural character of beging Eploring the notion, with regard to shame, guilt, and egalitarian
emotions risks assimilating emotion to thought. Myers (1986:106) makes versus'ftffiarchical societies, that "notions of the person, affective pro-
the point that "emotions are not.simply reactions to what happens, but cesses aiitl forms of society itself are interlinked" (1984:148). She con-
interpretations of an event, judgments [about situations]." Lutz and White cludes by arguing that we are "social persons" (1984:151) and that what
(1986:428) conclude that in the new work on emotions, "emphasis is is needed is to relate "lives of feeling to conceptions of the self, as both
shifted away from the question of whether a somehow decontextualized of these are aspects of particular forms of polities and social relations"
emotional experience is 'the same' or 'different' across cultures to that (1984:150). In suggesting/ as do many whose work on emotions con-
of how people make sense of life's events." Rosaldo (1984:137-8) argues siders the realm of the social, that there might be some correlation be-
that "thought is always culturally patterned and infused with feelings, tween emotion constructs and forms of society, she raises the somewhat
which themselves reflect a culturally ordered past," and sug$sts that troubling possibility of a new brand of national character studies.s This
feelings must be understood as "embodied thoughts" (1984:138). In can be avoided only if society is not conceptualized as a unitary body
privileging in their theories of emotion activities such as understanding, but rather as composed of individuals and groups with competing inter-
making sense of, judging, and interpreting, these theorists may be in- ests involved in relations of power, and if politics is not reified in the
advertently replicating that bias toward the mental, idealist, or cognitive notion of a polity.
that Lutz (1986) points out is such a central cultural value for us. I prefer to follow out a more dynamic strand of Rosaldo's argument,
A second and related problem with the cultural approach is that it one that carries more sociolinguistic assumptions, as a promising direc-
assumes that all humans are primarily engaged, like social scientists or tion for the anthropological study of emotion. She writes that "what
philosophers, in the project of interpretation and understanding rather, individuals can think and feel is overwhelmingly a product of socially
than other practices. Rosaldo (7984:139) outlines a trend in the social rrrganized modes of action and of talk" (L984:1,47).I would use this as-
sciences that has at its center the desire to grasp "how human beings pect of her argument against her interpretive bent to argue that instead
understand therrgelves and [how] to see their actions and behaviors as of enriching our concept of culture (1984:138), we need to break with it
in some ways the creations of those understandings." It strikes me that, by pressing harder on the question of social action and talk. In other
without in any way denigrating their philosophical inclinations or con- words, we need to examine emotion as discursive practice (see Abu-
sciousness, the people we study are just as likely to be interested in Lughod and Lutz, this volume, for a critique of the culture concept).
other aspects of living as in interpreting or undertstanding. Thus, rather '[his brings up the third problem with much cultural analysis, a problem
than relativizing Western psychology by calling it "ethnopsychology," that has received a good deal of attention in this volume and elsewhere:
thereby attributing the essentially contemplative project of psycholo. It generally models itself on theories of language as referential and com-
gizing, or for that matter anthropologizing, to everyone, we might want municative. This promotes a lingering concern with meaning, which al-
to bring into question, as Foucault (L970) does, the very peculiar and ways implies that it is the referents that are the object of study rather
historically specific developments of which the enterprises known as the than the speakers (see also Rosenberg, this volume). I would argue that

k
.

28 Lila Abu-Lughod Shifting politics in Bedouin looe poetry 29

the first step must be to ask how emotion discourses are deployed in see as their superior morality. But, as I will discuss, even this is begin-
social contexts. This would shift the concern from what Foucault has ning to change.
argued is characteristic of and widespread in the modern West - a focus Here is the love story. I was back in Egypt in 1985, visiting for the first
on what is said in discourse - to the more interesting and political ques- time since my initial fieldwork the families I had lived with for almost
tions of what discourse is, what it does, and what forms it.6 What we two years. It was early in the morning of the last day I was there' My
need to know is how discourses on emotion, or emotional discourses for host, the head of the family, with whom I had lived as a sort of adopted
that matter (see Abu-Lughod and Lutz, this volume), are implicated in daughter, was getting ready to drive me to Cairo to catch my plane. He
the play of power and the operation of historically changing systems of rummaged around in the pockets of his various robes and vests, looked
sociahl-derarchy (see also Trawick, this volume). in his briefcase, and finally in exasperation asked his children, who were
Chie can look at the relationship between emotion, social li[e, and power all standing around, if they knew what had happened to the cassette of
as Lrttz (1986) does by noting the ideological functions of emotion attri- pathallqAi-jbehi. His eldest daughter sheepishly went and got it from
bution (e.g., the labeling of women, children, primitives, and lower classes the ca&${te player she and her sisters often secretly listened to when he
as emotional "to justify the exclusion of these individuals from positions was u*fri. The kids put my suitcases in the trunk of his new Mercedes,
of power and responsibility" lLutz 1986:2941) or by looking at the actual I said my goodbyes, and we set off. As soon as we were on the desert
social contexts in which emotion discourses are deployed. To study the highway, he turned on the tape deck and said that I had to listen to this
ways emotion discourses are used, to focus on practice rather than tape. We listened. A man chanted, in a moving and pained voice, Poem
meaning, and to examine discourses rather than their putative referents after poem of the type called the ghinndwa.
are, it seems to me, the projects shared by the contributors to this vol- My host listened raptly, interiected exclamations of sympathy at the
ume.7 end of some of the poems, and elaborated, with intense and obvious
I focus, in this chapter, on the emotions or sentiments associ#ted with admiration, on some of the references in the poems. Among the poems
relations between men and women in a Bedouin community in Egypt's were the following two:
Western Desert, especially the discourses of "lo.ve." My argument pro-
ceeds by way of a Bedouin love story. The community of Bedouins about Patience is hard
which I write are part of a group known as the Awlad 'Ali who inhabit for my heart, so freshly wounded . . .

the area along the Mediterranean coast of Egypt west of Alexandria


into Libya.s I lived in this community froml978 to 1980, visited once for wa'r'al€h iq-gabr
jarl.ra jdid mazal khatn . . .
a month five years later, and went back for five months of fieldwork in
1987. Until about thirty-five years ago, those still living in the Western
I'd figured, oh beloved, that distance
Desert made a living mostly by herding sheep, growing barley, and or-
would be a cure but it only made it worse '
ganizing camel caravans to transport dates from the oases to the Nile
Valley. Now they^are involved in all sorts of activities, from the old one ni[rsab ya'aziz il-m6h
of raising sheep tdthe newer ones of tending orchards, smuggling, sup- yabgd li dwd nddh zidni .. .
plying construction materials, and speculating in real estate. They used
to live in tents. Now most of them live in houses, although they still My host explained that Fathalla, the young man reciting the poems
pitch their tents next to their houses and prefer sitting in the tents at (whose kinship relation to one of our neighbors he identified) had been
least during the day. They used to ride on horses, camels, and donkeys; in love with his paternal cousin and wanted to marqr her. Their fathers
now they prefer Toyota trucks. Although sedentarizing, they still proudly had first agreed to it but then got into an argument with each other. The
distinguish themselves from the settled peasant and urban groups of the young woman's father decided to refuse to give his daughter to the young
Nile Valley - the Egyptians - by their tribal organization and what they man. In despair and thinking that he might get over this more easily if

E
;"

30 Lila Abu-Lughod Shifting politics in Bedouin lozte poetry 31

he put distance between them, the man set off for Libya (where until feelings individuals expressed in their poems and the ones they ex-
recently many Bedouin men went looking for work). Some time after- pressed in their ordinary language communications about the same sit-
ward, the girl's father arranged to marry his daughter to someone else. uations.e This is not apparent from Fathalla's case, because his poems
When Fathalla heard the news, he composed and recorded these poems were out of context. But it is obvious from a more typical case I have
and sent the tape to the girl's brother, a cousin with whom he had grown written about, that of a middle-aged woman I called Safiyya.
up. Fifteen days after the wedding, when the bride came back to her When she told me about her divorce from the man she'd been married
family's household for the ritual postmarital visit, her brother played her to for almost twenty years, she showed the aggressive nonchalance I,d
the tape. She listened to it and when it was over, she gasped for air, come to expect as the typical attitude toward love and marriage. She
faintdland then fell over, dead. explained that she had never liked him and didn,t care when he di-
Thts story tells us a great deal about the politics of emotion discourse vorced her. But two days later, when a conversation between Safiyya
in Bedouin society. I will take up four aspects of it. Ffust, there is the and seqeral other women in her household turned to the whereabouts
matter of the poignant love poems and their relationship to other Be- tuf herq{rhusband, away on a trip at the time, she suddenly recited a
douin discourses on "love." Second, there is the import of the reactions poem tHht everyone knew was about her husband. In it she expressed a
to the young man's poetry and its apparent power. Third, there is the very different set of sentiments, especially her sadness at losing him.
fact of the cassette. In the conclusion, I will consider a fourth issue: the' 'l'his and several other poems she recited suggested a feeling of attach-
context of the particular telling of this story to the anthropologist. I nrent to her husband and were recognized as doing so by all the women
When Fathalla wanted to express the sentiments of love, he did so who heard them.
the traditional and formulaic medium of poetry. The way he expressed Like Fathalla, Safiyya expressed the sentiments of love in poetry but
-
these feelings and the medium in which he expressed them could rhe denied those sentiments in her ordinary conversation. There she
said to be distinctively Bedouin. One could easily say that tfiEy expressed the more characteristic sentiments of everyday male-female
shaped by the culture in which he lived. Fathalla expressed his pa relations in this society - which involved the denial of concern about the
feelings in a genre of oral lyric poetry that is so much a part of husband. In general, male-female relations among the Bedouins are
life and" so cherished by the Bedouins that on my first field trip I murked by distance. sexual segregation characterizes daily lif'e. people
ended up studying it. Reminiscent of fapanese haiku in its length dt,ny interest in love or sexual matters and avoid members of the oppo-
condensation of imagery, but more like the blues in emotional tone, pltc sex except close relatives. An important goal of the socialization pro-
is the poetry of personal life. It is by no means the only kind of Grns is to teach children, especially girls, to do this. I heard a girl confide
the Bedouins recite. They have many distinct genres of long rh ln her uncle's new wife. "To tell you the truth, I don't even know what
verse, usually composed by men, often specialists, and recited at thlr love is. I hear about it in songs and hear about this one giving some
gatherings. They also have special types of wedding poems and guy her necklace, that one her ring, but I don't know what they are
But poems of this genre, the ghinndwa, can be composed and recited talking about." The older woman responded approvingly, "That,s my
anyone, male or ffmale. Although the poems can be changed or $r1," Even married women deny any interest in their husbands, not to
sometimes at weddings, and as you could tell from my story, now Rention other men. women rarely use their husbands' names, referring
into cassettes passed from person to person, most often as I heard ta them as 'that one' or, if affectionate, 'the old man'; if they are being
they were just recited in the middle of ordinary conversations with fBrmal, they refer to him politely as 'the master of my household., At
mates. in front of others, they are formal and distant with husbands,
The poetry is rich in sentiment, so one might be tempted to nB no public affection.
the language of poetry in order to get at the cultural construction By the same token, men do not spend much time with their wives and
Bedouin sentiment. But there would be a problem with this proiect, talk about them, They are ridiculed if they show too much con-
I began to pay attention to the way people ueed this form of , When one man'B new bride ran away and he sulked and seemed
discovered an intriguing pattern: There waa a discrepancy between hia rclativee all teaeed or ecolded him. His mother said, "l,ly'hat,
:

Shifting politics in Bedouin looe poetry 35


34 Lila Abu-Lughod
not have as much stake in the system - like young men and especially
gitimate guise of marriage, although necessary for the reproduction of
women - will help perpetuate it, because their virtue or their standing
society and the perpetuation of lineages, is hard to deal with. It does
as moral beings, as good persons, depends on denying their sexuality.
not rest easily within this framework for social relations and is in fact a
As I hinted earlier, these sentiments of sexual modesty are situational.
threat.
They are important to display only in front of certain people - the elder
Love and the bonds it might establish between individuals are not just
male agnates. So, sexual modesty must be seen as a form of deference
threats to the framework that orders social relations, but are also talked
to them. The moral sentiments of modesty are part of a discourse that
about as threats to the solidarity of the paternal kin group, something
sustains and perpetuates the particular social system and the power of
often poted in the literature on patricentered societies from traditional
r'i.r* certain groups within it.
Ch\p td Zinacantan (Collier 1974;Wolf 7972).The Awlad'Ali view sex-
Conversely, then, the immodest sentiments of "love" are subversive.
ual bonds and the bonds of agnation as competing. Even more impor-
To expgess them is subversive of the social order and defiant of those
tantly, sexual bonds are seen as threats to the authority and control of
whos-fu,ftterests are served by this order. This element of defiance is
elder male relatives who represent the interests of the agnatic family
made 6ncrete in the story of Fathalla. In singing about his feelings of
group, control its resources, and make its decisions. At marriage, sons
love, he was, in a sense, defying the authority of his paternal uncle, who
begin to have a small domain of authority of their own, and daughters
had thwarted his desires and prevented him from mar4ring his cousin.
leave the domain of authority of their father and kin.
Because it carries subversive sentiments of love, one could consider
The threat marriage represents is counteracted at every point by social
the ghinnawa the Bedouin discourse of defiance. There is plenty of evi-
and ideological strategies. The marital bond is undermined in numerous
dence that poetrv is in general associated with opposition to the ideals
ways: Wom'en retain close ties to their paternal relatives, senior male
relatives control the choice of marriage partners, and sexual se8$Egation
of normal social life. This type of poetry is considered un-Islamic. The
pious shouldn't recite it or show any interest in it. It is also considered
ensures that husbands and wives spend little time together. Divorce is
unrespectable. Even the term "to sing" can't be said in mixed-sex com-
easy and polygyny possible. And the married couple is rarely economi-
pany without causing all-around embarrassment. People say that they
cally independent. Love matches are actively discouraged. One man told
are ashamed or embarrassed about singing in front of nonintimates, es-
me thatthe only way people who loved each other would be allowed to
pecially elders. Women told me never to share their poems with the
marry was if their elder male relatives or the girl's paternal cousins did
men. And in the past, older men avoided public settings like weddings
not know. Women often told me that love matches always ended badly
and sheep-shearing parties where young men usually recited this type
for the woman because she would not have the support of her male kin
of poetry. The most persuasive evidence of the oppositional character of
if her husband mistreated her.
poetry is who recites poetry and who avoids it. Although older men
I have argued that the cultural preference for patrilateral parallel cousin
occasionally recite them, ghinniwas are most dosely associated with youths
marriage is another such strategy (Abu-Lughod 1986). The Awlad'Ali
and women. These are the disadvantaged dependents who have least
frequently marry first cousins or other cousins on the father's side, and
to gain in the system as structured.
the male even haS,a legal claim to his paternal uncle's daughter. This
type of marriage may be upheld as the cultural ideal, because it provides
a means of defusing the threat of the sexual bond in this social system; Dialectics of deference and defiance
it subsumes the marital bond under the prior and more legitimate bond '[he existence of defiant or subversive discourses by those not in power
of kinship.
is probably fairly common in the world. We must take care not to ro-
The moral code that prescribes modesty is the most effective means of
manticize this rebellion by taking it out of the context of social and polit-
undermining the sexual bond. If the threat to the social order can be
ical relations in particular societies. This brings me back to my second
made to seem a threat to the respectability or moral worth of the individ-
general point: We need to consider Bedouin reactions to this poetry.
ual, then that order will be reproduced by the actions of individuals in
What may be specific to the Awlad 'Ali Bedouins is that even though
everyday life. The modesty code engures that even individuals who do

b
..

36 Lila Abu-Lughod Shifting politics in Bedouin loae poetry 37

this resistant emotional discourse goes against the system and makes his uncle's authority and ended up undermining the old man's control
the groups with power nervous, it is both culturally elaborated and pos- over the lives of his daughter and nephew and thwarting his attempt to
itively sanctioned. Poetry is a highly developed art and the Awlad 'Ali deny them what they wanted. Others, including my host, who was an
cherish and privilege it, in certain contexts listening intently to poems, older head of a lineage and a paternal uncle to many, were awed by the
memorizing, repeating, and being moved by them. Most of the adults I fatal power of this poetry. Partly this was because they recognized that
lived with suggested that poetry was the best thing they had to offer as the uncle had abused his authority, but partly it was because, in Be-
a cultural group. They thought of poetry as distinctively Bedouin, asso- douin eyes, the legitimacy of authority is always in question. Resistance
ciated with their noble past when they were politically autonomous, has positive valence, and Fathalla's love poetry was a kind of resistance.
toughr$rd independent. Of course, it is sadly telling that all Fathalla could do with his poems
Sirifrlarly, those who recite poetry, expressing those sentiments that was to thwart his uncle. He was not able to get what he wanted. The
challenge the social system and the authority of elders, are not just tol- storyls4hagic end - the death of the woman Fathalla loved - suggests
erated or not disapproved of but actually admired. This is apparent in the uffite power of the system and the futility of resistance. Perhaps
my host's reactions to Fathalla. He and the many who wanted to hear that is that made the tale so compelling and poignant. It may have
this tape over and over clearly admired this young man for his passion captured people's imaginations because it so vividly portrayed the com-
and for his ability to express it in poetry. They were moved by his poems plex relationship between love poetry and power in Bedouin society.
and awed by fhe power of his words.
I have argued that this ambivalence about love poetry - the discomfort
surrounding it on the one hand, and its glorification on the other - re-
The cassette
flects a fundamental tension in the organization of Bedouin social and
political life. It may be related to an uneasy recognition of the rffiy that l'he third aspect of Fathalla's story I want to take up is the somewhat
the system of hierarchy within the lineage and family, the one to which rurprising fact that his love songs were on a cassette. I had thought,
the sentiments of deference apply, violates the tenets of tribal politics, when I left Egypt after my first period of fieldwork in 1980, that the
where ttp paramount ideals are autonomy and equality. Day-to-day pol- tledouin ghinnawa was dying out. The adolescents I knew did not sing
itics, however, puts in the hands of elder male agnates control over re- ttr recite this type of poetry, nor did they seem particularly interested in
sources and power over dependents. This domination contravenes the It. They were beginning to listen to Egyptian radio, and it was from their
ideals that sustain the wider tribal system. It may be rationalized through nlothers, aunts, and grandmothers, and sometimes from their fathers
the elaboration of a moral code that justifies the privilege of elders and end a few young men, that I collected poetry. These adults offered one
dignifies the deference of dependents. But it is a contradiction. explanation for why poetry was dying out: They said that there were no
Love poetry, as a discourse of defiance, is seen as a discourse of au- ktnger any occasions for singing. There is a certain truth to their decep-
tonomy and freedom. Recited mostly by those slighted in the system - tively straightforward explanation to which I will return.
that is, women ary! young men - it is exalted because a refusal to be lf, however, it was the ideology of the political system, with its value
dominated is key tb tfreir tribal political ideology and so a key value, of nutonomy, that lent positive valence to expressions of love as defi-
even for these individuals. To love, or to express the sentiments of love, Ence, even when they came from below, then one would not be sur-
then, also signifies one's freedom. But to talk about sentiments or the prlae'd to find such discourses dying as the Awlad'Ali Bedouins' politi-
discourses on sentiments as signifying something - in this example, eal autonomy was undermined. This has been going on for quite a while
freedom - is misleading in that it suggests something too static and too ea the Egyptian state has sought, over the last 35 years, to introduce its
idealist. Love poetry as a discourse of rebellion is used to assert thie tuthority into the Western Desert, a process that has been underway in
freedom and is credited by others with tremendous power. Eedouin areas closer to the Nile Valley for 150 years or more. The Awlad
This is the other sense in which sentiments should be seen as political
'All have developed an imPressive array of strategies to resist,
subvert,
and reciting love poetry as a political act.l0 Fathalla's poems challenged lnd clrcumvent the authority of the etate, which they consider illegiti-

h :;. ffEE
.

38 Lila Abu-Lughod Shifting politics in Bedouin looe poetry 39

mate. Since they are not fazed by guns, and prison sentences carry no imprisoned. The domestic political divide now runs along gender lines,
stigma, it is even hard to intimidate them into good citizenship. whereas it used to be between elders on one side and women and young
There is, however, one process that began in their region in the 1970s men on the other. All men have access to the market and increasing
that, more effectively than government efforts to disarm them, school freedom of movement; all women do not.
them, put them in the military, license and register them, is progres- This shift in the political economy has implications for traditional love
sively undermining their resistance to the state: the gradual shift in their poems, which, as I have discovered, are not, after all, dying out' Be-
economic life from herding and commercial activities, including smug- douin love songs are taking on different meaning and force, having been
gling, to investment in land. Land reclamation efforts by the govem- given new life by the advent of the cassette. The Bedouins had said that
ment #stransforming some of their desert into agricultural land, which songs were dying out because there were no occasions for singing' In a
theyhlc increasingly buying and relying on for a livelihood, if not ac- sense, they were right. By the time I first met them, in 1978, they were
tually farming themselves. Land along the coast, on the other hand, has reciting$ve poems only in intimate social situations. As I later learned,
become valuable for tourist development, and many of them are doing howevffihe most important forum for love songs had been weddings,
quite well selling beachfront property. They are also fighting with each at which]roung men and women had sung within earshot of and some-
other over this land, which was formerly tribally held rather than indi- times to each other.l3 Those kinds of celebrations had stopped by then,
vidually owned. With this involvement in land, the Bedouins have be- and the weddings I attended were sexually segregated. The women sang
come enmeshed in the state's legal system, since they need to get titles only songs of blessing, congratulation, and praise, and the men did
and make claims through it.1l nothing but sit around. Today, locally made cassettes - copied, re-
This shift in the Bedouin political economy can be connected with copied, and sometimes sold for money - provide a new occasion for
what I see as a shift in the dialectic of deference and defiance in which song, as does a new kind of wedding celebration coming into fashion.
love songs are deployed. As the economic basis of the tribal systerrftrodes, At this new wedding, attended by invited guests but also attracting a
and with it the political underpinnings of the value of autonomy, the growing group of somewhat rowdy young men, the small-time stars of
older reality of mutual responsibility within the family and lineage is these low-budget commercial cassettes perform. Because of the public
changing. There used to be a complex division of labor, with resources nature of these occasions, where, unlike in the past, "public" includes a
managed by elders but not owned. Now private ownership puts tre- wide range of nonkin and complete strangers, women are absent. They
mendous control in the hands of patriarchs. Young men suffer, as I will 1re also, out of modesty, absent from the recording sessions where tapes
discuss, but those most dramatically affected are women. They are now rlre made. They make no tapes and no longer sing in public. No longer
economically dependent on men, having little access to money, and their having as much social and political support for defiance, the women also
work is increasingly confined to housework. With the moral value of scem to be losing one of the means for it - love poetry.
modesty still in force, these women who live in the new circumstances The poems sung on cassettes and at mtkrofAn weddings seem now to
of sedentarized communities, where they are surrounded by neighbors he part of a discourse of defiance by young men against the more abso-
most of whom argnonkin, must be more secluded, more often veile.d, Iute authority and economic control of their fathers and paternal uncles.
and less free to move around. 'l'his is a period when, at the same time, young men are beginning to
Older women comment on these changes, reminiscing about things have more possibilities for independence from the kin group through
they used to do that young women today cannot get away with - like wage labor and more knowledge than their fathers about the ways of
having rendezvous with sweethearts and exchanging songs with men at the state through their experiences in the army and school. A new sort
weddings and sheep shearings.l2 But they are also convinced that they of generational conflict produced by these transitional circumstances is
were more modest, a perception that I think relates to a sense that it was treing played out partly in the language of love.
more self-imposed. They often complain that their sons, husbands, I had unexpected confirmation of the new use of "love" on my visit
nephews, and grandsons harshly restrict the girls, not letting them go ln January 1987, when I was listening with friends to one of the latest
anywhere. Girls, for their part, are beginnlng to complain that they feel casoettes of popular Bedouin songs of a different genre. There was a

E ,j:#
:,

40 Lila Abu-Lughod Shifting politics in Bedouin loae poetry 4L

long and somewhat humorous song about the tribulations of a young animated our radio therapist's project.ls These assumptions keep mak-
man whose father and uncle had arranged three terrible marriages for ing it difficult to see how, for us, emotions serye as tokens in the con-
him with women he'd never met. The first woman turned out to be bald, struction of our subjectivity, how they bolster our belief in the truth of
the second dumb, and the third insane and violent to boot. In the final our individuality, and how all of this might also be political and specific
verse of his song the poet, speaking on behalf of all young men who to our place and time - that is, something worth malynng critically rather
have suffered the tyranny of such fathers and uncles, sings:14 than universalizing. 16
If instead we take discourses as the object of analysis, we can get at
My warnings are to the old man
something more interesting. I am not making a narrow plea for sociolin-
who imprisons the freedom of youths
guistics or the ethnography of communication, although they are also
i&o's forgotten a thing called love involved. "Discourse" is a concept that recognizes that what people say,
affection, desire, and burning flames
who's forgotten how strong is the fire of lovers Benerou*ly defined (which is, after all, what anyone is dealing with in
the anffiPological study of emotions), is inseparable from and inter-
how strong the fire of lovers who long for one another
penetrat€d with changing power relations in social life. There is a double
What's exquisite is that they're afraid
movement implied in this notion. First, social and political life is to be
they say, any minute my prying guard will turn up
!{een as the product of interactions among individuals whose practices
oh my father's about to catch us
rrre informed by available discoursesl second, language and culture are
The relationship between love and freedom in this song is complex - understood pragmatically rather than referentially. They are understood
because, although he does not want the elders to force loveless mar- as part of social and political life. Analyzing emotion discourses as dis-
riages on their children, the poet recognizes that what makes love ex. courses rather than as data for our own "scientific" discourses on emo-
quisite is that it is stolen - it is against the authority of elder a$ftates. In tion provides us with a technique for avoiding the false attribution of
other words, he wants the freedom to defy the elders abit, a freedom the project of psychologizing to others as it reminds us relentlessly of
he reminds them they used to want, but he does not reject the system the social nature of emotional expression.lT
as a whple or want to have love easy or open.
The continuity of form in love songs is consistent with this attitude.
A discourse redeployed
Unlike rock and roll, which some would argue played and plays a simi-.
lar role in our society, the protests occur in an idiom that the elders can If any further evidence need be offered for the critical importance of
appreciate: the poetry they themselves love and must respect, given their rt taining a sense of the always social character of emotion discourse,
own values. This is true even though they disapprove of the young men's eunsider the fourth and final aspect of this Bedouin love story: the con-
bare heads, occasional long hair, experimentation with drugs and li- text in which it was told. Fathalla's story was told to me, as I recounted,
quor, and general loss of fuasham. Everyone comments on this now - the [ry *y host, the man whose household I had lived in for two years. He
new brand of yorffiB men aren't modest in front of their fathers. Accord' played me the tape of those poignant love songs as I was about to depart
ing to the girls I iilk with, in front of their fathers these young men not 6gain for the United States. I had been absent for five years the first
only smoke cigarettes but, worse, they shamelessly play love songs on tlme, and they did not know when I would next return. I promised it
their cassette recorders. would be soon.18 Although my host and his family had begged me to
This is only a partial analysis of the shifting politics of Bedouin dis. Etay, and my host had gone as far as to offer to set up a job for me
courses on love, a complex subject on which I do not want to impose a directing a private school he would finance, I insisted that my life was
false coherence. Yet it should be sufficient to make clear my larger ana. ln nmrika and that it was not likely that I would come to live permanently
lytical point regarding the anthropology of emotion. As long as emo. ln Egypt. When he played this tape for me and told me its sad tale, he
tions remain the object of study, we can break with neither the idealism wa$ not interested in explicating Bedouin emotion concepts or in under-
and mentalism of the interpretive approach nor the assumptions that etnnding himself, but rather in impressing on me the force of poetry.

h
;

42 Lila Abu-Lughod Shifting politics in Bedouin loae poetry 43

Wasn't he, in a way, using the force of poetry on me? Of course, he (1983), Riesman (1977), Lindholm (1982), and others as examples of this
knew I was writing a book about poetry, and we often discussed poems. type of work.
5. tio-tn nuck Schieffelin and Ward Keeler, in personal communications, have
Was there more? Did he wish to move me, to resist my departure by voiced these sorts of worries even about their own exemplary work'
these songs and by telling me what effect they had had on another 6. I am paraphrasing Foucault on discourse (1972:218,229).
woman? 7. I am using "discourse," that admittedly slippery and overused tetm, not

I sensed that this may have been part of his intent when, two years iust as lin-guists do, to refer to the speech of individuals, but also in the
'Foucaultia"n sense of a range of culturally available and historically specific
later, as I went over Fathalla's love poems with my host's wife, a woman statements. See our introduction to this volume'
I always talked with about poetry and who was good at explaining poems, 8. For a fuller discussion of the Awlad'Ali, see Abu-Lughod (1986).
9. Much of the argument in this section was developed in Abu-Lughod (1985).
I heaf*something surprising. She knew the poems and knew Fathalla's
10. In other wordi, to have certain feelings or at least to exPress particular
stoff-but said she had not heard that the girl had died. In fact, she was sentiments becomes a political statement, if not a political act. Although in
fairly certain that she was alive and living with the husband of her ar- the.Story of Fathalla tlie poems seem to have the Power to kill, Bedouins
ranged marriage. usffi$just say that it.ttorr"t people or causes them-to change their actions.
Tha!i'.s, they see it as persuasive.
This incident can serve as a reminder that the emotional discourses When I wai there doing fieldwork most recently (rtr 1987), every Bedouin I
we might want to use for our anthropological discourse on emotion are knew seemed to have ilawyer. They still tried to resolve disputes through
hardly inert. They may indeed have a cultural context, but the more their own tribal mediation system, and my host, who is a mediator, was
strained to the limit with th6 number of tense cases he was asked to help
important thing about them is that they participate in social projects - resolve. What was happening, however, was that they had to work both
whether the larger ones of generational contests over power in an erod= through their own system and through the courts.
ing tribal system or the local and particular ones of a conversation be- t2. They Iak about an institution called the miilds, in which a young unrnarried
tween a Bedouin man and a youngish female anthropologist driving to woman would entertain all the eligible young men in a tent, challenging
them to respond to her songs.
Cairo in a Mercedes. t3. Like the disiolution of the discourse of women's weaving in North Africa,
argued by Messick (1987) to be related to the capitalist tran-sformation of
do--mestic-weaving, with the disappearance of -one occasion for song have
died the songs appropriate to it among the Awlad 'Ali. Sheep shearing-s,
Notes ' which used t6 be-occaiions for groups of young men to go from household
to household shearing the sheep, no longer occur, as professionals, mostly
This paper was completed while I was a member of the Institute for Advanceil from Sinai, have taken over thii work. The songs that accompanied sheep
Study, a unique institution to which I am grateful for many things, including shearing were more explicitly sexual than the ghinnawa, couching in innu-
support, through the National Endowment for the Humanities, for writing. Eail endo th-eir references to relations between men and women. These are no
lier versions were presented at the Anthropology Department Colloquia at J
longer heard, and I heard of no equivalently sexual geryg-s.
York University and the City University of New York Graduate Center, w The Arabic original, as sung by'Awadh al-Mdlki, is as follows:
questions from the audience helped sharpen my arguments. Several Several pe<
peoplei' nqrltra minni lish-shayib
including Timothy Mitchell, Catherine Lutz, and Buck Schieffelin, carefully read il[ habis hurnt ish-shab
and commented on drafts, and their suggestions, sometimes taken, sometimed wnasi [raja ismha trub
not, are gratefully acknowledged. As always, my greatest debt is to the Bedouiq w'aff wshog wndr thib
families in Egypt ttho let me participate in their lives and learn from ttrem. ya magwa n6r il-ghawi
Funding frori'NEH, Williams eolege, and the Fulbright Commission enabled ya magwa nir il-'ajgin
me to spend more time with them in Egypt in 1985, 1986, and1987. i illi ba'dhun mishtagn
1. For a comprehensive review of this literature, see Lutz and White (1986). r simatritha yagbo khayfin
2. This ethnocentrism is expressed beautifully in the title of one of Lutz's (1985) yg0l in-nagir sa'a yji
articles, "Ethnopsychology Compared to What? vsul in-nasir si'a vtug
3. For extensive references to this work, see Lutz and White (1986:417-20\t 15, i"?; ffiil;or tn,J iJ8uutm implicit in interpretive anthropology, see Asad
Also, for a wonderfully insightful and playful reflection on some of the (1e83).
problems shared by philosophers and anthropologists engaged in the cul. 16, Lutz (tgAO) makes some of these points, but I think that three further sets
tural translation of feelings, see Rorty's (1979:70-127) discussion of the An- of questions about the Euro-American emotion concePts she outlines need
tipodeans (persons without minds). to be researched. First, which cultural concepts are most salient, and does
4. They cite my work, that of Myers (1979, 1986), Appadurai (1985), Keeler thln pattern differ by subcultures? Second, when do certain ways of think-

h ,sEilE=E
:

50 Geoffrey M. White Moral discourse and the rhetoric of emotions 51.

In one of the recorded meetings, a man daimed to be 'angry' with his ing'angry' events to fit the avowed aims of disentangling while at the
sister's son for stealing betel nut from one of his trees. But because he same time promoting a particular point of view about those events. The
simply expressed 'ange{, without acknowledging either the asymmetry cffectiveness of emotion talk in accomplishing the pragmatic work of
or the close relation of kinship between himself and his nephew, others disentangling is enhanced by the fact that it is both indirect and ambig-
saw his talk of anger as inappropriate. Two other participants re- uous.
sponded not to the act of stealing but to the uncle's stated 'anger' toward As a moral idiom, emotion talk is indirect because it relies on presuP-
his maternal nephew. One of the listeners spoke to the uncle, saying: position and implication to state interpretations and evaluations of con-
tested events. The quality of indirection is particularly important in small-
This kind of talk that you are making is as if you are all separate scale, face-to-face communities such as that considered here, where overt
1^isiasopa\ people. It is better to speak to your nephew to teach him, public statements about others' behavior are often proscribed. The rhet-
like I do with my nephew. . . That kind of l'ar.9ry'7 talk can be
oric of qmotions offers a remedy. Relying on shared models of socio-
aimed at other people, but to our own nephews, our own children,
emotiqffirocess, speakers may pose and counterpose statements about
it is very bad.
their ovtfrr actions and emotions, leaving moral assertions about others
The speaker's reprimand mirrors that of the first example. Both focus unstated. The availability of an idiom with which to assert moral claims
on the social context of expressed 'anger': one between adult and childn indirectly while simultaneously acknowledging the overt agenda of re-
the other be'tween mother's brother and sister's son. This second ex- cstablishing solidarity is essential to successful disentangling.
ample is more explicit about the primacy of the social comPonent ("['an Related to the indirect quality of emotion talk is the fact that emotions
ger'] can be aimed at other people"). Given these assumptions about nre susceptible to multiple moral readings. The potential for ambiguity
emotion and morality, it can be seen that the experience of anger arises from the complexity of conceptual and situational factors con-
a dilemma here: 'Anger' expressed in the context of commu tributing to emotional meaning. A specific emotion attribution may have
is likely to be regarded as contrary to the ideals of solidarity, w & range of potential entailments, and just which implications are brought
suppressed 'anger,' according to the fok theory of misfortune, Poses tu bear in a given interpretation may remain ambiguous. The following
dangerJo the self and others (see White 1985a for more on this). enalysis pursues this point by examining the use of a particular emotion
As an arena for the sanctioned expression of emotion, disen t,tm, di'a nagnafa ('sad'), in disentangling discourse.s This concept is
would appear to be a culturally constituted solution to the dilemma ct'ntral to the disentangling activity because it is seen to follow from
suppressed 'anger'. Yet, as the preceding examples indicate, this is trnnsgressions that threaten interpersonal relations, but at the same time
as straightforward as it might seem. In disentangling as in other glves rise to attempts to repair those relations. Represented schemati-
texts, talk of 'anger' in close relations is likely to be disapproved' enlly, basic elements of the di'a nagnafa scenario may be sketched as fol-
disentangling context does not legitimate expressions of hostility t Iows:
are not tempered by ideals of solidarity and the goal of repairing
DAMAGE CLOSE RELATIONS (Other, Self)
aged relations. !f,hat it does provide is an opportunity for the crea
J
of a social reality'-in which 'angry' events are rhetorically transforr
di'a nagnafa ('sad')
and damaged relations symbolically repaired.
Talk of 'anger' is somewhat out of place in disentangling ',
REPAIR (Self, Other)
defined as they are as occasions for reestablishing community
rather than uncovering transgressions or imposing sanctions. The An lmportant source of moral ambiguity in talk of 'sadness' derives from
amples cited account for most of the 'anger' attributions in the the potential for focusing variously on the antecedent (transgression) or
meetings. Instead, one finds numerous examples of talk about' the implied response (attempt at repair). The former may constitute a
(di'anagnafa) and'shame' (mamaja). The following analysis suggests ehallenge to the other, whereas the latter emphasizes reconciliation and
these attributions are particularly well suited to the work of rclatedness.6

t ,gEH
Moral discourse and the rhetoric of emotions 53
52 GeoffreY M. White
in talk of importantly for this chapter, contemporary moral ideals are also cast
The following analysis shows that interpretive ambiguity largely in terms of Christian ideology.
real-
emotions facilitates the rhetorical transformation of socioemotional A'ara-speaking people sporadically (and decreasingly; White in press)
Specifically'
ity - a transformation central to the task of disentangling' cngage in a practice known as graurutha, or 'disentangling' (trom rutha,
reformulate those events so as
narrative constructions of past events 'undo' or 'untie'). Put briefly, disentangling is an activity in which fam:
totransmutethreateningandconflict-ridden'antger'rhetoricallyto ily members or village mates meet together to talk about interpersonal
'rid"u"''' to
Both 'anger' and 'sadness' pertain
solidarity-engendering conflicts and 'bad feelings' (di'a nagnafa). The point of this talk is to make
thesortsofproblematiceventsinwhichthetransgressionsofothers bad feelings public so as to defuse their destructive potential. The A'ara
in
i-pi.g" o., th" self. Disentangling institutionalizes a speech event
':l'1": believe, as do people in many cultures (e.g., Harris 1978;lto 1985; Strathem
in terms of 'sadness"
*t i"f,ip"ut"rs reframe 'angry tesponses 1968) that negative emotions that remain hidden may cause illness and
are narrated so as to
and related emotions. tn doing so, conflict events
highlight valued interpersonairelations and community :'ll1lY:ll: ,misfortune - ranging from personal injury to a poor catch of fish or fail-
basis of rhetorical qfore to locate domestic pigs in the forest. Furthermore, bad feelings are
analysis also raises questions about the experiential only potentially damaging to the self, they may harm others as well,
transformations accomplished in discourse' It is therefore in the interest of the community as a whole to repair social
discord and maintain emotional harmony. Disentangling is an institu-
The conceptualization of 'disentangling' tionalized means for achieving those ends. It provides an occasion in
which people are encouraged to talk about conflicts and resentments
SantalsabelisoneoffivemajorislandsintheSolomonlslands,witha that would otherwise be difficult and even proscribed topics of conver-
present population of more than 14,000' Half of the populafi:1:P,,"^?L:: xntion.
Cheke Holo (Whitel
Austronesian language known variously as A'ara or The significance of disentangling depends on a presupposed world of
in subsis
Kokhonigita, and Pio-unu 1988)' People engage p$marily understandings about persons and emotions. As already stated, a cen-
schemes to prodr
tence gariening while pursuing various agricultural
it ttre turn of the century' sociopoliti tral tenet of this world is that hidden 'bad feelings' arising from inter-
cash crops. Prior to pcrsonal conflict may cause illness and misfortune for the self and oth-
* "hur,g",
activity was organized larg:ely by relations of descent
and regional ali
the focal point t rs. The operative word here is 'hidden'. By implication, if negative
ments in which local big *"" ot'chiefs' (funei) were emotions can be 'talked out' and made public, they lose their potential
The idiom of descent is distinctly m
intergroup feasting anJraiding' fur causing harm. But this 'talking out' is not conceived of as catharsis
of grouP identi
trilinial. I, for*", times, the territorial locus descent
ln the sense of reexperiencing or releasing intense emotions "bottled
were made to a
was marked by shrines where propitiatory offerings up" (my metaphor) in the person, as would be expected, for example,
cestral spirits. ln the American model of anger (Lakoff and Kovecses L987). To the con-
cultural tra
Although economic changes have been slow in coming' trary, disentangling talk is conducted in a narrative mode of reporting
influence, especially Christianity
formationls associated with Western ttrt past events rather than enacting demonstrative displays of felt emo-
havebdenmoredramatic(White1988).Historically,themostsignifican tlon. For A'ara speakers, public talk about conflict in this context is itself
(now the Chu
agent of change was the Anglican Melanesian Mission att act of reconciliation. Individuated 'ange{, hidden from the restrain-
island po
oiMelanesia), which completed its work of converting-the lng moral judgments of the community, rrta! t in the local view, give rise
people' who had
lation just after the tum ofthe century' The Isabel kr a broad range of destructive behaviors (White 1985a, 1985b). By en-
from the western
,"rr"ruty victimized by marauding headhunters gaging in disentangling, speakers tacitly acknowledge the prior value of
of the nine
mon Isiands, eagerly received the mission at the end eummunity solidarity, ritually closing the conflict episode as a source of
century. Conversion entailed majorshifts in residential
putt"Tt
the mountainous
I
interior dlrruptive thou6hts and actions.
people formerly scattered throughout 'l'his interpretation of the local conceptualization of disentangling ie
grr*a to coastal villages of unprecedented size 1100-'-90 P":l]-:\ 1
evldcnt in a range of conventional metaphorc uced to talk about the
i-*rr rir" was (and .oi.,tir,r", to be) centered on the village chapel.

:
:,
..

54 Geoffrey M. White Moral discourse and the rhetoric of emotions 55

activity. These metaphors suggest that disentangling is thought abo gling gives a more complex picture of the multiple social and emotional
as a process of reoelation, of moving personal thoughts and emotio. realities created in context.
into community awareness. By examining a range of interrelated li
guistic expressions, it is possible to identify key images used to conce
The practice of disentangling
tualize the rationale of disentangling.
Presuming an image of the person as a container, two commonly [,ocal theories of misfortune reflect the indigenous rationale for disen-
metaphorical expressions describe disentangling as either'talking tangling as a desired and even necessary activity (White 1985a). These
(cheke fajifla) bad feelings or as 'opening up' (tora) so that thoughts a understandings (among others) constitute an intersubjective basis for its
feelihgs do not remain hidden. In the former expression, problema loint enactment. To some extent, a disentangling session achieves its
fedllngs are described as moving from inside to outside the person purpose simply by taking place. By virtue of their coparticipation, Peo-
that they become visible or known to the community. The directi plc signq,l their mutual commitment to repairing community related-
adverb fajifla ('out') signifies physical movement across,a boundary, ntss. I{foielrer, beyond the simple fact of its performance/ a disentan-
in leaving a room or removing an. object from a container. The gling se#ion creates a communicative environment in which a certain
metaphor also builds on a sense of revelation - of becoming known type of discourse is possible - specifically, narrative reconstructions of
virtue of coming into view. The verb 'open up', often used to problematic events.
opening a door or lifting the lid to a box, describes the sort of talk t Consider some of the ways in which the social organization of disen-
'opens up' persons in the context of disentangling. tengling supports its communicative goals and facilitates the transfor-
Metaphors of sight and visibility are further elaborated in a variety firtrtion of 'artgry' emotions. To begin with, the narrative mode is well
expressions used in this context. For example, disentangling ma :uited to the avoidance of 'anger' in the disentangling session itself' For
thoughts and feelings public by'bringing them to the surface$just as an event that has as its goal the alleviation of hostility, it would not do
turtle is spotted when it 'surfaces' (thagrn) for air. And, by to provoke further expressions of resentment. And, given that partici-
covering layers, disentangling reveals feelings that have been Pnnts are expected to talk about interpersonal conflicts,
the avoidance
al hostility is a difficult problem. The organization of disentangling as a
ffrunlgr'covered ovef (plohmo) by problems and reluctance to talk
them. iequence of narratives minimizes or suPPresses the kind of quick chal-
Metaphors of concealment and revelation represent a local a lenge-and-riposte that typically leads to confrontation and argumenta-
of the opposition of "public" and "private." Disentangling tfurn. The social genius of disentangling is that the conversational orga-
that the person is a bounded locus of experience that may or may not fllzation of the event is managed through the joint participation of all
communicated in discourse. The expressions mentioned previously Prr!$ent. No designated person or group
leads the meeting. Rather, var-
indicative of the local view that disentangling transforms itrus participants take turns urging one another to speak and, where
thoughts and feelings by placing them in a wider field of neccssary, jointly direct conversation away from problematic topics or
circulation. Pardcipants in disentangling invoke images of 'talking o furnrats (White in press).
and 'opening uflwith talk' to remind one another that iust by talking Uven the physical setting of disentangling works to defuse confronta-
a certain way, one may fulfill the overt agenda of disentangling, of Honal modes of interaction. The two meetings I attbnded were both held
ing thoughts and feelings public. Disentangling performances are It night in a large house with people gathered in the shadows both in-
orly seen as efficacious, they are efficacious in transforming that w Elde and outside. (The nighttime format may be more characteristic of
is 'hidden', 'inside', or'below the surface' to that which is'visible', ' hrger, village-level meetings than of smaller, family disentangling ses-
side', and 'above the surface'. However, as anyone familiar with !lona.) Not only would the protagonists not address their speech to one
family life of small communities might expect, there is more to disen lnother, they would likely not even see one another. Narratives are Pro-
gling than the overt agenda. A closer analysis of the practice of Cueed for a largely unseen audience, And yet, the unseen coterie of
:"
:

56 Geoffrey M. White Moral discourse and the rhetoric of emotions 57

listeners is essential to the successful enactment of disentangling (cf, range across varying levels of specificity in their narrative, thus leaving
Brenneis, this volurne). Furthermore, disentangling participants are room for differing degrees of ambiguity in their moral assertions'
just passive listeners. They actively collaborate to evoke narrative per- The possibility of finding both 'sadness' and 'anger' as plausible re-
formances from those known to be involved in not-quite-resolved Bponses to the same events is the basis for their substitutability - a fea-
flicts, and to orchestrate the interaction so as to minimize the potenti ture of A'ara understandings that lies at the crux of disentangling dis-
for antagonistic confrontation. course. By relying upon generalued understandings about responses to
Each narrative produced in this environment is an opportunity for rule violations to talk about'sadness' rather than'anger', a speaker may
forming socioemotional reality. Given the avowed purPose of dise then draw out differing implications for his or her response - interpreted
ghngtit'talk out' bad feelings, it is not surprising that the narratives a in terms of a scenario of repair rather than retribution. Combining the
pufiEtuated with overt talk of emotions - talk dense with moral signi schemas outlined earlier, one may visualize the potential for attributing
cance. In light of the fact that each emotion concept entails specific 'Anger' a4d 'sadness' to the same eliciting event, with divergent impli-
terpersonal information (Levy 1984; Myers 1979), what sorts of cations type of reaction inferred.
{i$he
*
talk recur in disentangling sessions? In my limited sample of TRANSGRESSION
sessions,T three emotions emerge as particularly salient 'sadness', 'shame' r'\
and 'anger'.8 For reasons that should now be apparent, the last di'a tagna di'a nagnafa
'anger' , appears less frequently than the former two. This finding is ('angry') ('sad')
sistent with the hypothesis that disentangling is concerned with J J
ically transforming 'ange{ in relations where overt tensions are not RETRIBUTION REPAIR
ily voiced. The following discussion takes up this hypothesis in
detail. A case of disentangling
A review of several conflict narratives recorded in two separate m
ings indicates that statements to the effect that "Event x made me di' An example of narrative drawn from a village disentangling meeting
nagnaf*('sad')" occur frequently as speakers seek to draw attention illustrates the process by which emotion talk is used to create an emer-
the problematic nature of event x. In my analysis talk of 'sadness'in gcnt social reality (a transcript of the entire meeting is given in white in
context, like talk of 'anger', presupposes some sort of morally q press). The case summarized here was the second incident taken up in
tionable action - an action that has harmed the speaker or so a meeting conducted one evening as people met to clear away possible
identified with the speaker. At a generalized level of understandi unresolved conflicts that might interfere with preparations for an up-
both of these emotions (di'a nagnafa'sad' and di'a tagna'angry') may Cgming feast. (Hidden resentments might, for example, cause failure or
seen to derive from forms of transgression. It is at a more specific lniury to those going out pig hunting or turtle hunting.) The meeting,
of cultural knowledge that they are differentiated in terms of distincti ln fact, served two purposes. with the disentangling portion of the eve-
about particular4fypes of transgression. Although both 'sadness' ning finished, villagers turned to a discussion of preparations for the
'anger' are evoked by rule violations, 'sadness' pertains to social ac upcoming feast.
that damage relations, whereas 'ange{ is regarded as a response to Discussion of this case in the meeting consisted primarily of just one
tions that harm or threaten the self or significant others. These di I0ng narrative by one of the people principally involved in the incident
degrees of generality in cultural understandings of emotion reflect t (whom I refer to as Tom). Tom',s narrative is then followed by a series
hierarchical organization of knowledge such that models are nested of shorter statements. Several others who figure prominently in the story
models, with reasoning about specific emotions drawing uPon more wcre not present at the meeting (a fact that was noted prior to Tom's
understandings about kinship obligations and the social order generally apeaking, as participants in the meeting attempted to get all the parties.
(see Quinn 1987:189). The nesting of cultural models allows speakers to Ittvolved to speak before realizing that they were absent)'

t
58 Geoffrey M. White Moral discourse and the rhetoric of emotions 59

Tom's account focuses on the way he and others responded to father, Gata. In the course of his narrative, Tom twice emphasizes his
revealed transgressions of two young people (his younger brother a younger brother's responsibility for his own injury. However, following
a young woman) who had been having an endogamous extramari this, he shifts attention to Fala's rejection of his brother, twice quoting
affat- His discussion is not concerned directly with the illicit relatio her utterance "Don't set foot in Paka againl"
ship, but with the ways people reacted to it and the implications of It is the moral significance of this rejection that becomes the focus of
reactions for their relations with one another. In other words, his is Tom's narrative. He brings Fala's act into moral scrutiny by placing his
narrative of metaconflict - articulating conflicting views of conflict in relationship with Fala in the foreground of his account. The evaluative
community. implications of Fala's utterance are delineated most clearly by Tom's
W.ful the mother of the young woman learned what was going characterization of their relationship in terms of close kinship. He says
sh&pparently told the young man to "never set foot" in her that Fala is "like another mother" to him. (Tom and Fala are in fact
(Paka) again. The young man then went off in a rage and at some poi related_4 maternal half-cousins. Fala's older age is the basis for Tom
cut himself with his bush knife. Tom later encountered his younger referrinfufu-her as "mother.") The idioms of kinship and food sharing
in the bush and elicited his story about what had happened. After trre invol$d to assert close relations among those involved in the entan-
ing his brother back up to their village (Holo), Tom decided to return glement. These images emerge clearly in Tom's narrative as he describes
Paka to see the people there and retrieve his brother's bag. Along his encounter with Fala's husband, when he first heard about Fala's ex-
way, he encbuntered the young woman's father, Gata, and talked wit pulsion of his brother. The narrative establishes the incongruity be-
him about what had occurred prior to his brother's injury. The tween that act and the fact of kinlike relations:
told him that his wife, Fala, had told Tom's brother not to set foot
Then he (Fala's husband) told about [the statement] "Don't set
their village again.
foot again in Paka."
Tom then proceeded to Paka, where he found Fala's mothff, Su
and apparently told her that if that was how they were going to talk [So I said] "Who said that man?"
"Fala."
his brother, then she and her family should not come up to his vi
"Fala! So, alright, [she] is just like another mother to me' You
Holo, either. These mutual rejections had the potential to involve wi
married her, but she was already [just like] my mother. There is
circles of people, since the planned Christmas feast was to be held
water, there is food, there are sweet potatos there [for me to share].
Holo. Tom's encounter with Sukhi was followed by a period of
'Don't set foot in Paka!' [she said to] my younger brother! He did
tainty as to who, if anyone, from Paka would participate in prepara
do wrong before they started thinking like this, so alright," I said'
for the upcoming feast. As in many such incidents in small co
But my thoughts became all confused ('knotted') about that'
ties, an initial transgression reverberated throughout the social
threatening to ignite a wider conflict between others related to the Note that Tom uses a metaphor of entanglement ('knotted' haru) to ex-
tagonists. press his response to the moral conflict created by his brother's
Tom's narrativ6 chronologically reconstructs his involvement in tronsgression and his relatives' subsequent rejection of him. The image
events followinglnis brother's self-infl icted injury, making extensive of entanglement effectively comments on the fact of a moral dilemma
of reported speech to give the account immediacy and validity. He without committing to a specific interpretation pihpointing the a8ency
tuates his retelling of these events by describing his feelings during and the responsibility.
episode. It is the emotions of 'sadness' and 'shame' that emerge Having established his sonlike relation with the woman Fala, Tom
clearly in the account, serving to reinforce and amplify Tom's i FOes on to assert that her banishment
of his younger brother made him
tations. ntnmaia ('ashamed') with her and, most immediately, with her husband,
Tom begins speaking by distancing himself from the illicit affair who reported the act of exclusion to him. Although a thorough account
his brother's self-inflicted injury. He then proceeds to narrate the ,shame' ie not possible here, it typically signifies some kind of
6f A,ara
quence of events leading to his first encounter with the young woman' miematch between context and event, usually involving a violation of

h
Moral discourse and the rhetoric ot' emotions 61
60 Geoffrey M. White

appropriate social distance. In this instance, Tom's athibution of ' uhown here as an inference drawn from a set of premises. As noted
upon hearing of Fala's remarks implies that her act of rejection implic r.arlier, in many cases these inferences pertain to instantiating a gener-
misread or undervalued the relations that should have obtained in t nlized schema. In the conventions adopted here, a specific proposition
context. To underscore this point, Tom characterizes his relations nbout emotion is shown in quotation marks and the hypothesized back-
Fala in terms of potent symbols of food sharing, invoking a comm g,round schema is indicated by capital letters. The proposition implied
used scenario of entering someone's house and helping oneself to hy instantiating the schema is depicted below a horizontal line as fol-
food.e krws:
The juxtaposition of this characterization and the attribution of 'sha GENERAL SCHEMA
induced by Hla's reiection calls attention to her act as one that pla +
inappropri6te distance between kin relations, thereby creating the "attributed emotion"
to repair the moral damage. This implication is then further reaffi
inferredi ition
by the attribution of 'sadness' - just the emotion that would be
in response to a transgression that damages close relations. Following the ge quoted earlier, Tom continues his narrative by
Each emotion attribution may be seen as a partial "filling in" or rrcalling his thoughts and feelings at the time he encountered the young
stantiation of a general schema such as those discussed previously. wt)man's father and heard him tell about the expulsion of his younger
the emotion portion of a scenario is instantiated, the listener may lrrgther. He characterizes his feelings as a mixture of 'sadness'and'shame':
inferences about how the events leading to or following from that It was the "Don't set foot in Paka" that made me ashamed with
tion are to be interpreted. In other words, the listener fills in other him. This man [the young woman's father] is like our father, we
tions of the schema through inference, even though the speaker has two [Tom and his younger brother] would go inside [their] house,
been explicit about those aspects of the events' these sweet potatoes were our food, these houses, these beds . ' '
Much of the "actiotl" in emotion rhetoric involves the instantiati These feelings were probably with him [Tom's younger brother].
that results from giving an interpretation to precipitating events' Because of this I was sad.
when Tom aseerts that the banishment of his brother made him ' + SAD
DAMAGE CLOSE RELATION
he is implying, in this context, that the rejection is an instance of a
"Fala's remark --+ sad"
tain kind of social action - one that damages close relations, that is,
sort of thing that evokes 'sadness'. And, even more importantly for Fala's remark: damage close relationlo
portrayal of the incident, he implicitly claims that his response of I was just ashamed from those ways, from the way "Don't set foot
off the offending parties was an attempt at withdrawal rather than here." [So] I came to the old woman [Sukhi] up there.
kind of 'ungry' getting even.
'litrn's talk of 'shame' and 'sadness' creates a context for rationalizing
The portion of Tom's narrative in which he talks most
about his emotional responses is given subsequently so as to look hls own actions. Most problematic is his telling Fala's group (through a
closely at the conceptual role of emotion rhetoric in building a desi €Brrvcrsation with her mother, Sukhi) not to bother coming up to his
interpretation. By combining Tom's overt statements about emotion Ylllngc for the Christmas feast. The attribution of 'shame' provides a
the hypothesized knowledge structures discussed earlier, the ana *lotivc for Tom's hasty words. The prototypical scenario for shame (not
seeks to identify implicit structures of emotional meaning' (For detailccl here) entails a response of withdrawal. He here implies that his
analysis, only attributions of 'sadness'are expanded in this way, but Entorks to Sukhi were more of a response of avoidance in the face of
White (in press) for a similar treatment of 'shame") Certain *he rocial distance created by Fala's rejection of his brother.
are useful for depicting interpretation as a Process of inference As '[om goes on with his narrative, he focuses more directly on his
ins 1973; Quinn 1987). Borrowing from the standard format for fullngs and motives prior to making those remarks to Fala's mother.
senting syllogistic reasoning, the implications of emotion attribution Hle tnlk of 'shame' is embedded in attributions of 'sadness'. At this point
:,
..

62 Geoffrey M. White Moral discourse and the rhetoric of emotions 63

in the narrative, the emotion rhetoric is oriented more to types of tion. 'Sadness' is the emotion that links Tom's reaction to the exclusion
sponse than to characterization of antecedent events. of his brother with his own confrontation with Fala's group - an emotion
The hypothesis that much disentangling discourse is concemed wi that follows from a concern for valued relations and is supposed to lead
the transformation of scenarios of 'anger' to a more desired reality to efforts at moral repair. Although Tom's statements in the preceding
examined in the following passage. In this portion of rom's narratir passage emphasize the types of action that follow from 'sadness' (and
as he continues describing his encounter with Sukhi, the 'shame'), he also reminds his listeners about the source of those emo-
schema can be seen to be lurking just behind the scenes. In fact, T tions in disrupted social relations by reiterating images of kinship and
here applies a tactic unusual to disentangling discourse: He explici food sharing. These images emphasize the social ties that are an essen-
poseq..ghe anger scenario as a possible altemative interpretation of tial element of the interpretation that replaces a scenario of angry con-
hu&y- words so that he can reject it from consideration (see eui frontation with an awkward withdrawal motivated by'shame' and'sad-
1987:185*6 for a similar example).11 ness'. p{ether this strategy works, of course, is uncertain. But it is not
so mucftSe truth value of his assertions as their coherence and force-
It's just that when my real father said that [reported Fala,s
fulness iff the disentangling context that constitute a successful perfor-
that I was sad and came and talked to the old woman [Sukhi]
mance.
here.
This example, drawn from a single disentangling narrative, illustrates
SAD --+ REPAIR DAMAGED RELATION the ethnofunctional beauty of disentangling as a predefined social occa-
"sad --+ Tom's remarks to Sukhi" sion in which people rhetorically mend minor tears in the social fabric
Tom's remarks : repair damaged relation before relations unravel further. In this context, direct talk of 'anger' that
leads to retribution seems to surface only in regard to minor incidents,
I was just ashamed is all. It was not [I didn't mean], ,,you,sall ond even then it may be problematic (White in press). Instead, disentan-
come back up. Don't set food i^ *y village, you all.,, I didn t I gling discourse plays on the inherent moral ambiguities of interpersonal
that.
conflict to characterize'angry' reactions as responses of 'sadness' and
ANGER --+ RETRIBUTION 'shame', thereby reframing events that have been problematic for per-
"Tom's remark not retribution" rons and communities alike. In short, disentangling - as ideology and
not anger irs social institution - works (however imperfectly) to bring conditions
of divisiveness and individuated 'anger' more in line with models of
That was just from my sadness, my own shame. nolidarity.
SAD -+ REPAIR DAMAGED RELATION
"sad ---> Tom's temarks"
Conclusion
Tom's remarks: repair damaged relation
Disentangling institutionalizes an activity in which community mem-
Now I'm in4iont of my mothers and sisters [in the village pa lrers ritually realign skewed relations. Talk of emotions such as di'a nag-
These houses should be for coming inside. These houses are nnfa is a culturally constituted means of periodicilly re-creating these
me to get a drink. These houses should be for taking a rest, nlignments while simultaneously advancing more covert agendas of moral
when my [fictive] father said that, Iwas sad about it. argumentation. The consistent attribution of 'sadness' and 'shame' to
DAMAGE CLOSE RELATION-+ SAD problematic events (seen by many as having evoked hostile responses)
"Fala's remark --> sad" ruggests that disentangling discourse rhetorically transmutes 'anger' to
Fala's remark : 'Etldness'. There is enough overlap in the 'anger' and 'sadness' schemas
damage close relation
ao that they may plausibly be applied to the same events, with the effect
In this passage/ the attributions of 'sadness'and ,shame, both of reconstruing conflict situations so as to emphasize reconciliation in-
to an emerging interpretation of Tom's response to hie brother,s ate'ad of retribution. And, by implication, the likelihood of suffering any

h
a

64 Geoffrey M. White Moral discourse and the rhetoric of emotions 65

of the broad range of damaging effects seen to follow from unresolved Notes
'anger'is reduced.
'l'his chapter is based on fieldwork in santa Isabel carried out in 1975-6 and on
Without either the shared model of disentangling or the institutional-,
rnore focused research on disentangling conducted during two months in 1984.
ized context in which to produce narrative accounts of conflict, neither support of the latter project by the wenner-Gre,n Foundation (grant no. 4248) is
the rhetoric nor the reality of emotion manifest in disentangling grdtifrly acknowledged. I arir also grateful to Jack Bilmes, Lamont Lindstrom,
be what it is. If one construes emotions as socioculturul institutions tha and Catherine LUtz for helpful comments on this chapter, as well as to numer-
ous others who have offered insightful remarks on a longer analysis of disentan-
depend on both the cultural model and the interactive situation for thei g,ling discourse (White in press).
meanin-g and effect, then analytic attention is directed more widely tha i, A third major thrust-in anthropological approaches to discourse not ad-
a strigt[f$erson-centered approach would suggest.l2 By recognizing t dressed directly represented in poststructuralist con-
dir6ctly in this chapter is that rePresen
cerns with the historical and political conditions necessary for statements to
are not simply expressed in social situations, but are in 1
obtain truth value and moral [orce (epitomized in the work of Foucault, e.g.,
"rr,oti5frr by the types of activities and relations in which they a
constituted 19811). '[take the sociopolitical preconditions of knowledge and power that
enacted, ethnographic attention may be given to the institu make tffirtain discouise possible to be different from the conceptual and
commurfrcative structures-through which that discourse takes on cultural
discourse practices that shape emotional meaning and experience (
and experiential reality. However, approaching emotion as discourse opens
and White 1986). up theiormer vantage point as a way of d-eline-ating the historical and insti-
Because of its evocative functions, emotion discourse is especially tutional determinant-s oi emotions as a culturally constructed political force
(see Abu-Lughod, this volume).
evant for theories of social action. To talk about or express emotion
As a protory-pe structure, the schema represents a typical.or default course
context is to expect to evoke a certain type of response in both the for socioemolional process, but not one that is seen as invariant or com-
and the listening other. However, analysis of the socioemotional pletely predictable.-Ethnopsychological knowledge of persons and action
ties so created is complicated by the fact that the analysis is likely to inuy be elaborated in a greit variety of ways and may still be consistent with
the basic shape of the event-emotion-response scenario- For example, D'An-
beyond the conscious awareness of the participants. In the case drade,s (1987) discussion of the American folk model of the mind postulates
tangling, participants appear not to recognize or acknowledge the tra a more detailed picture of conceptions of feelings and desires related to a
formative effects of disentangling narrative, referring instead to the range of mental and behavioral processes.
'Ihii chapter does not examine in detail the more_ specific sorts of relational
out' aspeet of disentangling. Even though participants acknowledge
information encoded in emotion concePts. Knowledge about the social con-
attempts of narrators to replay past events to their moral advantage, texts of emotions applies to the salience of certain emotions for certain rela-
folk theory of disentangling does not encompass the act of tions, as well as to iheir appropriateness of expression in those relations.
specific emotions. In disentangling, as in many emotive institutions, lior example, 'anger' in Isabel is widely regarded as a problematic.response
to conflici and trinsgression. Yet, according to cultural ideals, it is not ap-
nature and degree of participants' awareness of discourse functions propriately expressed in relations that are either close (e-g., among family
important elements in the institution's ability to achieve its inte incmbers;'or sharply asymmetrical (e'g., between a 'good' chief and his fol-
ends. Iowers).
Sce Wierzbicka (1986) for a discussion of translation issues in the domain of
Perhaps ironically, a discourse-centered approach such as that emotion. she draws attention to the need to decouple the cross-cultural study
lined here holds pgticular promise as a method for examining si of emotion from the unexamined semantics of English emotion terms. Al-
cances of emotional meaning unrecognized or unacknowledged by though the type of semantics outlined by wierzbicka is narrowly lexical (and
ticipants. It is usually taken for granted that the analysis of discourse rultinfttely tiiriitea by the absence of contextual and performative-informa-
tion), hei suggestion that translation work may_be-advanced by the use of
an avenue to identifying public constructions of socioemotional cxpiicit metaliirguistic models is consistent with the search for prototype
To this we should add that analysis of patterns of unspoken meaning ntructures in emotional meaning.
'l'he vernacular term di'a nagnafa (literally, 'bad heart' or 'bad feeling') is not
emotive discourse may also provide a means of investigating less
wcll translated by the Engliih gloss 'sad.' For a more extended discussion of
ble" transformations of personal experience. the local meanirigs of thls eniotion, see White (in press). English glosses
Tangry' used in this chapter for expository
Fuch as 'sad' and PurPoses should
not be read a8 a clalm of cross'cultural equivalence.
'l'he 'sadness' schema as eketched centers on the resPonse of the experi-
66 Geoffrey M. White Moral discourse and the rhetoric of emotions 67

encing subject. However, as Lutz notes in her discussion of Ifaluk emotion


theory $987:295-8), cultural models of emotion also encompass under- References
standings about the evocative power of emotions to elicit specific feelings
Abu-Lughod, Lta. 1986. Veiled Sentiments: Honor and Poetry in a Bedouin Society.
and responses from others. Thus, in the case of'sadness', not only does
Berkeley: University of California Press.
the self desire repair [REPAIR (Self, Other)], but so, ideally, does the other
Bailey, F. G. 1983. The Tactical Uses of Passion: An Essay on Pouter, Reason and
Self)1. Here the reciprocal quality of 'sadness' implies a
[REPAIR (Other, Self)].
TREPAIR Reality. Ithaca, NY: Cornell University Press.
i.,n.r..r.n".
convergence in goals, ind other, aicordine
soals. with both self and according to the schema,
Briggs, lean. 1970. Neoer in Anger: Portrait of an Eskimo Family. Cambridge, MA:
seeking to reestablish mutual relations. By implication, then, tall
Harvard University Press.
D'Andrade, Roy. 1987. A Folk Model of the Mind. In D. Holland and N. Quinn,
priate) response of the
discourse is based on four particular ses' eds., Cultural Models in language and Thought. Cambridge: Cambridge Uni-
7. M'*- i'sample" of disentangling
Myr.pample" diser
qio-ri3, two of which were tape-recorded.
versity Press, pp. 11248.
Epstein, A. L. 1984. The Erperience of Shame in Melanesia: An Essly in the Anthro-
8. C6ncepts of shame have received a great deal of attention in the cross-
pology of Affecf . London: Royal Anthropological Institute of Great Britain and
cultural literature on emotion (see, e.9., Epstein 1984, White and Kirkpa-
mamaia, figures impor-
Irelan{, Occasional Paper No. 40.
trick 1985). Although the A'ara term for 'shame', Foucault,@fthel. 1981. Th-e Order of Discourse. In R. Young, ed., Untying the
tantly in disentangling discourse, limitations of space preclude a detailed
Text: ,&Post-Structuralist Reailer. Boston: Routledge & Kegan Paul, pp. 4&-
discussion here (but see White in press).
78.
9. In this instance, the entering-a-house-for-food scenario is stated with con-
Freeman, Derek. 1983. Margaret Mead and Samoa: The Making and Unmaking of an
ditional past tense markers, indicating that the relations signified are
Anthropological Myth. Cambridge, MA: Harvard University Press.
tingent on their continuing acknowledgment by participants. Tom im
Gumperz, ]ohn. 1982. Discourse Strategies. Cambridge: Cambridge University Press.
that Fala's exclusion of his brother constituted a failure of such ackr
t-larris, Grave. 1978. Casting Out Anger: Religion Among the Taita of Kenya. New
edgment, thus creating shame in contexts where none would be evoked
York: Cambridge University Press.
conditions for the scenario were in force.
l{olland, Dorothy, and Naomi Quinn, eds. 1987. Cultural Models inLanguage and
10. For the sake of simplicity, the relation achieved by instantiating a gene
Thought. Cambridge: Cambridge University Press.
ized category with a specific action is indicated by an equals sign (: ). T
t{utchins, Edwin. 1973. An Analysis of Interpretations of Ongoing Behavior.
notation is intended as shorthand for the phrase "is an instance 6t" (
"is equivalent to"). It should also be noted that this sort of inference is Unpublished manuscript, Department of Anthropology, University of Cal-
ifornia, San Diego.
one compelled by the canons of propositional logic. The closest syllogis
1980. Culture and lnference. Cambridge, MA: Harvard University Press.
to that represented here would be something like: given A --> B, and
("sad') is true, therefore A. But, since reasoning from the consequent Ito, Karen L. 1985. Ho'oponopono, "To Make Right": Hawaiian Conflict Resolu-
tion and Metaphor in the Construction of a Family Therapy. Culture, Medi-
not a move compelled by logic (not a "strong" inference), mat
cine and Psychiatry 9:20'1.-17.
and others have legitimized this sort of inference as "plausible" reast
(see Hutchins 1980:56 for examples and discussion). This is the most
l,akoff, George, and Zoltan Kovecses. 1987. The Cognitive Model of Anger In-
herent in American English. In D. Holland and N. Quinn, eds., Cultural
mon type of inference in the examples discussed here.
Models in Language and Thought. Cambridge: Cambridge University Press,
11. I'm grateful to Naomi Quinn for drawing my attention to this example of
pp.195-221.
discourse strategy in which a proposition is rejected, even as the schen
l,evy, Robert. 1984. Emotion, Knowing and Culture. In R. Shweder and R. LeVine,
'eds.,
necessary to its interpretation is invoked to do so.
Cultural Theory: Essays on Mind, Selt' and Emotion. Cambridge: Cam-
t2. This notion of "emotive institution" is well illustrated in recent work bridge University Press, pp. 274-37.
suicide in Pacific societies. For example, Rubinstein's (1984) study of T
l,utz, Caiherine. 1985. Ethnopsychology Compared to What? Explaining Behav-
amu)unumwun s[rows that the emotions implicated in suicide events
ior and Consciousness Among the Ifaluk. In G. White and |. Kirkpatrick,
constituted in ftrticular communicativ" fo'.rnt and culturally patter
eds., Person, Self and Experience: Exploring Pacific Ethnopsychologies. Berkeley:
scenarios of conflict. The indigenous conceptualization of these
University of California Press, pp. 35-79.
appears to bridge categories of affect and action. The study of this
1987. Goals, Events and Understanding in Ifaluk Emotion Theory. In D. Hol-
related practices in the Pacific (see, e.g., Freeman 1983:218-22 on Sar
land and N. Quinn, eds., Cultural Models in Innguage and Thought. Cam-
bridge: Cambridge University Press, pp. 290-312.
l,utz, Caiherine, and Geoffrey M. White. 1986. The Anthropology of Emotions.
White in press).
Annual Reaiew of Anthropology 15:405-36.
Myers, Fred. 7979. Emotions and the Self: A Theory of Personhood and Political
Order Among Pintupi Aborigines. Ethos 7:343-70.
1988. The Logic and Meaning of Anger Among Pintupi Aborigines. Mar (N.S')
23:589-610.

h
:'
..

68 Geoffrey M. White
Poole, F. I. P. 1985. The Surfaces and Depths of Bimin-Kuskusmin Experiences
ol "Anger": Toward a Theory of Culture and Emotion in the Constitution
of Self. Paper given at the 84th Annual Meeting of the American Anthro-
pological Association, Washington, DC.
4. Engendered emotion: gender,
Quinn, Naomi. 1987. Convergent Evidence for a Cultural Model of
Marriage. In D. Holland and N. Quinn, eds., Cultural Models in
power, and the rhetoric of
Universi Press, pp. 17T92.
and Thought. Cambridge: Cambridge Universify
rldo, Michelle. 1980. Knowledge and Passion: ll
Rosaldo, llongot Notions of Self and emotional control in Amerrcan
Life. Cambridee: Cambridee University
Cambridge: Cambridge Universitv Press.
nstein. Donald. 1984. Self-Righteous
Rubinstein, Self-Riehteous Anger, Soft-Talk and Amwunumutu
^/
discourse
Sgrcrdes of Young Men. Paper given at the 83rd Annual Meeting of
.;lifierican Anthropological Association, Denver, CO.
Schieffelin, Edward. 1983. Anger and Shame in the Tropical Forest: On Al CATHERINE A. LUTZ
as a Cultural System in Papua New Guinea. Ethos'l,l:181-91.
Shweder, Richard, and Robert LeVine, eds. 1984. CultureTheory: Essaqs on Mi -,t
ln Wesftf:academic discourse, emotions have begun to move from their
Self and Emotion. Cambridge: Cambridge University Press.
Strathem, Marilyn. 1968. Popokl: The Question of Morality. Mankind 6:553*62, culturallf assigned place at the center of the dark recesses of inner life
Valentine, Charles A. 1963. Men of Anger and Men of Shame: Lakalai and are being depicted as cultural, social, and linguistic operators. In
psychology and Its Implications for Sociopsychological Theory. the process, we can ask not only about the cultural foundations of things
1:441-77.
Watson-Gegeo, Karen, and Geoffrey M. White, eds. In press. Disentangling: construed as emotional, but about the organizing category of "emotion"
Discourse ot' Interpersonal Conflict in Pacific lsland Societies. Stanford: Stan: Itself. One important aspect of that category is its association with the
University Press. fumale, so that qualities that define the emotional also define women.
White, Geoffrey M. 1985a. Premises and Purposes in a Solomon Islands
psychology. In G. White and J. Kirkpatrick, eds., Person, Self and
For this reason, any discourse on emotion is also, at least implicitly, a
Exploring Pacific Ethnopsychologies. Berkeley: University of Califofriia discourse on gender.
pp.328-65. As both an analytic and an everyday concept in the West, emotion,
1985b. 'Bad Ways' and 'Bad Talk': Interpretations of Interpersonal Conflict
llke the female, has typically been viewed as something natural rather
a Melanesian Society. In J. Dougherty, ed., Directions in Cognitioe
ogy:Utbana: University of Illinois Press, pp. 345-370. than cultural, irrational rather than rational, chaotic rather than ordered,
1988. Symbols of Solidarity in the Christianization of Santa Isabel. In G. S, rubjective rather than universal, physical rather than mental or intellec-
ders, ed., Culture and Christianity. Westport, CT: Greenwood Press, pp. tual, unintended and uncontrollable, and hence often dangerous. This
31.
ln press. Emotion Talk and Social Inference: Disentangling in a Solomon Retwork of associations sets emotion in disadvantaged contrast to more
lands Society. In K. Watson-Gegeo and G. White, eds., Disentangling: valued personal processes/ particularly to cognition or rational thought,
Discourse of Intupersonal Conflict in Pacific Island Societies. Stanford: Stan 6nd the female in deficient relation to her male other. Another and com-
University Press.
peting theme in Western cultural renditions of emotion, however, con-
White, Geoffrey M., and ]ohn Kirkpatrick, eds. 1985. Person, Self and Exper
Exploring Pacific Ethnopsychologies. Berkeley: University of California Pr trauts emotion with cold alienation. Emotion, in this view, is life to its
White,
te, Geoffrey, ryancrs Kokhonigita,
Geottrey, Fgancis KoKnomglta, anoand Hugo Pulomana. 1988. Cheke abecnce's death, is interpersonal connection or relationship to an
Dictionary. Car{berra: Pacific Liiguistics.
Unemotional estrangement, is a glorified and free nature against a
Wierzbicka, Anna. 1.986. Human Emotions: Universal or Culture-Specific?
ican Anthr op olo gist 88:584-9 4. $aekling civilization. This latter rendition of emotion echoes some of
the fundamental ways the female has also been "redeemed," or alter-
iatlvely and more positively, construed (Lutz 1988).
In this chapter, I will explore how emotion has been given a gender
h come sectors of American culture and, in the process, make two re-
arguments. First, I will demonstrate that local or everyday lay dis-
rEe on emotion explicitly and implicitly draws links among women,
nation, rebellion, and emotion by examining interview conver-

L
..

72 Catherine A. Lutz Engendered emotion 73

about emotions, one of the most common set of metaphors used is we can ask what the rhetoric of control might accomplish for the speaker
that in which someone or something controls, handles, copes, deals, and what it might say to several audiences (see Brenneis, this volume).
disciplines, or manages either or both their emotions or the situation At least three things can be seen to be done via the rhetoric of emotional
seen as creating the emotion. For example: control: It (1) reproduces an important part of the cultural view of emo-
I believe an individual can exercise a great deal of control over their tion (and then implicitly of women as the more emotional gender) as
emotions by maintaining a more positive outlook, by not dwelling irrational, weak, and dangerous; (2) minimally elevates the social status
on the negative, by trying to push aside an unpleasant feeling. I'm of the person who claims the need or ability to self-control emotions;
getting angry and like I said, he's over being angry, more or less and (3) opposes the view of the feminine self as dangerous when it is
reversed, that is, when the speaker denies the need for or possibility of
4[ifopea it and he expects me to also. Wel] we don't have the same
ieihper, I just can't handle it that way. control of emotion. Each of these suggestions can only briefly be exam-
And in a more poeticturn, one person mused: ined.,. {
firsffiis rhetoric can be seen as a reproduction, primarily on the part
sadness . . . dipping, dipping into that . . . just the out-of-control-
of womin, of the view of themselves as more emotional, of emotion as
ness of things.
dangerous, and hence of themselves as in need of control. It does this
People typically talk about controlling emotions, handling emotional situ- first by setting up a boundary - that edge over which emotion that is
ations as well as emotional feelings, and dealing with people, situations, uncontrolled can spill. A number of people have noted that threats to a
and emotions. dominant social order are sometimes articulated in a concem with diverse
The notion of control operates very similarly here to the way it does kinds of boundaries (whether physical or social) and their integrity (e.g.,
in Western discourses on sexuality (Foucault L980). Both emotionatty, Martin 1987; Scheper-Hughes and Lock 1987). One of the most critical
and sexuality are domains whose understanding is dominated& by a boundaries that is constituted in Western psychological discourse is that
biomedical model; both are seen as universal, natural impulses; both are between the inside and the outside of persons; individualism as ideol-
talked about as existing in "healthy" and "unhealthy" forrns; and both ogy is fundamentally based on the magnification of that particular
have corne under the control of a medical or quasi-medical profession boundary. When emotion is defined, as it also is in the West, as some-
(principally psychiatry and psychology). Foucault has argued that pop- thing inside the individual, it provides an important symbolic vehicle by
ular views of sexuality - as a drive that was repressed during the Victo- which the problem of the maintenance of social order can be voiced. A
rian era and gradually liberated during the twentieth century - are mis- discourse that is concerned with the expression, control, or repression
leading because they posit a single essence that is manipulated by social of emotions can be seen as a discourse on the crossing back and forth of
convention. Rather, Foucault postulated, multiple sexualities are con- that boundary between inside and outside, a discourse we can expect to
stantly produced and changed. A popular discourse on the control of see in more elaborate form in periods and places where social relations
emotion runs functionally parallel to a discourse on the control of sex- appear to be imminently overturned.
uality; a rhetoric of6ontrol requires a psychophysical essence that is ma- This rhetoric of emotional control goes further than defining and then
nipulated or wrestled with and directs attention away from the socially defending boundaries, however; it also suggests a set of roles - one strong
constructed nature of the idea of emotion (see Abu-Lughod and Lutz, and defensive and the other weak but invasive - that are hierarchized
this volume). In addition, the metaphor of control implies something and linked with gender roles. Rosaldo (1984) notes of hierarchical soci-
that would otherr,rrise be out of control, something wild and unruly, a cties that they seem to evince greater concern than do more egalitarian
threat to order. To speak about controlling emotions is to replicate the tunes with how society controls the inner emotional self and, we can
view of emotions as natural, dangerous, irrational, and physical.2 ndd, with how one part of a bifurcated and hierarchically layered self
What is striking is that women talked about the control of emotion controls another. The body politic, in other words, is sometimes repli-
more than twice as often as did men as a proportion of the total speech cated in the social relations of the various homunculi that populate the
each produced in the interviews.3 To help account for this difference, human mind, a kind of "mental politic." When cognition outreasons

h
i

74 Catherine A. Lutz
Engendered emotion 75

who do. Women, more than men, may speak of control because they
and successfully manages emotion, male-female roles are replicated.
are concerned about counteracting the cultural denigration of them-
When women speak of control, they play the roles of both super- and
selves through an association with emotion. "I think it's important to
subordinate, of controller and controllee. They identify their emotions
control emotions," they say, and implicitly remind a critical audience
and themselves as undisciplined and disciptine both through a dis-
that they have the cooler stuff it takes to be considered mature and ra-
course on control of feeling. The construction of a feminine self, this
tional. It is important to note that, as academics, I and the graduate
material might suggest, includes a Process by which women come to
students who conducted the interviews may have been perceived as an
control themselves and so obviate the necessity for more coercive out-
audience in special need of such reminders. The speakers would have
side control
lirr " been doing this, however, by dissociating themselves from emotion rather
The;?j;is the example of one woman in her late thirties; she talked
than by questioning the dominant view both of themselves and of emo-
aboui the hate she felt for her ex-husband, who began an affair while
she was pregnant and left her with the infant, an older child, and no
tion. q
AltliffiBh women may have less access to a view of themselves as
paid employment:
masterftll individuals, a common aspect of the cultural scheme that is
So I think you try hard not to bring it [the feeling] out 'cause you available paints them as masterfully effective with others on joint tasks,
don't want that type of thing at home with the kids, you know: particularly interpersonal or emotional tasks (social science versions of
That's very bad, very unhealthy, that's no way to grow up. So I this include Chodorow L978, Parsons and Bales 1955). This subtly alters
ttfnk now, maybe I've iust leamed to control it and time has changed the meaning of the rhetoric of control; knowledge of what the feelings
the feeling of the hate. are that "need" control and of what control should be like is perceived
The woman here defines herself as someone with a feeling of hate a and described as a social rather than an individual process. For example,
portrays it as dangerous, primarily in terms of the threat it posesto one woman says: "If you're tied in with a family, . . . you have to use it
own children, a threat she phrases in biomedical terms (i.e., for guidance how you control your emotions." This is the same woman
healthy"). She replicates a view that Shields (198n found prevalent in whose central life problem during the interview period was coping with
snrvey oJ twentieth-century English-language child-rearing manuals; her husband's ex-wife and family, who lived across the street from her.
is the danger that mothers' (and not fathers') emotions are thought 'l'he regular, friendly contact between husband and ex-wife has left her
present to children. In addition, this woman's description of her very unhappy but also unsure about what to do. The ambiguity over
essentializes them as states; as such, they remain passive (see who ought to control or regulate what is evident in her description of an
L987 on the feminization of love) rather than active motivators, a argument she had with her husband over the issue.
to which we will return.
I was mad. I was mad. And I said, "I don't care whether you think
In other cases, people do not talk about themselves, but rather
others (usually women) of the need to control themselves. These
I should [inaudible word] or stay in this at all, it's too, and cause
I'm going to say it." And I said, "How dare you tell me how I'm
stances also serve fo replicate the view of women as dangerously el
supposed to feel," you know. Bob [her husband] would say, you
tional. Another rt#ou*rpoke about a female friend who still grieved
know, "You got to live with it" or "You got to do this" or "How
a son who had died two years previously: "You've got to pick up a
dare you tell me this, I don't have to put up with anything" or "I
go on. You've got to try and get those feelings under control." The "
don't have to feel this way because you tell me I have to feel this
in this statement is a complex and multivocal sign (Kirkpatrick 1
way." You know, it was, in that case Robin is his ex-wife, "and
and directs the admonition to control simultaneously to the gri
you have to just kind of deal with it," you know, "all the problems
woman, the female interviewer, the speaker herself, no one in
that she presents in your own way." And it was almost sort of like
lar, and everyone in a potential audience.
saying "You're going to have to like it." Well I don't. I don't, you
A second pragmatic effect of the rhetoric of emotional control is a
know. And for a year and a half he kept saying, you know, "You're
to have the ability to "rise above" one's emotions or to approve of

t
r- a

76 Catherine A. Lutz Engendered emotion 77

going to have to like it, this is the way it's going to be, you're going he had to over-intellectualize problems and explained that he worked
to have to do this, you're going to have to have, be, act, this certain against that tendency because
wayi'you know, act everything hunky-dory, and it wasn't, you It wasn't that I wanted to cut off my emotions, I just didn't, they
know, and I was beginning to resent a whole lot of things. I, I, I would get out of control, and I found that the more I tried to sup-
resented him for telling me I had to feel that way when I, I wasn't press them, the more powerful they would become. It was like this
real fond of the situation. I didn't like it. When I would tell him big dam that didn't let a little out at a time, it would just explode
that I didn't like it, it was "It's your problem, you deal with it." I all of a sudden, and I'd be totally out of control.
didn't like that, that made me really angry because I was'saying, The question remains, however, of the validity of seeing these latter
'&fep me out here, I don't know how to deal with this."
, seemingly resistant uses of the rhetoric of emotional control as "oppo-
This woman is frustrated with her husband for failing to join her in a sitional" forms (Williams L977) within that system.s This is certainly a
collaborative project of "dealing with" her feelings of resentment. Here dangqftgs rhetorical strategy, caught as they (we) are within a hege-
control is given away to or shared with others. This strategy of control monicffiscourse not of our own making. The opposition to self-control
is more complex and subtle than the simple self-imposition described in will most likely be absorbed into the logic of the existing system and so
other parts of the transcripts so far; it aims to control both the emotiong come to equal not resistance but simple deficiency or lack (of control).6
of the self and the attention and assistance of the other. Note also that A possibly oppositional intent may have collaborative outcomes to the
she speaks of "resenting" ot "rtot liking" (relatively mild terms of dis' e'xtent that the denial of self-control is taken by most audiences as a
pleasure) the overall situation but is most incensed ("mad, mad, mad") deficit and a confirmation of ideas about women's irrationality.
about her husband's assumption that she ought not to feel a ce The culturally constructed emotionality of women is rife with contra-
way. She asserts the right to "feel" unhappy about her predicagent bt diction. The emotional female, like the natural world that is the cultural
is clearly defining that feeling in the standard contemporary sense of source of both affect and women, is constructed as both pliant (because
strictly internal and passive event' Nowhere in the interview does weak and a resource for use by civilized man) and ultimately tremen-
explicitly state or appear to imply that she wants, intends, or ought dously powerful and uncontrollable (Strathern 1980).7 Emotionality is
act in cdncert with those feelings. what is being controlled or dealt wi the source of women's value, their expertise in lieu of rationality, and
therefore, has already been defined as a relatively innocuous feeling ra yct it is the origin of their unsuitability for broader social tasks and even
than an action tendency. a potential threat to their children.
Finally, the rhetoric of emotional control can also be employed in There are vivid parallels between this and the cultural meanings sur-
idiosyncratic and "reversed" ways that may intend or have the effect rounding colonialism that Taussig (198q and Stoler (1985) have de-
at least minimally resisting the dominant view of emotionality, and ncribed. Looking at early-twentieth-century colonists' views of the local
of women. A few people, for example, spontaneously spoke about Columbian labor force, Taussig describes their alternation between fear
problem of emotional control, thereby evoking the whole schema nnd awe of Indians who were perceived as dangerous and powerful
have just been lo&king at. They went on, however, to define "con figures, on the one hand, and disgust and denigration of their perceived
in a way that entailed relatively minimal constraints on emotional wcakness and lack of civilization, on the other. Taussig describes the
munication. One woman, a twenty-eight-year-old bank teller, said: process as one in which a "colonial mirror" "reflects back onto the col-
me explain control. It's not that you sit there and you take it [some onists the barbarity of their own social relations" (1984:495).In a (cer-
of abuse] and, you know, I think controlling them [emotions] is tnirrly less systematic or universally brutal) way, a "patriarchal mirror"
them out in the proper time, in the proper place." Perhaps more cnn be conceptualized as helping to produce the view of women as emo-
cally, some women (as well as one of the gay men with whom I tlonal - as dangerously "eruptive" and as in the process of weakly
denied that they had the ability to control some or many of their "trrcaking down." A "paradox of will" seems consistently to attend
tions.a One man in his twenties critically described a previous dominating relationships - whether those of gender, race, or class - as
F
Engendered emotion
78 Catherine A. Lutz 79

the subordinate other is ideologically painted as weak (so as to need and has been attributed by the biomedical research community to hor-
protection or discipline) and yet periodically as threatening to break monal imbalances in the women who suffer from it. The syndrome has
ideological boundary in riot or hysteria. Emotion talk, as evident in t
been used to explain a host of emotions ranging from irritability and
transcripts, shows the same contradictions of control, weakness, mood swings to depression, anxiety, and panic attacks. A number of
strength. Given its definition as nature, at least in the West, emoti feminist critiques (Archer and Lloyd 1985; Fausto-Sterling 1985; Gottlieb
discourses may be one of the most likely and powerful devices by w 1987; Whatley 1986) have pointed out the weakness of the evidence for

domination proceeds. this syndrome. Assessment of women's mood is usually based on retro-
spective self-report via questionnaires (one popular version being titled
|!:r," the "Menstrual Distress Questionnaire"), which allow women to draw
Thbengendering of emotion in science ttn cultural knowledge about the relation between gender, emotion, and
Demonstrations of the political, moral, and cultural bases of W hormoqes. Conversely, studies that disguise the purposes of the ques-
science have been made convincingly in a number of natural and tionnd{psshow no significant premenstrual mood changes. The putative
fields (e.g., Asad,1973; Fausto-Sterling 1985; Haan, Bellah, Rabinow, therap&tic effects of hormone injections are taken as primary evidence
Sullivan L983; Sampson 1981). In like fashion, it can be argued tl of the female hormonal basis for mood changes, but these studies have
the sciences of emotion have been, in a significant sense, a product Rot been "double-blind." As Whatley argues, this biomedical discourse

their social context. In particular, the academic literature on emotion on emotions and gender may "cause us to ignore the fact that our pre-
be considered a form of political discourse on gender relations menstrual mood changes . . . may also correlate more closely to a monthly
of the marked associations between the two domains. That li eycle of low bank balances than of hormonal fluctuations" (1986:L83).
thus arises out of and reenters a field of power struggles for the de Moreover, the emotional symptoms of premenstrual syndrome can be
tion of true womanhood. As Haraway (1986) has said of Amffiican !€en as a discourse on both the good and the deviant woman, on the
matology, it can be seen as "politics by other means," and in the necessity of her emotional suffering and the abnormality of, especially,
emotions, it is most centrally a politics of gender by other rl€€tll3r her anger or irritability (Gottlieb 7987), both common symptoms at-
examining several examples of studies of emotion, we will see that hched to the syndrome. Normative academic and clinical work on pre-
research over the years in biology, psychology, sociology, soci Renstrual syndrome focuses on the emotionality of women as both
tics, and other fields has been implicitly based on everyday cultural
,o, ur,& yet as a "symptom" in need of a cure. This research draws
tinking women and emotionality, and that this research moves the entrenched cultural view of emotions as sited in females, as nat-
assumption of these cultural premises to their "proof." Most std in essence (like but independent of the "naturalness" of females),
as irrational or pathological when they occur.
about these studies is the number that naturalize the purported
differences by attributing them to biological or necessary and unlv Thls line of research follows from and reinvigorates the cultural model
features of the female ,lt" ir, physical and social reproduction, I Which women are more emotional than men because they are more
briefly examine6ieveral areas of research, including the analysis to the biological processes that produce emotion. Wombs, men-
menstrual syndrome and mood, sex differences in the recognlt lon, and hormones "predic(' emotion. A more tacit part of the
facial expressions of emotion and in aggression, and studies of th! ral logic connecting women and emotion may arise from the view
tive components and concomitants of motherhood' Feminist n as biologically inferior both because they menstruate and be-
the.y are smaller, weaker/ and lack a penis. When viewed as a form
a number of these latter fields have been intensive, and I will
ynical chaos or "breakdown," emotion is one other form of biolog-
them while extending the analysis to the domain of emotion,
Studies of the relationship between mood and hormonal weakness suffered by women.s
focused on women's (rather than men's) cycles and in the number of people in the interview study just described sponta-
discovered the hormonal disease of Premenstrual syndrome, ardorlated related ideas about the relationship between women,
drome is characterized by both phyeical pain and mood emotion, and pathology. In Beveral caoeo, they referred to
..

80 Catherine A. Lutz Engendered emotion 81

research as the authoritative source of their assertions, although my aI- Another line of research, on sex differences in aggression, also draws
gument is that the relationship between everyday and scientific ideas on cultural views of emotion and women. This happens, first, because
about women and emotion is dialectical rather than an idea system im- aggression, at least in the Western cultural view, is seen as retrospec-
posed hegemonically on a previously blank or very different lay model. tively predictive of anger (Montague 1978). Anger is the one emotion
According to one woman, a forty-eight-year-old telephone operator, that is exempted in everyday discourse from the expectation that women
"women have been known to have different reactions to the same situ' feel and express more emotion than men. It is in fact every emotionbut
ation at different times of the month. And that's been a study. I've seen anger that is disapproved in men and, conversely, expected in women
where some women can be downright dangerous, they could be poten- (Hochschild 1983). This gender stereotype has been shown to have been
tial killei's." thoroughly learned by American children as early as the preschool pe-
Ar$ther field in which some attention has been paid to sex differences riod (Birnbaum, Nosanchuck, and Croll 1980, cited in Shields D8n. A
is the study of facial expression of emotion. In one sociobiological ac' recenf,*$zidely accepted, and often cited set of studies makes the parallel
count, female emotionality is a product of evolution. Babchuk, Hames; claim t@Ve demonstrated a relationship between levels of the "riirale"
and Thompson (L985) interpret studies showing that women are better hormonE testosterone and aggression. Fausto-Sterling (1985) demon-
able than men to read facial expressions of emotions in infants. In their strates the weakness of the evidence for this claim and questions why it
view, this is the result of women's long history of being the primary has been taken up so enthusiastically by so many.
caretakers of infants and the reproductive value of using these facj The echoes of the lay view in the scientific are followed by the echoes
cues to detect infant distress. This argument is implausible on ma of the scientific view in the lay on this point as well. A professional
grounds, not the least of which are the redundancy in infants of woman in her forties in the interview study commented on the associa-
expression and other cues to discomfort, and the theoretically at tion between aggression and gender: "So far the research shows that,
equal value of facial expression recognition skills for the prehistor{& yes, little boys are inherently more aggressive than little girls. . . . I think
who, in many evolutionary accounts, were engaged primarily in de it bothers me that there's a sex link with aggression. There are a couple
ing the female and infant against threatening and dissembling outsi of sex-linked ones that bother me but . . . but I can't do anything about
In addition, one of the central studies that demonstrates female su it."
ority in decoding facial expressions of emotion (Hall 1978) has A number of studies that use the cultural logic of engendered emotion
reanalyzed and shown to account for less than 4 percent of the vari on physiological differences to account for emotional ones than
l'ocus less
between individuals in facial expression recognition skills (Deaux L on universal functions and roles. In particular, they draw on the notion
cited in Shields 1987). of women's reproductive role and the nurturing role and emotions that
Despite its obvious problems, this account of the evolution of nupposedly naturally accompany it. From ethological bonding theory
expression identification is a story with some power, as it draws on (llowlby 1969) to some schools of feminism (e.g., Ruddick 1980), focus
trenched cultural narratives about women, motherhood, children, ls placed on the natural or inevitable emotional concomitants of moth-
love. Here, the firgt premise is that women are more attuned to emotiofl erlrood (rather than fatherhood), including particularly the positive
in themselves and others. Unlike the premenstrual syndrome studieg* emotions of love, caring, and attachment. Bowlby follows the prevailing
however, female emotionality is celebrated here, with emotions taki cultural emphasis on women's emotional qualities when he focuses on
on their positive sense of the interpersonally engaged, the unalienated thc emotions of women and their children. He wants to explain the in-
Women's emotionality becomes a skill and an asset. It is significant tensity of the bond between mother and infant, and roots that explana-
the sociobiological account focuses on the use of that asset to detect tlon in an instinctual need for attachment in the infant and fear of sepa-
tress (rather than, for example, threat). Distress, of course, calls for nur'' ration. Feelings of love for the child on the part of the mother are
turance, whereas other facial expressions (in either infants or adulte) ttaturalized (cf. Scheper-Hughes 1985), and disastrous consequences are
might call for flight or defense, but only the former behavior is norma. ehronicled should the infant fail to receive sufficient quantities of mother
tive for women and mothers. love. These two facets of Bowlby's approach provide the carrot and stick

t
Engendered emotion 83
82 Catherine A. Lutz
men. Hochschild's ideas contribute to a breaking down of the dichot-
of natural instinct and psychological harm to the child as reasons omy of emotion and thought; they can also extend the notion of wom-
continued emphasis on the need for emotionality in women'
en's double day of domestic and wage labor as women are required to
Ruddick (L980), on the other hand, identifies "resilient good h
contribute both emotional and cognitive labor in both paid and unpaid
and cheerfulness," "attentive love," and "humility" as among the cen"
spheres. In this and other feminist analyses, gender and emotion are
tral features of maternal virtue that follow from (rather than
related through the relations of production. For Hochschild, emotion is
the task of parenting and, by frequent correlation, the task of being
a personal resource that women must self-exploit more than men. It
male. From these perspectives, women are more deeply embedded
nonetheless remains a psychophysical fact, socially manipulated, rather
relatiqq,ghips with others (with the mother-infant bond as the prima
than a discursive practice that constructs women as more emotional than
,t a tn" primary cause). This interpersonal engagement wi men.
"^rrnpit,
others is what produces emotion, which is here defined as responses
In sup, social science disciplines women and their psyches. It con-
others with whom one is involved. From the perspective of feminis
structs"ffibtion as an individual and intrapsychic phenomenon and evi-
male individualism is antithetical to the experience of emotion (see
clences tfie same concem as lay discourse with the emotionality of women
Chodorow 1978). * its frequency, its intensity, its virtues as an emblem of female gender
The differences between these two perspectives on mothering
identity, but most of all, its danger and implicitly the need for its control.
emotion are, of course, crucial. Bowlby-style bonding theory natura
the connection between women and affect through evolutionary
and is continuous with earlier theorizing about the elevated moral Personalization
of women achieved through their divinely assigned and natur I now return to the question of how these cultural notions about the
embedded mothering skills. Feminist theory most often idenlifies
cmotionality of women, articulated in scientific discourse, are related to
social division of labor rather than nature as the ultimate source of
everyday discourse. The rhetoric of control that we first looked at was
emotional differences. Interestingly, however, both kinds of
rhown to reflect, in multiple and complex ways, relations of power be-
on emotion elevate women (the first to a domestic pedestal, the
tween men and women, and to reflect them in ways that can be said, in
to seff-[steem and/or the ability to resist patriarchy) by focusing on
large measure, to reproduce the "emotional female." By looking closely
itive emotions such as love and by using "emotion" in its positive
of some more microscopic aspects of the interview talk, however, we
mantic sense of connection and disalienation'
c.rn see that gender differences are minimal, afact that may speak to the
Yet another version of the cultural view of women as emotional
gaps and fissures in the construction of a hegemonic discourse.
found in the Parsonian normative construction of family roles, in w
ln two of the series of interviews, people were asked, first, to describe
women are the "expressive expert" and men the "instrumental
recent experiences with each of several common emotions and, second,
(Parsons and Bales L955). These competencies are seen as an outc
Io talk about how they feel about their work and family lives. In an
the domestic-market spheres in which the genders differentially
nrralysis of a sample of 286 randomly selected interview statements that
ipate. Compare &ris notion, however, with the contradictory view
lirclude direct reference to emotions, I have focused on the degree to
iromen,s emotional impact on the family noted in the interview examl
which the statement "personalizes" the emotion experience - that is, on
and the child-rearing manual themes described earlier. The point m
a variety of ways emotions, even as they are discussed, can be distanced
be that women are expected to be experts in noticing and attending
from the self. It might be expected that women would use more person-
the emotional needs of others (also per Bowlby), not their own, w
elizing and immediate syntactic forms if they operate following the cul-
are rather objects of control or suPPression because they, unlike
tural model in which women are more emotionally expressive and have
emotions of other family members, are defined as dangerous'
e more emotional self-identity.
Hochschild,s (1983) important feminist revision of Parsons and
l'ersonalization, or a nondistancing discursive strategy, was indexed
scheme paints emotion less as a skill than as a form of labor. women
by four speech patterno (eee Table 4.1), which will now be discussed.
socially assigned a much heavier burden of emotional labor than
t

L
Engendered emotion 85
84 Catherine A. Lutz
1. The present tense rather than the past or conditional tense (e.9., "I
Table 4.1. Personalization in syntax gcf [or am] angry whenever someone talks to me that way" compared
with "I was very angry"), is used. Tense obviously does several things
'
-1.' Present Tense
;;i talks to me that way'"
to the meaning that audiences can make of a statement about emotion.
Srt lor amf angry whenever someone First, it can move the emotion experience farther away from or closer to
Others
"lwasvery angry." the self or another in time. Second, it can either generalize or particular-
lz,e the experience; the use of the present tense, for example, can often
2. Exoeriencer of the emotion discussed
Seif, as subiLct of emotion experience lnclude the implication that the emotion is habitually experienced by the
..Wmvery anxious about it'" Subiect. On both of these counts, the stereotype would lead us to expect
$elf, as obiect of emotion experience more use of the present tense by women speakers. In fact, there is no
"if r rnuiattg me an1ry jusi talking-about this'"
iemite, or gender un specified) differeqpe between male and female speakers in the interview sample in
- ;'t person"(male,
Other
tiforlrn*ur r"ry annoyed"with me foi going into that field'" the. us$Ethe present tense. If anything, men as a group make slightly
Unspecified (lrrsignilicantly) more use of it.
"Ii was a very strong feeling of hate'"
,;L"ain"tae'vetope8 a certiin amount of hate toward that individual 2, Another element of a personalizing strategy might include the use
because of the fact that he . . '" 9f syntactic patterns that more directly portray the speaker as the exper-
-'; - erhotion as an abstract entity go hand in hand' I would say'"
None l€ncer of the emotion. Statements were coded as portraying the exper-
W"U, f,ute and frustration usual$
" 'Love'would be, I think, a good citchall phrase because ' ' '" kncer as the self, as another person (male, female, or gender unspeci-
fred), or as leaving the experiencer unspecified (e.g., "lt was a very strong
3. Cause or elicitor of emotion
of hate" or "And that developed a certain amount of hate toward
Self
"Thev were anzrY at me." '{*' individual because of the fact that he . . .") or the emotion as an
"riu* nna of fifgled and made her even angrier'" entity with no particular experiencer (e.g., "Well, hate and frus-
Other person tion usually go hand in hand, I would say" or " 'Love' would be, I
"I hite herbecause she was mean enough to tell me that'"
"fm deathly afraid of dentists." nk, a good catchall phrase because . .").The category of self was
Event her broken down by whether the self was portrayed as subject or
"The most anxious moment I had ' was my first Performance with the
of the emotion experience (e.g., "l'm very anxious about it" com-
Choral SocietY."
"I hate going out unless I really have to'" red with "It's making me ar.gry just talking about this"). The belief in
Obiect s emotionality might lead to the expectation that women would
';He loves books. often portray the self (particularly the self as subject rather than
Unspecified
as the experiencer of emotion, whereas men would portray the
"Lots of little things are frustrating'"
"i riunittalk anyirore, i Jurt scrJaming to begin with' when I'm really as the experiencer or leave the latter ambiguous.
angry'" In thc interview sample, itis not significantly more common for women,
&
4. Negation the.ir discussions of emotion, to focus on the experiencing self as the
"I [br she] wasn't angry." versus the object of the emotion, nor is it more common for men
l€nve. the experiencer unspecified or abstract. In addition, neither
nor men are more likely to portray others as opposed to the self
the experiencer of the discussed emotion. Women and men speak
allke than differently in this sample when discussing the experien-
€f emotions.
Btatementa about emotion usually contain an implicit or explicit
, that is, they epecify the cauae (uaually by opecifying the object)
;
F'

86 Catherine A. Ltttz Engendered emotion 87

of the feeling. Personalizing strategies might include identification male and female styles of thinking, behavior, and speech - and then
either the self or, secondarily, another Person as the ultimate cause often found in self-fulfilling fashion. The absence of differences is more
the emotion (rather than the use of syntactic patterns that obscure or rignificant given the syntactic nature of the evidence examined; Shiba-
to identify the cause). Statements were coded as portraying the cause tttoto (1987) has concluded that gender differences that are not a re-
either the self (e.g., "They were angry at me" or "I just kind of sponse to audience expectations about particular gender identities are
and made her even angrier"), another Person (".8., I hate her more likely to be found in syntactic patterns of use because they are
she was mean enough to tell me that" ot "l'ir,:. deathly afraid of typically outside of our awareness and hence of our easy manipulation,
tists':), an event (e.g., "The most anxious moment I had . . . was my unlike semantic patterns such as those having to do with the notion of
perJofinance with the . . . Choral Society" or "I hate going out unless I "control" we examined earlier.
haie'to";, an object (e.g., "He loves books"), or as leaving the
unspecified (e.g., "Lots of little things ate frustrating" or "I can't talk .w'
Concludqn
more, I start screaming to begin with, when I'm really angry") (cf.
manoff 1983). Given the associations between gender and affect I Itr a[ sotieties, body disorders - which emotion is considered to be in
earlier, we might expect that women more than men would see this society - become crucial indicators of problems with social control
people as intimately involved in their own emotion experience .lrld, as such, are more likely to occur or emerge in a discourse concern-
themselves as evoking emotion in others, rather than seeing events hg social subordinates. Foucault has made the claim that power creates
triggering emotion in themselves or failing to specify a cause. The stxuality and its disciplining; similarly, it can be said to create emotion-
strategy can be associated with the view of emotion as nonsensical,' Elity. The cultural construction of women's emotion can thus be viewed
rational, or without ascertainable cause. In fact, there are no signi flot as the repression or suppression of emotion in men (as many lay-
gender differences in the use of personal versus impersonal cfrrsal prople, therapists, and other commentators argue) but as the creation
bution, nor do women use self versus other attributions more than of cmotion in women. Because emotion is constructed as relatively cha-
ptic, irrational, and antisocial, its existence vindicates authority and le-
4. Finally, a number of statements about emotion in the in
are essentially denials of emotion in the self or the other (e.g., "I [or gltimates the need for control. By association with the female, it vindi-
wasn't arrgry").The stereotype might lead us to expect more eates the distinction between and hierarchy of men and women. And
in generalfrommen andmore negationof particularkinds of female- tht, cultural logic connecting women and emotion corresponds to and
emotions (which include most emotions except anger) by men and lhores up the walls between the spheres of private, intimate (and emo-
male-linked emotions (notably anger) by women. Here again, thrnal) relations in the (ideologically) female domain of the family and
and men's speech are indistinguishable in terms of the proportion ptrblic, formal (and rational) relations in the primarily male domain of
emotion states that are negated as they are discussed. tht' marketplace.
The absence of extensive differences might be attributed to the s Itubin has remarked of sexuality that "There are historical periods in
nature of the pegple interviewed, all of whom agreed beforehand to which [it] is more sharply contested and more overtly politicized"
with a stranger about emotion. The results are consistent, however, 119U4:267). Emotionality has the same historical dynamism, with shifting
a study of gender differences in emotion language use by Shi gender relations often appearing to be at the root of both academic and
(1983), who did a similar analysis of the tape-recorded natural by struggles over how emotion is to be defined and evaluated.l0In other
sations of a number of American college students and married words, the contemporary dominant discourse on emotions - and partic-
and found few differences in male and female conversations that Ulnrly the view that they are irrational and to be controlled - helps con-
cluded reference to emotions.e The results are also consistent with itruct but does not wholly determine women/s discourse; there is an
trend in studies of psychological and linguistic sex differences in Att(.mpt to recast the association of women with emotion in an alterna-
eral, which have tended to show far fewer differences than flve feminist voice,
both expected - on the basis of cultural stereotypes about Feminist treatmente of the queatlon of cmotlon (e.9,, Hochechild 1983;
r j

88 Catherine A. Lutz Engendered emotion 89

propositional networks underlying the sensibility of sentence order, it is


]agger have tended to portray emotions not as chaos but as a
1"987)
irosiiUte to draw inferences about the kinds of models individuals
are using
course on problems. Some have contested both the irrationality and br, perhaps more aptly, to draw inferences about the kinds of inferences
passivity of feelings by arguing that emotions may involve the listeners can make ibout what the speaker has left unsaid but likely wants
understood.
cation of problems in women's lives and are therefore political. Talk
3. There are 180 instances in those parts of the women's transcripts analyzed
anger, for example, can be interpreted as an attempt to identify the so far, and 85 instances in the men's, with each set of transcripts being of
istence of inappropriate restraint or injustice. Sadness is a di approximately equal length.
the problem of loss, fear on that of danger. By extension, talk about I hive found Wirotard'J (1985) analysis of the nature of hegemonic and
oppositional forms of language use very productive in formulating what I
control of emotions would be, in this feminist discourse, talk about have to say here.
suppibssion of public acknowledgment of problems. The emotiona 5. Martin 1rr87; has examined the American discourses on reProduction and
mSic might then be seen not simply as a mythic construction on the wome.,is bodies and has rigorously uncovered the contradiction between
a vietrv of uterine contractions during childbirth as involuntary and a view
of some arbitrary cultural dualism but as an outcome of the fact of tht.woman as in fact in control of the labor Process. The women she
women occupy an objectively more problematic position than does inte'Wwed about their birth experiences spoke very similar$.to the women
white, upper-class, Northern European, older man who is the descfilied in this chapter aboutlheir sense of control over the physical P-Jo-
cess and over their cries of pain and pleasure during labor and birth. She
exemplar par excellence of cool, emotionless rationality. According
notes a class difference, hotiever, with middle-class women speaking with
feminist analysis, whether or not women express their problems more approval of control than working-class women. We might then ex-
are emotional) more than men, those women's audiences may pect min also to express more concem with and-approval of control of
message that is an amalgam of the orthodox view and its feminist bmotion, which is not the case here. This is certainly a problem worthy of
more study, particularly a delineation of what kinds of control of which
testation: "We (those) women are dangerously close to erupting domains afpilar to emerge from what kinds of experience within hierar-
emotionality/pointing to a problem/moving toward a social critique. chical systems.
Ackno*tedgment of one's emotionality may mean very different things to
female and*male audiences. Women may announce to each other shared
identity and solidarity, while asserting difference, submission, or defiance
Notes when making similar statements to men.
An ealier version of this chapter was presented on the panel "Emotion Abu-LughodYs (1986) study of the Awlad 'Ali represents the most detailed
Discourse" at the annual meetings of the American Anthropological Assc and elo{uent example of irow, in another cultural system,_ the particular
'emotions
sreatly fron
Chicaso, November 18-22, 1987. The draft has benefited greatly
tion. Chicago,
tion, from kinds of allocated to and voiced by women articulate with other
comments of Lila Abu-Lughod and Steven Feld. The research on which aspects of their ideological and social structural positions.
chapter is based has been conducted with grants from the State Universib Tliis group of studies obviously follows in the tradition of centuries of ex-
New York Foundation and the National Instltute of Mental Health and with pert ixplanations of hysteria. Although there have been many- versions of
help of many people. Kathryn Beach, Robin Brown, Paula Bienenfeld, q ihe expianation (such is one nineteenth-century account that diag-nosed its
Walter Komorowski assisted in interviewing and transcription, and the exp origini as an empty womb and a childhood where the restraint of emotion
analytic work of Angela Carroll and Marion Pratt helped give the chapter wai not taught lSmlth-RosenberglgT2)), they have been organized around
form. the connectibn between female physiology and mood.
1. The actual p{ocess by which these models of gender and emotion are Shimanoff (1983) found that male and female speakers did not differ in the
quired is a ffiscinating but unexplored question. We might expect thr number of affect words they used, in the tense, valence, or source (similar
includes, in part, the child's reasoning from the culturally assigned to the notion of "elicitor" uied here) of statements about emotion' She did
ity and control of the male teacher to a lack of emotion, the latter find, however, that males made more reference to their own emotions than
already having been learned to require "strength" and "control" to m to those of other people when compared with females.
The resurgence oi int-erest in emotion in the late 1970s and 1980s across the
- in other words, to generalize from the dominant position of males social sciences may in part be the result of the feminist movement's reva-
presumed lack of emotion (a process that might also have occurred in
teachers' views of themselves). lorization of all things traditionally associated with women (Margaret Tra-
2. The method used in looking at the transcripts draws on recent wick, personal communication). Changing Sender relations may also be at
ments in the "cognitive" study of cultural meaning. These focus on the the robt of the reinvigoration of a long-standing Western discourse on the
of extended and relatively natural conversations for the cultural knr value of emotional eipression; the current debate pits expressionists, for
or cultural models (Holland and Quinn 1987) evident, if not always whom healthy emotions are vented ones, against those who would dismiss
itly stated, in them. By looking at such things as syntax, metaphor, or the latter as '7self'lndulgent" or "immature." This debate no doubt draws

t
j

F i

90 Catherine A. Lutz Engendered emotion 91,

ina complex way, in each concrete context in which it occurs, on the in Morocco. In H. Medick and D. Sabean, eds., lnterest and Emotion. Carn-
der ideologies and conflicts of the individual participants. bridge: Cambridge University Press, pp. 103-28.
Martin, Emily. 1987. The Ideology of Reproduction: The Reproduction of Ideol-
ogy. Paper presented to the Upstate New York Feminist Scholars'Network,
References . September.
$untagu, Ashley, ed. 1978. The Learning of Non-Aggressioz. Oxford: Oxford Uni-
Abu-Lughod, Lila. 1986. Veiled Sentiments: Honor and Poetry in a Bedouin versiw Press.
Berkeley: University of California Press. Far*,,nr, Talcott, and Robert Bales. 1955. Famity, Socialization, and lnteraction Pro-
Archer, John, and Barbara Lloyd. 1985. Sex and Gender. Cambridge: ('(fss. Glencoe, IL: Free Press.
University Press. rldo, Michelle Z- \984. Toward an Anthropology of Self and Feeling. In
Asad" Talal, ed. 1973. Anthropology and the Colonial Encounter. New Y ll. Shweder and R. LeVine, eds., Culture Theory: Essays on Mind, SeIf, and
rltiihities Press. l.motion. Cambridge: Cambridge University Press, pp. 137-57.
Ba6thuk, Wayne, Raymond Hames, and Ross Thompson. 1985. Sex rldo, Renato. 1977. The Rhetoric of Control: Ilongots Viewed as Natural Ban-
in the Recognition of Infant Facial Expressions of Emotion: The l dits ryrd Wild Indians. In B. Babcock, ed., The Rettersible Woild: Symbolic ln-
Caretaker Hypothesis. Ethology and Sociobiology 6:89-101. ucrsiffi in Art and Society. Ithaca, NY: Cornell University Press, pp. 240-57 .
Bimbaum, D. A., T. A. Nosanchuck, and W. L. Croll. 1980. Children's ln, G1ffie. 1984. Thinking Sex: Notes for a Radical Theory of the Politics of
types About Sex Differences in Emotionallty. Sex Roles 6:435-43. 1 Sexudlity. In Carol S. Vance, ed., Pleasure and Danger: Exploring Female Sex-
Bowlby, ]ohn. 1959. Attachment and Loss, Vol. 1. London: Hogarth Press. r
Cancian, Francesca. 1987. Looe in America: Geniler and Self-Deoelopment.
- uality. Boston: Routledge & Kegan Paul.
lUddick, Sara. 1980. Maternal Thinking. Feminist Studics 6:70-96.
bridge: Cambridge University Press. Snr;rrort, E. E. 1981. Cognitive Psychology as Ideology. American Psychologist
Chodorow, Nancy. 1978. The Reproduction of Mothering. Berkeley: U 36:73043.
California Press. fthepcr-Hughes, Nancy. 1985. Culture, Scarcity, and Maternal Thinking. Efhos
Deaux, K. 1984. From Individual Differences to Social Categories: Analysio l3:291-i17.
Decade's Research on Gender. American Psychologist 39:10!-16. ftheper-Hughes, Nancy, and Margaret Lock. 1987. The Mindful Body: A Pro-
Fausto-Sterling, Anne. 1985. Myths of GenderiBiolo{ical Theories of Womefl [cgomenon to Future Work in Medical Anthropology. Medical Anthropology
Mez. New York: Basic Books. 4qi
Quarterly l:,641..
iexuality,
Foucault, Michel. 1980. The History of Sexuality, Vol. 1. New York
York: Vintage.
V r , irmoto, lanet. 1987. The Womanly Woman: Manipulation of Stereotypical
Gottlieb, Alma. 1987. American PMS:: A Mute Voice. '--' Paper r--"-"'-" at'
- -r-^ presr
presented nnd Nonstereotypical Features of fapanese Female Speech. In S. Philips,
annual meetings of the American Anthropological Association, Chica S. Steele, and C. Tanz, eds., Language, Gender, anil Sex in Comparatiae Per-
Haan,n, Norma, Robert Bellah, Paul Rabinow, and William Sullivan. 1983. ipactiTre. Cambridge: Cambridge University Press, pp. 2549.
Science as Moral Inquiry. New York: Columbia University Press. Ehlr,lds, Stephanie A. 1987. Women, Men and the Dilemma of Emotion. In
Hall, J. L978. Gender Effects in Decoding Nonverbal Cues. Psychological I'. Shaver and C. Hendrick, eds., Sex and Gender. Newbury Park, CA: Sage
85:845-75. l'trblications, pp. 229-50.
Haraway, Donna. 1986. Primatology Is Politics by Other Means. In Ruth Ehlrnunoff, Susan. 1983. The Role of Gender in Linguistic References to Emotive
ed., Feminist Approaches to Science. New York: Pergamon Press, pp. I
States. Communication Quarterly 30:17 4-9.
Hochschild, Arlie. 1983. The Managed Heart: Commercialization of Human Ith-llosenberg, Carroll. 1972.The Hysterical Woman: Roles and Role Conflict
Berkeley: University of California Press. i n 1 9th-Century America. Social Resmrch 39:652-78.

Holland, Dorothy, and Naomi Quinn, eds. 1987. Cultural Models in languagl , Anne. L985- Perceptions of Protest: Defining the Dangerous in Colonial
Thought. Cambridge: Cambridge University Press. American Ethnologist 12:642-58.
u matra. American
Sumatra.
lagger, Alison. 1.987. Love and Knowledge: Emotion as an Epistemic Ree( Strnthcrn, Marilyn. 1980. No Nature, No Culture: The Hagen Case. In Carol
for Feminists.&4s. in possession of author. Department of Philosophyr MacCormack and Marilyn Strathern, eds., Nature, Culture, and Gender. Cam-
versity of Cincinnati. hridge: Cambridge University Press, pp. 17+-222.
Jordanova, L. I. 1980. Natural Facts: A Historical Perspective on Sciencg umrig, Michael. 1984. Culture of Terror - Space of Death. Roger Casemenls
Sexuality. In Carol MacCormack and Marilyn
SexualiW. Marilyn Strathern, eds., Nature, l)utumayo Report and the Explanation of Torture. Comparatiae Studies in So-
ture, and Gender. Cambridge: Cambridge University
Univr Press, pp, 42-59. dtlu and Histont 26:467-97.
Kirkpatrick, |ohn. 1987. Representing the Self as 'You' in American Dis< Whntteji, Mariannd. 1986. Taking Feminist Science to the Classroom: Where Do
per presented at the annual meetinss
Paper meetings of the American A Anthropo Wd Go from Here? In Ruth Bleier, ed., Feminist Approaches to Science. New
Association, Chicago. York: Pergamon Press, pp. 181-90.
Lutz, Catherine. 1988. Unnatural Emotions: Eaeryday Sentiments on a Ml Wlfllamc, Raymond. 1977. Marxism and Literatute. Oxford: Oxford University Press.
Atoll and Their Challenge to Western Theory. Chicago: University of Wrxrlnrd, lCittrryn. 1985. Language Variation and Cultural Hegemony: Toward
Press. an IntegraHon of Sociolinguistic and Social Theory, Amqican Ethnologist L2738-
Maher, Vanessa. 1984. Poesession and Dispoeseseion: Maternity and

t
I'opographies of the self 93

are also anchored in the New Testament, where, for the first time in
rn history, a major normative claim was made about the separa-
5. Topographies of the self: of act and actor, intention and action, "inner states" and "outer
praise and emotion in Hindu India " Although these Christian ideas have undergone many changes
apawned many conflicting intellectual offspring, Western discourse
continued to build on this normative break. What has gradually
ARJUN APPADURAI :d is a complicated repertoire of discourses about the "individual,"
"telf," and "personalityi'with the last term serving as the most
+u- tt, and most technical, bridge between the first two.
An cxploration of this Judeo-Christian story and its many twists and
gv-fr the last two millennia is beyond the scope of this chapter' It
tigq4:ver, seem possible to identify an elementary Western topog-
Topographies of the self ofthe "person" (a relatively neutral term for the anthropologist)
This chapter explores a specific modality of verbal interaction in which the biologically anterior "self" (where the intertwined pro-
lndia - praise - in order to construct an argument about the v of ontogeny and phylogeny play themselves out), through the
of the relationship between language, feelings, and the of the trajectory of "personality development," becomes a
the self in human societies. In contemplating emotional life in though idiosyncratic moral unit, the "individual." This view
turally specific setting, it seems important to note that emotions i ktnd of master trope within which more specialized discourses (re-
linguistic life and a public and political status that frequently en therapeutic, and legal) contest each other on matters of detail.
formulaic modes of expression. Yet, it also seems that emoti8ns, topography is anchored in a spatial image of layers, of which the
other phenomena, appear to have a basis in embodied experience, ive bedrock is seen as simultaneously the simplest, the most gen-
inclining us to see them as rooted in some elementary biophysi nnd the most directly tied to the somatic side of personality. Thus,
ertoire that is both limited and universal. To ignore completely Western metatheories of personality are doomed to remain paro-
ond aspect of emotion is to run the risk of deconstructing emotion since they ask cross-cultural questions without any consciousness
gether as a distinctive phenomenon to be investigated. In the thcir constraining master image itself needs to be interrogated be-
that follows, I will try to show that praise in Hindu India is a form ecrious comparative questions can be asked (Lutz, this volume).
erned by the regularities in performance of a culture-specific I would suggest that such topographies of the sel{, whether or not
ventional activity, like many other linguistic activities. But I will nre articulated in elaborate cultural discourses (and metadis-
to show that praise in Hindu India is, in Bourdieu's sense, a ,n), are variable cultural phenomena. We need to deepen our un-
improvisatory practice (Bourdieu 1977), which depends on a pa nding of. this variation if we are to retain the force of the insight
topography of thgelf that underlies its public expressions. plified throughout this volume) that emotions are discursive pub-
This topography, properly understood, leads us to a second forms whose special Power does indeed draw on embodied experi-
Much recent discussion, several chapters in this volume, and my , without implying any parsimoniously describable universal bio-
previous work (Appadurai 1985) take the view that our current substrate. But this demonstration cannot, at least for the
sense about intention and expression, about "real feelings" as o ropologist, be primarily experimental or deductive. It has to be eth-
to "voiced sentiments," about superficial statements that conceal phic, and anchor itself in interpretations of existing forms of per-
and "inner" emotional states is, as is so often the case, merely oill in particular cultural settings, as a way of exemplifying alter-
bodied doxa misrepresented as general theories about the topographies of the self. In analyzing the pragmatics of praise in
between affect and expression (Abu-Lughod 1986; Lutz 1988). Such India, I seek to sketch the outlines of one such culturally specific
are not only part of emotion talk in the contemporary Anglophone y. In particular, I ehall argue that praiee ia not a matter of. ditect

92
j
L

94 Arjun Appadurai Topographies of the self 95


communication between the "inner" states of the relevant persons, l,tter being seen as mortal vessels of universal sovereignty. The praise
involves the public negotiation of certain gestures and responses. of kings thus is a perfectly acceptable and highly developed cultural mode
such negotiation is successful, it creates a "community of sentiment" n'd is to be seen in various types of panegyric and eulogy, starting with
volving the emotional participation of the praiser, the one who is prai the earliest literatures. such praise, which also tends to trr" formulaic
and the audience of the act of praise. Praise is therefore that set of nnd the hyperbolic, can sometimes take the form of an extended open-
ulated, improvisatory practices that is one route to the creation of lngportion of inscriptions recording a royal act of generosity.
munities of sentiment in Hindu India. These written prefaces (sanskrit: praldsti) have not been extensively
rtudied, since historians have generally been more interested in the acts
&:""
The"gineatogy of praise in the Indic context or events recorded in the main body of the inscriptions.l The rhetoric
of
tlrcse praise texts is complex and is historically and regionally variable,
As is the case in Christianity and Islam, so in the Hindu world the but soqftgeneral points can be made. In most of these ;praise pr"faces,,
adigmatic or prototypic act of praise is the praise of the divine. In thc cenft5l objective is to identify and grorify the reignin[ sovereign and,
Hindu case, as opposed to that of Islam and Christianity, there is a when not the same as the sovereign, the donor whose act is being re-
bination of subordination and intimacy, which Babb (1986) has c'rded. The identification is usually done through genealogical and
called "hierarchical intimary." The praise of the many incamated clrronological statements, often involving chains oi rru*"r, that stretch
of divinity in Hinduism is a central, highly developed part of the track to divine ancestors. The genealogical claims are themselves a form
process. Sfitra (a term whose linguistic features I shall discuss of glorification. A typical strategy is to present an extended series of
refers to a ritualized and usually textualized recitation of praise. It nnmes, some of which are titles (airufu). The Dravidian word airufu,
major part of the relationship of devotee to divinity in all the r4ajor which in historical contexts generally means 'title' or'emblem,, has the
ditions of Hinduism. In some contemporary contexts where SariStrit ln()re general etymological connotations of honor, pedigree, panegyric,
left a strong imprint on the regional languages, the technical, ritual ettd praise. All these names, many of which have identifiable meanings,
cept of sfitra imbues more everyday usages. Thus an everyday T arc themselves expressions of positive qualities, potencies, or achieve-
word fer praise is stottiram. tncnts.
Although this is not the place for an extended history of the of course, there are many cultures in which praise is an onomastic
and practice of sfitra in Hinduism, four things are worth noting principle. In south India, royal names or titles, often self-conferred, are
its cultural construction. First, it makes praise a ritual offering. c,mplex words or sets of words that refer to specific recent acts of valor,
it puts praise into a formulaic and an aesthetic framework. Third, gt'lrerositSr, or piety. These complex, self-conferred titles are both rec-
main device is description (often through hyperbole) of the rtrcls and advertisements of royal achievements. when the donor whose
qualities of the god or goddess in question. Fourth, praise in the I
ar'tion is being recorded in an inscription is the same as the king who is
is associated with the public expression of the emotional bonds of Lrt'ing praised, these praiasti constitute a culturally appropriate formula
otee and deity. ffi1tra thus is a mode of praise that is ritual, for self-praise for those who rule. Even here it is possiut" tt ut there are
hyperbolic, and emotional. This ritual mode underlies a very large :ar'rcd models, such as the self-celebrating epiphany at the end of the
of the corpus of devotional (bhakti) poetry in both North and South Blrngnaad Gita, in which Krishna stages an enormous spectacle
dia. What is important about this cultural paradigm is that it i of his
powers and forms to Arjuna.
both interaction and assessment: Thus praise involves intimacy 'l'hese pra1dsti forms tell us that
the pragmatics of praise in the Indic
the subject and the object of praise, while also implying a certain eontext has something to do with boasting, boosting, competing, elevat-
tance. I now tum to some further implications of the involvement ing, and inflating, In the Indo-Aryan languages, there is a semlntic dis-
praise with the attitude of a devotee to divinity. tlttction between those forms deriving from the sanskrit root stu (whose
As with other ritual and rhetorical forms in Hindu India, there ll Vt'dic context placea it explicitly in a rituar and hymnal context) and
praise a logical concatenation of attitudes to divinity and to royalty, thrrst' forms (such aa the Hindi and Gujarati procansa) deriving from the

t
F a

96 Arjun Appadurai Topographies of the self 97

Sanskrit 6ansa, which has the far more secular connotations of move from highly styhzed, fixed forms of expression to everyday ones,
mendation, applause, fame, and glory. In the standard dictionary of we enter a cultural field of greater ambiguity.
Dravidian languages, of the twenty-one relevant entries, five are
mantically neutral, but of the remainder, twelve have very secular Praise and dependence
tones (again having to do with flattery, boasting, elevating, and p
cizing), and only four are clearly linked to adoration, worship, One consequence of the resilience of the value placed on the formulaic
and so forth (Burrow and Emeneau 1961). praise of superiors is that flattery is a prominent part of everyday public
It is, of course, dangerous to infer contemporary semantic r bchavior in India. In this respect, praise (or, more exactly, flattery) is
frora;diich etymological patterns, especially given the complicated one of the standard means not simply to mark hierarchy but to mark
hist#its into various dictionary entries. Neverthelessr, dependence. Flattery is typically part of the rhetoric surrounding a "big
"o-p."rrl"d
seems reasonable to argue that, at least in the Indo-Aryan la !tlan," {vhether the big man is a politician, a film star, a business mag-
there are fzpo clusters of meaning surrounding acts of praise: a pri nate, dqilflage headman, or a charismatic religious leader. Of course,
cluster derived from Vedic ritual and hymnal practice and a euch a ihetorical stance occurs in many societies, but it has a particular
cluster pertaining to praise in everyday life. In the Dravidian case, flrrvor and structure in Hindu India.
relative priority of these two clusters is less clear, and it is concei Earlier, I observed that the stdtra combines the texture of ritual, aes-
that the ritual or worship-oriented sense of praise is a product of thetic, hyperbolic, and emotional modes of expression. Public flattery is
action with North Indian traditions and practices, and that the perhaps most easily seen in the politics of contemporary India. The praise
classical Tamil universe is oriented to praise as panegyric rather t hr.aped upon Indira Gandhi by her followers, themselves often leaders
praise as worship. gl substance in their own regions and contexts, frequently shocked out-
But it is also clear, in both the Indo-Aryan and the Dravidian gide observers. Such flattery was an important part of public demonstra-
that there is no sharp break between the two semantic clusters and tions of dependence and adoration (touching her feet, asking her bless-
each derives some of its pragmatic force from the other. The two lngs on various enterprises, anticipating her moods, and so forth) that
fields converge in the praise (or self-praise) of kings, where the eould easily be construed as simple, self-interested groveling. But in In-
matic lines between praise as adoration, praise as honor, praise as dla, these demonstrations are an important part of a tradition of adora-
ship, praise as flattery, and praise as naming are difficult to draw tion of and interaction with the glory of the superior.
197q.2 This pragmatic sense of praise could be extended from What we see in political behavior can also be seen in less formal con-
all patrons, as in the following typical benediction appearing at the texts involving superiors and their clienteles, where flattery takes on
clusion of each section of an eleventh-century jaina text from Delhil lmmense proportions. Today, given how far removed many of these
which the poet praises his patron, a ]aina merchant (Cohen 1979:1.21", €()ntexts are from the authority of ritual and the splendor of the court,
"May Sri Nattala, whose fame, appearing like the moon, shines on lndians themselves have developed a certain ambivalence about public
earth; because of g6hom bards have become seekers of the wish- flattery and the forms of dependence that it implies. One sign of this
tree; by whom speech is uttered free from error; who is beloved of Embivalence is the colloquial Hindustani term drnmdra (literally, 'spoon'),3
people, illustrious, pious, and incomparable, be victorious!" which is applied in a derogatory way to those dependents of any celeb-
The patron $ajamana) of any ritualized or aesthetic activity is the rlty who specialize in echoing, praising, and transmitting his or her views
uational incarnation of the god-king. Thus, the ideological and wlthout any whiff of critical independence. Yet, such hangers-on, whose
forms associated with the praise of patrons (which we have from a *tle function is to provide a continuous flow of praise, are an integral
variety of paintings, texts, sculptures, and inscriptions associated Part of most groups surrounding celebrities in India. It may appear at
Hindu "art") can be seen as paradigmatic of the attitudes expected first sight that this "groupie" phenomenon is an accompaniment of star-
inferiors toward their superiors in all domains of life. But in making dom and power in many cultural settings. But I would argue that its

L
j'
F ."

98 Arjun Appadurai Topographies of the self 99

special diacritic in Hindu India is its unconscious anchoring in the of reputation building have an excellent chance of benefiting the trans-
logic of worship. mitter as much as the subject of the act. Also true of such small com-
For those who lack the privilege of constant contact with a su munities of interaction is the distortion of messages that occurs, and
there is another, less direct mode of praise. This mode involves p eredit for a reputation-building act for the publicist may not be regis-
the good deeds, capabilities, powers, and reputation of the tcred appropriately. But the hope that your superior will hear of your
somebody else, outside the physical presence of the superior. praise, and thus think better of you, is not crudely manipulative (as with
not, therefore, a direct act of adoration, it is not flattery either. elternative topographies of the self, which permit both acts and judg-
applie_g to major rulers, deities, saints, patrons, and "stars" of ments of hypocrisy),abut a mixture of adoration and expectation of re-
sort" Ifihvolves singing their praises, not so much in the form of ward that characterizes Hindu ritual generally. Praise of the superior is,
andltialiUes as in narrative forms, focusing on specific achievem therefore, part of a complex series of acts of mutual benefit that charac-
the superior to which the speaker is able to testify. It is a form of terize t{e ethos of Hindu worship itself.
tation building (for both the speaker and the big person spoken of) But isn do superiors construe their own capability to praise their
cultural logic has hardly been explored. lnferioriland dependents? In general, direct praise of inferiors and de-
In this indirect form of praise, too, there is a special bond betweerl pcndents is as uncommon as the praise of superiors is common. There
speaker and.the object of praise, for its narratives frequently put ere stories in the Hindu tradition in which deities express their admira-
speaker in a privileged relationship to the superior. This privilege tlon for particularly remarkable devotees and acts of devotion. But in
two forms. The first involves the action worthy of praise, which everyday life, it is not in good taste for the superior to sing the praises
the speaker as a worthy recipient of the superior's favor. Stories of dependents directly. But here we need to discriminate among several
the miracles worked by a saint, the gifts given by a philan dtuations.
boon granted by a specific deity, the financial help given at a ln the domestic situation of most Hindu households/ parents do not
crisis by a politician, the secret lifesaving act by a filmstar - all prnise their children directly, for this would be seen as inviting the free-
usually involve the speaker as a direct beneficiary. The second f,oating malevolence of the "evn eye," a topic on which I shall short$
privilep is epistemological: It marks the speaker as someone who havc more to say. But although it is considered important for parents to
something special and who has the privilege of passing it on. deprccate the virtues and abilities of their children to outsiders, such
and transmission, insofar as they underlie such indirect forms of But$iders are certainly not prevented from seeing or hearing the children
of a superior, are fundamental parts of the cultural construction of' $enrunstrate their skills or accomplishments in drawing, singing,
utation in contemporary India. In this process, praise, witness, ghoolwork, or domestic arts. Such demonstrations raise difficult dilem-
and narrative come together. This configuration, which today fian for guests or observers, who have to calculate their responses care-
all sorts of celebrities, stars, and big men and women in the fully. To withhold all praise is tantamount to an insult. To praise too
relies on spontaneous narratives of praise that continue (as in the rectly and in too fulsome a manner raises the specter of the evil eye.
to generate texts:6toriesz po€filS, songs, and prayers of praise. The solution generally is to separate the act from the actor and to praise
of praise here is not direct flattery but publicizing, which is accomplishment as if it were distinct from the actor. Thus, praising
to the emotional satisfaction of the superior but to the increment product or performance (the painting, the cooked dish, the song,
or her following. rlance) rather than the performer is perfectly appropriate. This stric-
Sometimes, in the making of such reputations, the teller of thc applies only in the physical presence of the object of praise, where
(if you will, the publicist) tells it in the full knowledge that word elangers of the evil eye are greatest. When the child is physically not
or her acts will eventually return to the superior who is the nt, there is greater latitude, although even here the problem of the
such praise. Such strategies of circumlocution are typical of small eye is not entirely absent.
munities (families, neighborhoods, devotional groups, offices, From the point of view of the guest or visitor in a domestic setting,
groups and so forth) where circles of communication are finite, and y loooe conotrainto on the pralse of children are much more
Topographies of the self 101
100 Arjun Appadurai
for its practice. Begging is a
stringent in regard to wives, and women more generally. Especially |lilce for it and a recognized cultural idiom
ieg,itimate (even if devalued) activity for two reasons. First, it partakes
adult males who are not regularly involved with the household,
of women, whether direct or indirect, whether of person or of ol the emotional and ritual ethos of worship, with the beggar playing
mance, must be extremely muted. Even thanks for hospitality, whi tlrc role of the worshipper. Second, in its open admission of the radical
in any case complicated (Appadurai 1985), are rendered to the male clt,pendence of the beggar on the goodwill of others for his subsistence,
rather than to the female hostess and cook. If praise is ever directed ll r-.arries some of the moral authority of asceticism, and in many cases
,,l,cBgars,, and ,,ascetics" are not clearly distinct categories. The wide-
the women of the household, it has to be brief, casual, and
orien(gd, rather than oriented to the female or females of the house. a1,r.iud fear of both beggars and ascetics is tied up with the magical Power

thgJdtter case, not only is there the danger of the evil eye, but also th ril words, in which blessing and cursing are two sides of the same coin
are the sexual implications of praise, which endanger conventi (lt'ter van der Veer, personal communication)'
modesty in the South and of honor, reinforced by the impact of nlthEugh there is a great deal of regional and situational variation in
in the North. Women are generally expected to be invisible h,,w E6p;ing is practiced and perceived, the central verbal tactic of beg-
good things and good works, for which they cannot easily be pra g,rrs is'Ib ales s aid.praise their human targets while asking for food or for
The evil eye, which I have now mentioned several times, n,,,,.r"y. This practice could be labeled "coercive subordination," for in
applies to vulnerable creatures, who are dependent on the nurtu ttlessing and praising their (potential) benefactors, beggars seek to trap
and protection of superiors. The best examples of this category are tht,rn in the cultural implications of their roles as superiors, that is, in
dren and domestic animals such as cattle. When such vulnerable thil obligation to be generous. The coercive element is also expressed in
16r, opei way in which beggars reveal their
desire to benefit from the
tures are especially beautiful, innocent, weak, or fortunate, they are
grtirter good fortune of their fellows. In India, when desire is so open,
as natural magnets of envy, desire, and ill will, with fuzzy Mrt
c'an ,nali.e be far behind? Indeed, in crowded public situations,
beggars
between these sentiments. Beautiful children, healthy cattle, a
€arr often be seen (and heard) to express their resentment of
rebuffs by
daughters, and devoted sons are typical targets of these emotions.
the evil eye is invoked largely in those situations where the mrrrnbling abuses and curses (the inverse of praise and blessings), par-
nology of emotions is ambiguous, where acclaim might mask envy, tk,ularly when they have been rebuffed in a cruel or insulting manner.
itude might disguise desire, and congratulation might conceal ill lk,gging is an exaggerated and intensified enactment of forms of de-
There is no explicit indigenous theory about the relationship p*rr.lence and types of interaction that are widely institutionalized in
emotional ambiguity and the evil eye, but it seems to be a itrrlian society. The parallels between benefactors and gods, the coercive
interpretive link. Inrl reciprocal implications of praise, the link between praise and the
David Pocock has noted, in a West Indian rural setting, that the derrirc for the good things in life, the fine line between desire and envy,
eye is a culturally organized interpretation of envy and that, tlre hidden threat of abuse beneath the external profession of praise, are
cally, it is most likely to be imputed in those situations where all lirctors on which beggars depend to exploit the emotional ambiguity
who should be ftual are not so in fact" (1973:39). Pocock also that links praise - through envy - to the evil eye. The ambiguity lies in
(and
that najar (the evil gaze) is not to be feared, in a hierarchical the lact thit praise can be a celebration of dependence o/ a subtle
€uercive) complaint about it. Beggars, who rightly see themselves as
between those whose status is different and clearly defined. But
reason to suppose that this view cannot be generalized, and that havirrg little (transactionally) to lose, given the speed, number, and im-
link between emotional ambiguity and the evil eye is a more subtlc pern,irality of their encounters, frequently rely on this ambiguity in their
look threatening while they utter paeans
To make this case, it is worthwhile to examine begging as a social lrat.tiees. Thus, some beggars
nomenon in India. bf Pririse; or they exaggerate the mechanical and impersonal formalities
Begging, like many other activities in India, is an organized thcir verbal routines; or they violate canons of physical contact by
tional activity. Although it is increasingly a phenomenon of pa kruelring their targets on their arms or thighs rather than on their feet or
tion and proletarianization in urban India, there is a recognized cu n6t.lt illl. ln playing with this gestural and verbal repertoire while utter-
j.

F ..

102 Arjun Appadurai Topographies of the self 103

ing formulas of praise, beggars deliberately enact the ambiguity of (even social identities are disparate); or when contractual
if their larger
emotional relations to their potential benefactors. Thus their 3rrur,g"-"r,ts are being negotiated; or, most revealingly, in the transac-
mance underlines the fact that praise is not a matter of direct Hons and negotiations that accompany marriage. I shall use this last con-
nication between the inner emotional states of the parties involved hxt to explo-re the link between praise, assessment, and the world of
of a publicly understood code for the negotiation of expectations !hings.
obligations. This public play of affect involves a distinct topograp Mirriage negotiations in Hindu India, especially among upper-caste
the self, and a related aesthetics of transaction, that can best be
Etoups that
have elaborate dowry practices and other related gift trans-
ciated by considering two further arenas in India: the world of a ir,io.,r tied up with affinity, belong to a class of events that I have else-
and fhE.world of appreciation. tVhere called "tournaments of value" (Appadurai
1986). They are agon-
Btftbefore going on to discuss these two arenas, it is worth rem ,$tic encounters, outside the sphere of normal economic practice, where
that there are many contexts in which praise is not seen as a ve tio4s take placeover objects of prime
"*:, **,tll}::llf:
the sentiments of the evil eye: The praise of superiors, the *r* i*a&rtions for day-to-day economic well-being. In the particular
inanimate objects, and the praise of nonintimate inferiors are &ru.nrmeH of value constituted
&r*rrru*"ffi by Brahman Hindu marriage in Tamil
Tamil
of such contexts. Thus, if you praise as ir good worker a farm fttuth India, praise is an important rhetorical component of the status
marriage
who works for another man, you might do so in a patronizingway, @ntests that characterize both the negotiations that precede
(Appadurai 198L). Praise has special
this is not regarded as endangering the worker. We might conclude, {May 1986) and the marriage itself
when objects or persons are seen as relatively invulnerable to envy, polilical overtones in the context of marriage because marriage always
sire, or ill will, or when the emotions appropriate to them are Lrvolves the effort of bride givers (in tndia, by definition, of lower sta-
relatively unambiguous, the evil eye is not thought of as a likely hr) to dignify their temporary deference to the bride-receiving family.
paniment of acts of praise. But the praise of objects brings tf$ to *rit, tt u giftr tt rt accompany marriage and the individuals involved in
distinction between praise as adoration and praise as appraisal. ftr. marriage are subject to very complicated mutual assessment.
ln the first place, a subtle battle of assessment is involved on the very
Bfrit occasion when members of the prospective groom's
family visit the
Apprai$al and the world of things the entire rhe-
horne of the potential bride. On these delicate occasions,
as appraisal
I have so far discussed praise as interaction in two modalities: one torical posture of the bridegroom's party can be summed up
ritual in orientation and whose paradigm is the worship of the (or e,viluation, or assessment). The home of the bride's family is as-
and one that has to do with applause, publicity, and reputation, i€ssed, as €lle the things on display within it, including furniture, wall
jewelry
seems typical of the followers of leaders, big men, and big women. cecorations, and utensils. More personal possessions, especially
also on
there is another axis of discrimination in regard to praise, and here tnd clothing worn by the female members of the household, are
must distinguish praise as a mode of interaction from praise as a but not least, the prospective bride and her
dlnplay for issessment. Last
The exchange of words
of assessment. Byolooking at praise between putative ( lecomplishments are displayed for assessment.
in the context of ni:arriage) and between unequals, I shall-equals
show that gn ru.h occasions is a very subtle commentary on this traffic in things,
reveals some of the ambiguities inherent in all public codificationa lmages, and sense imPressions.
affect.s lriir the potential bride's family, there is usually a maximum effort to
In India, praise as assessment is most often seen in the rela ht their house, their possessions, and their carefully decorated daughter
tread a thin
between equals. It arises in contexts where praise as adoration and lpeuk for themselvei. When they themselves speak, they
bride
as applause are equally out of place. It is also prone to be iinn between humility (which is expected of them as prospective
especially in regard to things. Praise as assessment is likely to occuf The latter is accomplished subtly by re-
$ve-rs) and self-advertisement.
contexts with a commercial component, either generically, as in the &rrtng to well-placed relatives, influential friends, and their own eco-
texts of shops, bazaars, and markets, where buyers and sellers are iomic stability. More important, they make comments from time to time

t
:
F :

104 Arjun Appadurai Topographies of the self 1.05

Here
that allude to the culinary (or musical or housekeeping) skills of lum can, and often do, precede more pointed evaluative inquiries.
daughter or of the women of the family in general. On the bri praise is a gentle, polite way of introducing potentially impolite queries'
side, there is a greater sense of command and assurance and a ' lt is r"orih noting that the more "petsonal" the things in question, the
brazen tendency to assess their environment, induding the human more careful the acts of praise and assessment must be.
What can easily
in it. Questions about material possessions set the stage for later, be said or asked about a radio, a cat, a bicycle, or
a cigarette lighter can
nosed negotiations about what the bride's parents will do for leus easily be said or asked about a piece of iewelry, a piece of clothing'
daughter. Even those positive comments made by the bridegroom's gr a utensil in active use. In these latter cases, acts of praise begin to
in praise of the material lives of their hosts are meant to be acts of ipproach praise of the owner and to raise the possibility of envy or de-
praltgl, signs of evaluation. lire airectla at this person. This marks the boundary at which acts of
ht the marriage itself, a good deal turns on the material display of Praise cannot lightllbe
made or construed' Yet, praise of any thing or
dower: the utensils, the jewelry, and the dothing provided by the as a display
Dossessicn (whether made as a grudging act of appraisal or
family for the couple and especially for their own daughter. bf *ur,r#anthusiasm) is preferable to direct praise of a person, unless
sessment becomes an instrument of status politics in the most 'pers&
g5f is an obvious superior. With superiors, praising their Posses-
manner. Members of the bridegroom's family often do not hesi llons would most likely be taken amiss, since it implies an inappropriate
assess openly the possible weight of the metal in utensils, the hea Cvaluative stance on the part of the inferior. By extension, the positive
of silk saris, and the possible cost of specific pieces of jewelry. BHsessment of the material possessions of an inferior
would be regarded
actions, and the often deprecatory verbal accompaniments to t elther as patronizing or as sarcastic, but in either case as not in
very good
the negative side of praise as assessment and are frequently the taste.
point of bitterness between the two families, which can result in ln general, the praise of things exemplifies one mode of praise, which
term discord and sometimes in the abuse or death of the bfftde.6 le cvaluative, whereas the praise of persons exemplifies
another mode,
public, detailed, and intensive assessment of material objects, as have qualities that link
which is interactive. But since persons and things
just as
the social reproduction of groups, has other general implications thcm, every act of evaluation has something interactive about it,
Breclg:nridge 1986) that lie outside the scope of this chapter. evcry interactive act of praise is the product of some sort of appraisal.
It is sufficient to note that praise as assessment of material This situation both resembles and contrasts with that of more egalitarian
is most intense in the context of dowry-centered marriages, but it Focieties, where there may be explicit sanctions against creating
emo-
also be seen in a wide range of social interactions between persons (Cath-
tkrns of awe or envy in others by showing them one's possessions
are, roughly speaking, in the same social class. Thus, whenever erine Lutz, personal communication). In both cases, what
is important
and social groups reach the economic level where part of their to note is thit praise is governed by regularities of discourse and embod-
hold inventory of possessions is clearly for display and not simply lecl strategie, tf i.,tu.u.tion that do not assume anything critical about
subsistence, their social interaction acquires the dimension of a thc ,,inner" states of the actors. What the relationship then is between
More than elseyhere in the world, in Hindu South Asia it is cu expressions and emotions can best be seen by turning to the domain
of
acceptable to sc$utinize, and preferably to handle physically, the aesthetics and Performance.
possessions of the person you are visiting; to inquire after the "
biographies" (Kopytoff 1986) of their possessions; and, if possiblel
ascertain their original cost. There is also an equally acceptable set r{ppreciation and Performance
devices to fend off such questions, to give vague replies, and to ,t,lrroughout this chapter, I have noted that praise, in both formalized
offering information to the questioner. But the form of the and eiryday settingi in India, has something of the formulaic
and the
is culturally sanctioned, and neither side can easily take offense. hyperbolic about it. To the observer-analyst, it often appears exa8ger-
too praise plays an active role, for statements of admiration and aied, formal, and unrelated to the emotional interior of the person who
r :
."

1,06 Arjun Appadurai L07


Topographies of the self

praises. This problem of the emotional authenticity of praise (sr his or her own emotions in a particular formulaic, publicly
lrvine, this volume) is best tackled by looking at the major area rrstood, and impersonal way. The object is to create a chain of com-
lndians have reflected on emotion and appreciation: the ications in feeling, not by unmediated empathy between the-emo-
aesthetic theory of rasa, the master theory of aesthetics in Hindu rl
,,interiors,, of specific individuals but by recourse to a shared, and
As every student of Indian poetics and aesthetics is aware, ively fixed set, oi public gestures. The creation of shared emotions
per-
peculiarly elusive concept, partly because its assumptions are thus unyoked from the emotional authenticity of any particular
ferent from those of Western common sense regarding the n,s feelings. Praise in Hindu India partakes of this set of assumptions
between feeling, gesture, and performance. yet the theory of :rning performance and feeling. Praise is measured by the "com-
tairfrs,clues not just to Indian ethnopoetics but also to praise in itv ofientiment" it evokes and creates, and not by the authenticity
lil€and the topography of self that underlies it. Before I use the theiink between the private (or idiosyncratic) emotions of the praiser
of rasa to illuminate the pragmatics of praise, I will outline its rd the opject of his or her Praise.
in the Indian leamed tradition. ln theti$ssical theory of rasa, particularly as formulated by the elev-
rth-cen#ry Kashmiri Saiva theorist Abhinavagupta, the relation
lndian aestheticians have singled out eight feelings (bhaaa) that be-
sons experience in their lives: love, mirth, grief, energy, terror, lcn emotion and aesthetics takes an elaborate form and is expressed
anger/ and wonder (see Brenneis, this volume). In the poetic lnteresting ethnopsychological terms (Gerow L974:22C-1). The theory
and of
each of these is transformed into a corresponding mood (rasa), a based on intuition, on the generalization of character, of event'
Gerow as a statement of
alwed, impersonal feeling capable of being understood by other , and has been described by Edwin
in similar states. In drama, these moods are expressed in a antirealism. In Abhinavagupta's formulation, rasa is a transcen-
understood set of gestures, and both the dramatic performance Lmode of apprehending the work of art, to which normal modes
of
critical analysis involve the appraisal of these gestures. ffie reness are obstacles. Gerow (1974:224) has this to say about the "in-
quences of this appraisal for dramatic performance are neatly of emotion" in Abhinavagupta's view of the aesthetic resPonse:
by A. K. Ramanujan (1974:117-1,8):
The entire d.rama has now been translated from the theater to the
The actor, as in a Stanislavsky school, must study the audience; the theater is no longer "obiect,"but pretext for the in-
stances and expressions that are functions and reflections terior play whose success is nothing but a state of mind' cleverly
mood, even glandular secretions of tears and contractions erroked through suggestion, realized as those latent aspects of the
solar plexus: one feels grief because he weeps,
;oy because hh. audience's emotional being that are the common and recurrent
glows and his eyes dilate. It is a form of physical imagining, heritage of mankind. These aspects are implied by and present in
the story of the village idiot who found the missing donkd everyemotionalcircumstance,everyconcreteemotionalsituation,
imagining where he would go if he were a donkey. The but are never, in ordinary life, grasped in themselves' apart from
produces tears and gestures; cannot the gestures their specific determinants. It is the function of the play, of linguis-
emotion?,&rd the reader and the spectator in his tum goes tic art, so to free the very conditions of emotional life; and it is
the incipient gestures and tensions in himself: the mood r precisely in this sense that the rasa is not a concrete emotion fuhat:a)'
condition in which the reader or spectator reconstitutes hit Lut rather the inversion of an emotion; the specific determinants
analogous private, incommunicable, and forgotten feelings intt of the emotion (place, time, circumstance, etc') are so cast as to
impersonal expression. They are transmuted into the mood. appear themselvls as function of the latent emotional state, and
he enjoys, and thus he can enjoy, for example, grief. are generalized.
state of
Let me gloss this discussion of the portrayal of emotion in Ind Gerow also points out that the theory of rasa, which posits a
etics by noting that it has a special set of pragmatic assumptionr. { consciousness more real than the work of art itself, has clear philosoph-
key assumption is that the actor evokes certain feerings in thL viewcr leal linke to the theory of levels of reality contained in the
philosophy of

L
108 Arjun Appadurai Topographies of the self 109

Advaita Vedanta. Rnsa rr.ay be said to anticipate and prefigure because it is rooted in the general understanding of praise as a key to
(spiritual liberation). But later in the history of Hindu reciprocity between superiors and inferiors; partly because of the hidden
particularly in the writings of the Bengali Vaisnavas of the sixteenth threat of the evil eye; and partly because "communities of sentiment"
tury, a curious reversal occurs whereby m1kia "has not only become can be created in India by the skillful orchestration of specific gestural
ideal open to all men, recast as the perfection of the most human elements without reference to the "inner feelings" of the actors. The fact
relations, love, but this new 'emotional' transcendence, bhakti, has that both the beggar's perfgrmance and his audience's resPonse appear
come the essence of rasa" (Gerow 1974:226). mechanical and unsentimental does not disprove my interpretation, any
Gerow's statement of both the classical theory of rasa and the in more than the ritualized collective wailing accompanying death in many
sion Efemotion on which it is founded - and the reversal of the I societies suggests that the participants are in no way sentimentally in-
in Bilhgali Vaisnavism, where mOk$a itself. is seen as an emotional volved. In fact, the more deeply shared the ethos and the code, the more
an etemal version of the experience of rasa in art - provides the basis mattert"{fact, or "mechantical," the performance can afford to be.
making a plausible interpretive link between the vagaries of rasa tn thffitiation of "communities of sentiment," standardized verbal
in Indian history and the significance of praise in Hindu life. To and gestilral forms are used, and there is no assumption of any corre-
tiate this intuition, I will return to the topic of begging, which was spondence between the words and gestures and the internal emotional
cussed'earlier in relation to the evil eye. This large interpretive leap world of the "actor." What matters are the emotional effects of praise,
rasa theory to the cultural logic of begging in India is justified which, when it is properly "performed," creates a generalized mood of
begging is a highly organized performance and, in its most odoration or admiration or wonder that unites the one who praises, the
forms, has a large audience. Beggars are often consummate actors object of this praise, and the audience, if there is one. At the same time
have developed their own style - composed of verbal, gestural, and (and here the bhakti connection is relevant), a special emotional bond is
esthetic elements - for approaching their potential benefactors. ereated with the object of praise itself. But the emotional landscape im-
though the benefactors are not always willing participants in these plied by such acts of praise is not built on the idiosyncratic, biographical,
performances, both positive and negative responses to beggars have rxperiential , "irtrrer" feelings of Western common sense. It is consti-
ceptable cultural forms. Since the "interaction rituals" (Goffman tuted of the emotional effects created by the public negotiation of the
involved in begging are public and highly orchestrated, they are words and gestures of praise.
completely removed from the arena of aesthetic performance. The Of course, the verbal and gestural forms of praise in everyday life do
ond justification for using begging as an everyday extension of the rrot have the aesthetic rigor or the ritual predictability of art or worship.
derlying logic of rasa theory is that begging, too, involves the Thus, the emotional impact of specific acts of praise in ordinary life can
negotiation of emotional expressions. often be weak, ambiguous, or, for lack of a clear frame, unfocused. Still,
The beggar's praise is no more intended to represent his "inner" when praise is directed at a benevolent superior, something of the aes-
ings to his audience than are the gestures of the actor on the thctic and emotional communion implied by the concept of rasa is, I
stage. As in ras&theor/, what beggars do, by drawing on a suspect, present. To the extent that formulaic public praise of superiors
negotiable set of expressions, is to draw the audience (and their ln considered credible and pleasurable, it is probably due to this cultural
tial benefactor, who is sometimes the sole member of the audience) ettnception of the construction of emotional and aesthetic satisfaction.
a "community of sentiment" whose pragmatic consequence (if the Praise in Hindu India is one aspect of the critical evaluation of texts,
formance is skillful) is that the benefactor bestows some favor on fit'rsons, and deeds, and it is an important part. Such criticism is in-
beggar. Why can we not simply say that the beggar is a flatterer terrded to deepen, rather than to abort, the social bonds between the
plays upon the ego of his target? Because begging in India usually Eubiect and the object of criticism. It suggests an idea of emotional and
volves not only a tale of woe (this is hardly a unique feature of Eesthetic communion between audience, artifact, and ultimate reality,
in India) but also a fairly elaborate performance of "coercive which differs from those assumed and created by most varieties of post-
tion." Such coercive subordination, when it does work, does so Renaissance Weetern critical theory. A great deal more accounts for the

t
:

110 Arjun Appadurai Topographies of the self 111

Indian attitude toward traditional objects, whether human or tions of eighth- to sixteenth-century,south India, parts of which were suc-
than can be discussed here. Ideas about originality and authorship, cessively rinder Pallava, Chola,.and Viiayanagara rule'
1, I am grateful to Kirin Narayan for drawing this source to-my attention.
mentary versus criticism, and the inherent prestige of the past all l 3, I havE neither heard nor read anything about the etymology of this term. I
need to be carefully worked out to account for the Indian i,citical,, susfect that it implies the idea o("teeding-.'praise to one's superiol as well
per. The idea that praise is a complex devotional, evaluative, and u, dh" d"rogatorj, even polluting, identificition of one human being with
an instrumint put into another human being's mouth'
active act should be an important aspect of such an account.
ihe subject of hypocrisy, on which I am currently {o.rki1g, brings together
still, it is important to recall that praise in ordinary life is not a ,.l .rr*bi,. of corirpticated cross-cultural issues involving the topography
of
matter of communion, devotion, appreciation, and adoration. We the self, the problLms of staging and,representation, and the authenticity of
flgeen already that praise can, in relations with equals and inferiors, public expreisions, which cannot fully be engaged here'
l*' --more ambiguous rhetorical device, refl fhe etyniological roots of the Englisl, yotl "praise" are clearly economic
ecting envy, inappropriate and iniply asiessment in a framework that involves commerce, calculation,
or anger. But even in these cases, it is still a form that is *nd e*inar,ge. This original sense of the English word "praise".has been
having effects on the extralinguistic world. Directed at the wroig .firt in most"contempo."ary uses, but there is a dimension of praise that is
f,idt unrelated to it, even in contemporary Western practice'
or as a direct expression of inappropriate ,,personal,, emotions, p i"iefer here to the phenomenon of "Uriae burning," which, especially in
dangerous, duplicitous, even damaging. But whether it makes or North India, has dra'wn a great deal of attention both in the,press and among
the bonds between the speaker, the audience, and its referent, fcminist groups. It is widely agreed that these deaths are often the end point
,if u t.ui"Etory of abuse of #oilen by their husband's families, fueled by the
never h neutral, descriptive act. It always has the ability to affect
t rnassivL inflition of dowry demand-s in contemporary urban India'
create things in the world. The everyday challenge for Hindus is tg
sess correctly the meanings of the forms of praise they witness
predict prudently the results of the forms of praise they produce.
Praise is thus not a matter of linking the emotional "dfteriors,, of
leferences
tors by breaking through the public veil of language, of gesture,
communication. It is rather one of the varieties of improvisatory Ahil-l,ughod, Llla. 1986. Veiled sentiments: Honor and Poetry in a Bedouin society.
Bcrkeley: University of California Press.
t[at, in Hindu India, can create sentimental bonds q,rit" ird"pender Alrl,rrclurai,'Arjun. 198f. Gastro-Politics in Hindu South Asia. American Ethnolo'
the,"real" feelings of the persons involved. Such bonds are part of gist 8, 494-511..
politics of everyday life, and such politics is cultural and not lrl'dS. Gratitude as a Social Mode in South India. Ethos 13:236-45'
1rt86. Commodities and the Politics of value. In A. Appadurai, ed., The social
since its messages and its media are publicly expressed, construed,
l,ifa of Things. New York: Cambridge Universitl Pre^ss,.pp.' 3!3'--.
Fsl,b,'t,. i. 13lE6. nedemptiae Encountels: Three Modern Styles in the
appraised. Hindu Tradi-
lftrrr. Berkeley: University of California Press'
Bhur.,ii, A. lgZO'. Pilgrimag6 Sites in Indian Civilization. In J. W. Elder, ed',
(ltwters in Indian"Ciailil,ation. Dubuque: KendaUHunt, pp' 83-126'
Notes E,r,,r,ti"l, pierre. 1977. Outline ot' a Thiory of Practice. Cambridge: Cambridge
[,lniversity Press.
This chapJgl$ dedicated to Thomas Zwicker, who died in Ahmedabad on Bre,,k,.,lridje,'Carot. 1985. The Social Use of Everyday Ob-jects in. Hindu South
tober 29) 1985. It was first delivered at a conference on the emotions in =-'
lndia.ih'L. Ctranara,ed.,EssaysinHonorof Pipul'layakar. Delhi:AgamKala
I organized by Pauline Kolenda and owen Lynch in December 19g5. I am lrrrtkashan, pp. 57-66.
to all ure
Lu du the parucrpanrs
participants at tnat
that conterence foi their suggestions, and espe
conference tor Eurri,*, i., u,",a't*eneau, M. B. 1961,. A Draaidian Etymological Dictionary. Ox-
Frederique Marglin, whoserved_as
who served as commentator on thllchapt";. i r* uflo
thlf chapt"i. i;; forcl: Clarendon,
to- Owen Lynch for his detailed critique of an earlier ar"ff Fi"uln
i;h"fri gof,uir,-if .-iiZS. The Pasanahacariu of Sridhara: An Introduction, Edition and
t Abu-Lughod, Richard Cohen, Dilip Gaonkar, Catherine Lutz, andireter ilirnnnlation of the Apabrahmsa Text. Ph.D. dissertation, University of
and'ireter vtn
vtn
veer for helping me to sharpen what was previously
Veer. previouslv
lry a ratner eror'r
diffirsc at
rather dlttuse
diffuse !k.nrruylvania.
1. The word praiasti is derived from the Sanskrit 6ansa,
Sansa, which in turn
tur €erow, ti.t'win, 1974. The Rasa Theory of Abhinavagupta and Its.Applications'
for theterm prasansa,.a s.tan+Jd lexeme for,praise, in several modern lrr li, Dimock, cd., Thc Literatures'of lndia: An lnlroiluctittn. chicago: univer-
rlly of Chicago Press, PP.216-27.
rnan, tlrvingi'1967. Intiiaction Rituale, New Yorkl Anchor Books'
112 Arjun Appidurat
Kopltoff, Iqo1, .l-98!.!he Cultural Biography of Things, In A. Appadurai, ed,,
The social Life of Things. New York: Cah6riage University preii, pp. 6Lgl,'
Lutz, Catherine. 1.988. l-Innatural Emotions: Eoeryday sentiments on a iiicroneshn
Atoll and Their Challenge to western Theory. Chicago: University of chicago 6. Shared and solitary sentiments:
the discourse of friendship , play,
Press.
May, L..1986. Arrlnging Marriages: Negotiation and Decision-Making in
India. Ph.D. dissertation, University of Pennsylvania.
Pocock, D. 1973. Mind, Body and Wealth.- O:xford: Bisil Blackwell.
Ramanujan, A. K. 1974. Indian Poetics: An Overview. In E. Dimock, ed.,
and anger in Bhatgaon
Lrteratures India: An lntroduction. Chicago: University of Chicago
pp. 115-18. -of
DONALD BRENNEIS

ru"
i.(--
This chapter is intended to complement - in a highly tentative way -
current n$search on relationships between discourse and emotions. I have
been grfficl to the perspective I will propose by the people with whom
I have wirked: rural, Hindi-speaking Fiji Indians. I will provide a brief
ethnographic account of some of the notions through which they inform
their own understandings of talk and sentiment. The methodological
core of my approach is a consideration of a local rhetoric and aesthetics:
In what terms do Bhatgaon villagers interpret, evaluate, and shape their
experience, and how are such practices enacted through discourse? My
discussion will draw primarily on village men's accounts and behavior.
This is not, howevet, due solely to my being a male ethnographer in a
somewhat sex-segregated community. Rather, as I will later argue, these
local theories of emotion and experience - and the ways in which they
take shape in ongoing discourse - are critical elements in the definition
and politics of gender in Bhatgaon.
Before moving to a discussion of the Fiji Indian case and its possible
implications, I want to highlight two areas in which recent ethnopsy-
chological research and work in the study of language and emotion or
affect have led to important reformulations that are particularly relevant
to my argument. First, it has become increasingly evident that under-
standings of the locus and genesis of emotional experience vary consid-
& erably across cultures. In contrast to the usual Western notions of the
locus of emotion being within the individual, for example, in much of
the Pacific, "emotion words . . . [are] statements about the relationship
between a person and an evenf' (Lutz 1982:113). As Myers (1979) and
others have demonstrated, there is often a critical relational dimension
in local theories of the emotions. Indeed, "feelings" often provide a so-
cial rather than an individual idiom, a way of commenting not so much
on oneself as on oneself in relation to others.

113
u 4t...,..-t'f'
ll4 Donald Breniirels
L986;Lulz1987)' speak-
mes8ate. In other caoes (see, e'8', Abu'Lughod
they employ in pursuing
Second,wehaveamuchsubtlerunderstandingofeomcofthecritical
A number of stud' ers, intentions and the discursive strategies
ways in which language and emotion are entwined' thoseendshavebeenhighlighted;therhetoricalmotivationofemo-
of emotion vocabularies in dif' In both kinds of study' how-
ies have focused on seLantic dimensions iio^"f p"*ormances is a central concern'
with local terms'
ferent cultures (e.g., Ltttz L982; Myers 1979)' Starnng u,r"r, ti individual speaker - whether
expressive or calculating - occu-
the varied ways in which discourse " attention has been paid to the
such studies illumine considerably pies center stage. Littie overt or detailed
Other studies have fo' how they might affect
apparently concemed with emotion operates' of how those messages are received and
other communicative de' I""1U."t
cused more on those pragmatic, prosodlc' and their auditors.
vicesconventionallyasso"ciatedwiththeexpressionofemotioninpartic. Letmetakeseveralsomewhatextremeexamplesfromtheethnomu.
ularcommunities.Howdospeakersevincetheirfeelings?Throughwhat possible problems with a solely
(Mithun 1982' cited in Irvine sicological literature to suggest some
"plops, squeaks, croaks, sighs and m9ans" sender-focused model' Qawwaliis a
genre of Sufi Muslim devotional music
feelings conveyed? The 1983)' Professionalqawmli mu-
1982),,te,*mention only a f"ew possibilities' ate common in pakistan uri mdiu (Qureshi
is not limited to speakers; scholars as
;""t,# of expressiv" d""i""t have considered sty- sicians use a range of performance
devices not to exPress their own sense
&rr"rr" as Osgood (1960) and Besnier (1g}g' n'd') t"figiq"t ecstisy Uut to draw it forth from members
of the congrega-
un-
listic feature. of tfr" written encoding of emotion' "f
;;;.^fiF*h"^r"1.'"' can succeed at this only when they remain
and culturally valued ex-
Irvine(1982)providesaparticularlyhelpfulconsiderationoforalex. moved rfiovers, their music evoking powerful
di'"""ion ofihe descriptive and analytical
pressive devices, as well u, u Similariy' performers at the Kaluli gisalo are
questionstheysuggest.Oneissueraisedbylrvine(1982:34)isthatof
;;;;"", in their listeners' their effect on their audience's emotional re-
what she calls'"sincerity'" How do we' whether
as outside observers or ;il;ny concerned withmusical, and performative features properly co-
a true account of a sponses; "when textual,
community members, know that we are hearing
ui"r"u, someone will be moved to tears"
(Feld L982:215)'
in societies are clearly not limited to
,p"uk".', ieelings? The issue becomes especially complicated
,

are often used to The evocative possibilities of performance


like the Wolof, where intermediaries such as griots music. Shifting back to language' it is not
always speakers Yh: T" moved;
sitSearby'
op."rr their clients' outrage while those for whom they speak rather, they often tJengender in others particular kinds of ex-
expressionless (Irvine L982:40)' ";;P;
perience. Self-expression is noiat issue'
but the discourse is no less
Boththeemotionvocabularyliteratureandthatconcernedwithex- "emotional."Aconcernwithspeakers'theirrhetoricalintentions'and
pressive-devices raise questions -aboyt the
representation of emotion in they deptoy emotional idi-
emo- the communicative devices through which
iir.orrri. In the lexica^l approach, the focus is on discourceasabout
emotion' oms to pursue their goals, however'-must
be complemented by atten-
literatu'e focuses more on discourse theories and understandings
tion. The expressive tion to their audien"J, una to those local
thatis,astheenactmentorrePresentationofparticularfeelings.Inthis thatinformtheirresponses.Ifemotionalrhetoricworks,howdoesthis
a more general premise active inter-
consideration of the multiple roles of discourse, happen? Audiences are not solely targets;
they are' rather'
is about some-
of the ethnography of speaking is evident: Language pr"t"tt, critics, and resPondents'
,wftitft
and con-
iii.,g, doe, s[rr,Lthir,g, and is something in itself; the content performers are themselves dispassionate
-

related' Such instances - in


auciof emotional communication are integrally buttheirlistenersmoved-areextremecases.Iusethemtosuggestthat,
affecting' and con-
in most emotion discourse, language is expressive'
{s
influencing to some degree
Emotion discourse as relational practice stitutive, displaying a speaker'i state and
is often a relational dimension
that of his or f,", uidi""te' |ust as there
Acentralfeatureoftheprecedingstudiesisthattheirexplicitfocusis also, I suggest, might emotion
ol tlre senders to local conceptions of the emotions, so
almost exclusively on questions oiemotional expression' process'
dir.orrr" tretpfutly be conceptualized as an interactional their own
an implicit assumP-
of affect-lad"n *"rrugls. There often seems to be Even if an audience ,"-uir,, silent, members
construct ex-
between the inter-
tion that the problemiti" i,,'"' lie in the relationship perienceofwhatoccurs,attributingparticularStates,degreesofsincer-
nalstateoftheoriginal,,fee|e{,andtheformandcontentofhisorher
Shared and solitary sentiments
tt7
L16 Donald Brenneis
privileged position. Emotions are not solely "abotJt" the world; they also
ity, and so on, to speakers or other senders. Radway's (19M) thoughtful
irelp cJnstitute it. Similarly, emotion discourse is not solely referential.
examination of women's practices in reading, evaluating, and respond-
ing to romance novels is a particularly striking example of the constitu-
It ii more than a vehicle for the representation of feeling and for the
comprehension of the world through emotionally framed schemata. A
tiJe work that audiences do. Similarly, several more ethnographic stud-
consideration of audience theories can help illumine the ways in which
ies (Duranti and Brenneis L986) demonstrate the very active roles that
emotion discourse not only argues but also is felt and catalyzes re-
audiences play in events as diverse as Zinacanteco gossip sessions and
sPonse.
California church services. A concem for audience resPonse is implicit
Iwouldsuggestthat,justasethnopsychologicalresearchhasillu-
in manptneatments of emotion discourse, as in Irvine's discussion of the
mined the cultural construction of emotion and experience, a considera-
"sincd#)' question mentioned eatlier, where the use of griots "may
tion of emotion discourse needs to take into account the social aesthetics
perhaps weaken the affective load" (1982:40)' How the "load" is evalu-
of parti4rlar communities as they bear upon questions of experience and
utua only be understood by taking an audience's or teceiver's Per-
"ur, into account. In many cases, audiences become more active, p".ro.tffia, expressiveness, and evocation' In some communities' a
spective
sender-fBcused model may at times be appropriate; in others - or at least
speaking as interlocutors, expressing their own feelings, subtly trans-
in some domains within those groups - evocation rather than expression
formingfpreceding speakers' constructions, or evincing their disbelief,
may be at the core of local theory. one would exPect, in fact, that in
sympathy, or aPPreciation.
" most societies a complex of notions would obtain. R t"ly, I expect, would
In a recent urti"tu (Brenneis 1987),I have suggested the notion of "so-
some relational dimension be absent'
cial aesthetics" as a rubric for those shared theories through which
One potentially valuable focus for such analysis is the communicative
members of the village audience evaluate the style, coherence, and ef'
event, lspecially those events that local people consider to be salient,
fectiveness of particular events. Such a social aesthetics delineat"t*ul-b"-t!
.or,r"qr"rrtial, or beautiful. As white suggests in his chapter in this vol-
at times inchoately, the standards to which a verbal performer is held
accountable (see Bauman1977; Hymes L975) and in terms of which a
,*u, i" need to learn considerably more about the "articulation of eth-
nopsychologies of emotion with social institutions." Communicative
particular rhetorical strategy either succeeds or fails. Such aesthetic the-
are indeed critical points for such articulation, as already ele-
tries of 6h"r".r." and beauty are linked to ethnopsychological notions "rr".,i, (1986)'
gantly shown by, among others, Rosaldo (1984) and Abu-Lughod
of personhood, emotion, expression, and experience, as well as to local - In the remainder of this chapter, I outline briefly some aspects of social
ways of making sense. They also, to quote Herzfeld, "correspond intel-
aesthetics and emotion theory in Bhatgaon. I am particularly concerned,
u6tly with local social theory, with indigenous ideas about meaning,
as Bhatgaon villagers themselves are, with the aesthetic, experiential,
u.,a *itft criteria of style, relevance and importance" (L985:xv)'
and emotional dimensions of communicative events'
such a social aesthetics informs the ways in which audience members
make sense of and experience others' performances. The efficacy of dis-
course - ritual, political, emotional - lies largely in how compelling its A Fiii Indian aesthetics
audience finds ii t&be; central to a local social aesthetic are the criteria
The system of social aesthetics in Bhatgaon has a number of distinctive
that underlie such determinations. This focus on "experienced" dimen-
elements. First, the underlying system focuses primarily on the organi-
sions of discourse further addresses a recurrent problem in anthropo-
zation of performance as social practice. As such, performance almost
logical considerations of emotion that Abu-Lughod points out in her
results in inevitabllinvolves more than one actor; one can rarely conduct emo-
ch-apter in this volume: an implicit cognitivist inclination that
,,assimilating emotion to thought." "Feeling" is represented primar- tional discourse by oneself. This premise underlies a common aesthetics
our
for verbal and musical events alike. A second distinctive element of
ily as a way of interpreting and understanding the world, even if an
Bhatgaon social aesthetics is that it makes an explicit link between the-
initiaUy ,rrp"itir,g ot Altho,rgh there is an important cognitive dimen-
"' orieJof performance and the language of emotion and experience' It is
sion in the notion of social aesthetics - and in the Bhatgaon villagers'
indeed difficult to seParate ethnopsychological from aesthetic notions;
model that informs it - it occupiee neither a clearly demarcated nor a

.f;G:
1L9
118 Donald Brenneis Shared and solitary sentiments

espe-
in articulating the bases of their enjoyment or appreciation of particular elements that instantiate it. An evening's Session of conversation,
events, villagers also articulate their sense of self and experience. cially when the talk turns to talanoa ('gossip'), is a prime occasion for
Third, aesthetic theory in Bhatgaon is internally complex and multi- such playfulness. Similarly, the raucous but ritually critical singing
of
valent. Although the means through which different experiences are ef- cautal hymns during tne uoti festival is a ritual event definitively linked
fected are relatively invariant, the characteristics of those contrasting ex- with tamashabhaw. The conventional features of tamashabhaw - those that
periences differ in systematic ways. Rather than embodyrng a particular engender its experience - include a fast interactional tempo, a partici-
focal principle, as, for example, honor in thirteenth-century Iceland put.-ory structure in which a number of people take part, and
a-system of
(Bauman 1986) or manhood in rural Crete (Herzfeld 1985), a Bhatgaon turn allocation among performers allowing for considerable overLap. When
social aesth$es holds that different orienting values come into play in such elements are present in an event, it can be "tead" as playful; the
different liinds of events. The Bhatgaon situation also differs from that stage is set for the experience of social pleasure. That it is social Pleasure,
of Egyptian Bedouins described by Abu-Lughod (1985), where two ,,dis- tha-t is, an exg:rien"" thut can be had only in the comPany of
and with
crepant discourses on emotion" expressing quite different sentiments the cooperath&Pf others, is again critical'
coexist. The contrasting experiences encompassed by Bhatgaon local the ihatgaffi aesthetics and the system of ethnopsychological theo-
theory, rather, are all included within the same broader discourse. ries intertwlned with it are social theories in two senses. First, they are
learn-
In Bhatgaon villagers' discussions of emotions, two features are sa- shared understandings, the emerging product of ongoing social
lient. First, the word in local Hindi for emotion (bhaw) is the same as that ing, negotiation, andhodification' Second, they include necessarily in-
for gesture or display. Bhatgaon social aesthetics explicitly links feeling teractive behavior for their enactment and experience' Important emo-
and display. Second, none of these "feelings" seem to be individually tional experiences in Bhatgaon are located in events themselves, especially
experienced, at least as people spoke of them. Bhaw ('feehngs') are not in the constellations of persons and performance styles conventionally
viewed as internal states. Rather, moods seem to be located in everiB associated with their different varieties'
themselves.
Bhaw is most frequently used in compound constructions in religious
Anger: the solitary sentiment
discourse, as in prembhaw (prem 'amity'), which carries the multidimen-
sional meaning of (1) a situation of interpersonal amity, (2) the display To suggest that only shared, socially constructed emotions are given
full
of the mutually respectful and amiable demeanor that embodies this am- value as bhaw is not to claim that Bhatgaon villagers do not recognize
ity, and (3) the experience of that state. Prembhaw is definitively associ- individual feeling. Personal emotional states - individualized, transient,
ated with the weekly meetings of religious groups and linked through situationally ,pu"ifi" - are devalued, but they are acknowledged obliquely.
those events with such perfornance genres as parbacan ('religious Typicalofsuctpassingsentimentsisanger,theHindiwordforwhich'
speeches') and bhajan knaoali ('kaoaali-style hymns'). It is also said to qiiro, ul"o*"urr, ,fist'. The link between feeling and gesture is evident
characterize the relatively rare pancayats, public occasions for the media- here in a particularly concrete iconic way'
tion of potentially conseluential conflicts, organized by the religious as- True bhaw are 'built' or 'made' (banana, knrna). In the religious talks
sociations. Moral didactiEsm - the willingness to teach and be taught - definitively linked with prembhaw ('amity'),. for example' speakers fre-
is a critical component of. prembhaw. Clearly defined solo tums, a focus quently use the phrase prembhaw banana cahiye ('we must build amity') in
on moral and spiritual improvement - on the message - and the willing- d"rcrifir,g the goals of the worship sessions. Similarly, an oral invitation
to drop b! a friend's house is often defined as a chance to'make'
tamash-
ness of others to attend to what an individual is saying or singing are
among the features encoding prembhaw and enabling its experience. It is abhaw-('fin'). On the other hand, qussa ('anger') - never recognized as
critical to note that these are features that cannot be enacted by an indi- qussabhaw-'happens'toone(qussalaqye)'Itisanindividualphenome-
vidual alone; the discourse of amity is necessarily interactional. non, the experience of which is implicitly considered to be beyond one's
Another frequently discussed bhaw is tamashabhaw, best translated as own control, although its expression may be restrained. The source of
a situation of playfulness or fun, the aseociated experience, and those qussa isconsidered to lie in specific situations. For example,
the discov-
..

120 Donald Brenneis Shnred and solitary sentiments L21,

ery of a neighbor's bullock browsing in one's rice field often leads to Such solitary sentiments as anger, however, are seen in Bhatgaon as
anget, as does a direct accusation of some wrongdoing on one's part' individually enacted. Such individual enactments are also relatively un-
Some persons are considered more likely to express qussa than others. ambiguous; a fist, whether literal or figurative, requires little interpreta-
One who is constitutionally Prone to manifest anger is considercd akara tion in Bhatgaon. Further, such interpretation is neither called for nor
admi ('hard man') or knra aurat ('hard woman'). Such persons are also allowed. Responding to anger is difficult and potentially dangerous. The
characterize d as sidha ('straight'); they are considered to be dangerously social-individual dimension is critical in sorting out richly elaborated
direct. Although qussa may'happen' to anyone, not all individuals will and downplayed emotions in Bhatgaon. Beyond this, however, the shared
reveal it. or individual nature of the discourse embodying these emotions is criti-
Levyft98a) has suggested that certain emotions may be "hypercog- cal to their recognition and experience.
nizedlin a culturei that is, to quote Ochs, they are "the objects of con-
siderable attention and knowledge . . . [and are] richly expressed within i"
Event, e@tion, and gender in Bhatgaon
the culture" (L986:254). Other emotions are, in contrast, "hypocog-
nized," receiving little attention. Although Levy's original focus was on Focusing*bn those communicative events in which bhaw hgwe critically
the relative salience and degree of elaboraticin of emotions within partic- is effective not only as a heuristic device or starting point for under-
ular cultures, his initial distinction has been refined in a variety of ways' standing emotion discourse in Bhatgaon. Such events are indeed the
Besnier (lg}g, n.d.;, for example, has shown how an emotion that may only contexts in which these elaborated, coperformed emotions are Pos-
be hypocognized in some media or performance genres may be quite sible. Here the intimate link between local theories of emotion and the
marked in others; letters home in Nukulaelae encode considerable affect forms of discourse associated with them is central. Prembhaw - the shared
that is noticeably lacking in face-to-face conversation and other oral amity that both promotes solidarity and redresses the damage caused
*&
practices. by individual anger - can be 'built' only through joint effort and, usu-
Clearly, such emotions as amity and playfulness, those recognized ally, only in particular kinds of events, especially those sponsored by
and constructed as bhaw in Bhatgaon, are hypercognized, the core of a religious associations. Although tamashabhaw can be enacted in a wider
rich complex of theories and practices. Anger, individually enacted and range of contexts, from public festivals such as Holito evenings at home,
expressed, is asocial and in some senses accidental. It is taken as hap- it also requires coperformance.
pening to people, some of whom may be constitutionally more susceP- Central to such events is their "participation structure" (Philips L982),
tible to it than others. Anger is hypocognized, recognized only in pass- the range of likely or appropriate performers and the roles they are ex-
ing comments, and not lexicalized as a salient and valued emotion, that pected to take; gender is a critical determinant of such structure in Bhat-
is, not treated as abhaw. gaon. Active participation and, in most instances, even audience mem-
There are two levels at which this particular distribution can be inter- bership in the occasions with which prembhaw is definitively associated
preted. One is in terms of the kinds of emotions themselves. It is critical are, by and large, restricted to males. Only among the reform Hindu
io note that the hygercognized emotions require joint performance and group - and then only at special meetings - are women expected to be
construction, whei6as such trivialized sentiments as anger are defined present or encouraged to speak. Gender is the basis of exclusion (see
as individual. Beyond this, however, there is no "social" and socially Foucault 1972:2'1.6; Lederman 19M:103), not onlyifrom these events but
valued variety of anger. Anger poses real dangers for villagers. That it from the highly valued emotions that can only occur in them as well.
is hypocognized and is not a subject for shared enactment fits well in a The participation structure of the events linked to tamashabhaar is some-
community where a tender and easily endangered egalitarian ideology what less restrictively ordered. During the Holi celebration, for example,
informs the daily life of adult males. both men and women are expected to participate, although only men
A second interpretive tack is to consider the discourse involved. Bftaar perform the raucous songs central to the festivities. Even domestic en-
demand shared performance; in their formal and stylistic characteristicE, actments of tamdahabhaw are limited by avoidance rules precluding the
they instance and index the varioue kinde of Bociability that define them. participaHon of younger women; in those "fun" evenings in which I was

re
..

Shared and solitary sentiments t23


122 Donald Brenneis

involved, female participation was generally limited to women con- socially pleasurable but injudicious gossip of men (Brenneis L984). Women
sidered my elders. are not considered prime offenders here, perhaps because men are more
Several points are critical here. First, the egalitarian ethos informing likely to disregard what women have to say than the commentary of
adult male relationships in Bhatgaon is rarely considered to extend to other males.
women. Some male villagers, particularly members of the reform Hindu Finally, the social lives of women in Bhatgaon are in many ways oPaque
congregation, claim that women are men's spiritual equals and should to males, whether villagers or ethnographers. It was inherently impos-
be afforded the chance to develop their capacities through joining in sible for me to participate in all-female events or to interview a wide
worship. In practice, however, a concem for sustaining equality is - from range of women in any depth. Such extended discussions as I was able
the maleigillagers' point of view, at least - solely a male matter. In the to hive with appropriate female friends - as defined through fictive kin-
manag6rnent of disputes, for example, insults against women are rede- ship - suggest several tentative conclusions. First, the women with whom
fined as attacks on their male kin; it is male standing that is endangered I spoke.aSpeared to share without evident irony in the discourse con-
(Brenneis L980). cerning@. They saw those hypercognized emotions as indeed valu-
As in those North Indian communities from which their forebears em- able. Secffid, I received no inkling of a complementary set of emotion
igrated, Bhatgaon villagers reckon kinship patrilineally. There is little theories held by and concerning women. This could reflect my inability
genealogical depth, however, and the large corporate kin groups com- to work extensively with women. Emotional notions linked to such themes
mon in North tndia do not exist. Marriage is village exogamous and res* as nurturance and compassion might well figure strongly in female the-
idence virilocal. Young women marry into Bhatgaon and find them- ory and practice, but I have no evidence.
selves, initially at least, strangers in their husband's households and in ir, rr*, then, women are, by and large, excluded from the particular
the social world of village women. The first few years of marriage are events linked with the highly valued and inherently social bhnw.ln ad-
generally spent in a joint household in which the new wife is subf&ct to dition, the discourse of these elaborated emotions in Bhatgaon can be
her mother-in-law and her older sisters-inJaw. However, although ini- jointly created only in such contexts. The bhaw depend on shared dis-
tially isolated, young wives get to know women beyond the household .o.r.r" for their expression and experience; women are much less likely
fairly rapidly. They may, for example, accompany their husbands or fe- to take part in such discourse and, consequently, in those emotions cul-
male affines on visits to other households or take part with other women turally defined as worthy and beneficial. From the standpoint of local
in preparing special feasts or attending weddings. theory, women rarely have the opportunity to build or create such ex-
Oldest sons and their wives and children are most likely to remain periences. This suggests two major implications. First, if one is con-
permanently with the son's parents; most younger brothers establish cerned with emotion discourse as constituting commentary on social re-
new households within three or four years. Men generally attribute such lationships, women's exclusion from 'amity', for example, serves as an
household division to enmity between the coresident sisters-in-law. A idiom for their exclusion from the socially valued world of religious com-
frequent comment is bese qussa, bese jagara ('too much anger, too many munity. In an odd way, that this religious community is primarily male
fights'). Even if twqcoresident brothers are known to have been dis- is almost incidental. The gender-based distinction is hidden in a puta-
agreeing about a shared household budget, for example, their wives' tively gender-indifferent view of the emotions, a practice that leads to
apparent antipathy is publicly defined as the real cause of the split. this diitinction's being taken for granted rather than overtly articulated'
Women's anger - not men's arguments - leads to disruption in such second, if one takes emotion discourse as emotion, as helping to shape
situations. and constifute individuals' lives, women may well be excluded from these
It is not that women are thought to have a monopoly on anger; it is, socially valued experiences.
rather, that they are considered to have fewer emotional or practical Although both men and women have such hypocognized feelings as
alternatives to its expression than do men. Although social disruption anger ,happen' to them in situations where they are not emotional actors
within the family is often attributed to women's anger, another common bul passivl recipients, it is primarily men who can jointly engender
source of trouble - particularly between unrelated individuals - is the friendship and play, In contrast to the striking split in Western ideology

.,:iEiifi*=
125
and solitary sentiments n
Shnred ,:d^^ ^,
124 Donald Brenneis
between reason and emotion
and the corresponding association
of emo-
the dis'
'"'**lt*;;'"1.s*##H'l*{;ffs#ilffti#{tFfl1ff i{j*
i" ottutguo"
ihtt volume)' neither
tion with the female ii;;;' control the x'ff ;
occurs' Instead'
tinction nor the Ast'iiution "t""-g"tt"tully
t""t'ffy significant emotion is realized'
-":j;eg;5;;1ff5"p;;1:,i:tr2#,r.?l,x'v"nlaffi Eth-
discourses through *iJt' on Iraruk' American
or Emotion words
l-,,#,*#ffi;'!il?T#"in
Note Diii.t;i',e'"-,1':if ;:;;:X;H;:;'?E*
"'u:i"'i|,?ilfiiriioyi:i;;fi
rwouRr3}iketothank$tlTlrsl':"rTi,Cf;Ti'liix-'iti"&li1'l-?i:rsilcjfl
ilffil'{H'iliftit;ffi;;t,' o.,
"u.ri",
a'u}i'.-s"'u'isevmour a'r
-*'f#*i:?P'iTi';llff Yr*i:i,r!#'T;txtl;::''l:'l?'i3n1H:
Milrerprovided the .rii*'i" *rrl.i., *y r"iti"i
esPecrall f,:llii
""a"iriandi'ngs :1,11; ss."i
dimension first emerge'di'uo-i'"ur'io"t
ffi:;ffi;ilb"" sEg"i provided
otirer P.itzer colleagues'
useful criticism'
::,ffi#ffi i*i$*,i'dryIf-#rfi#ffiffi
Lolg":8:
Schieffelin and E' Ochs'
eds''
;
References

iil7ii*"
.

Abu-Lughod, Lita' 1989: Y^"::i


American Ethnologts.t
6t.
arrd the Sentiments of Loss
ttnnoto gist tztzq>nL'
12,2a5a7'
iiilia sirii.Z*t'
in a Bedouin Society'

n Bed ouin Society. Berkeley:


Honor and Poetry in
Univer-
:,[Ef fflT#-'}Hffitrh':;t',#l;#t^fr ,,Hfffr
H*1-#"*,#:F"-^t*itrrty*,X*!Htri*$'*iir+'rs;
iffi
,#';*":iltit:*."i:lXlii"?;.o",llT:i"lf :ff ll"'i#,1;.'JXTH$'verbat
Tilifi-v"l?Ji#:l?i#;iltityn:",:n''f"iffii:f ;:#:;"F'tkt're
rsil. P"rf.i."*ce and Honor in
99:131-50'
Besnier, Niko. 19g9. Literary
'
and Feelings: The Encoding
of Affect in Nukulaelae

Arrect on Nororu"ru". Ms,


Department or Anthro-
'i
:{ffi *ruyg$reAlpn*f*t*-rii-i,*
ilon' Catttbidg'
"i:ti*ri:t3l"",?1;".
jf Choi ce : M alice and Mischief
in
-* [1;""*\".y ffi
'fl.
"
t" r',"' ol .o"f f :l'"

"!!;ifn:ffifiYii-i{#it?*{:,:,:::::::i:"ilI;;
1987. Performing Pas
' ;;;i'"
a" d;;;;",v' ""it' Ethnot o-sist 14: 236-50'
rssa' The Audience as Co'
Duranti, Allessandro"ffi; 6;;ie'
f-t""""i''

U:- j;;;e":;;tril;;;ini';iio,,,ou,,,ontan-
^:,lii:rJl'rW:;#;';:"12:i,Tff'y,,*rweepins'Poetics.andsonginKatuti
1""*rufrlltt#fr
5:qiil,I'"5ffiL3*tyl:l"Xi{ii.!":*":':':':":-*actetanK'
Villase. Princeton
"".,'#f -Dell' 1975' eiliiiiit;';gh ittto pti'rotill"te' Lt D' Ben-Amos and
,
Hvmes, , o"a''diiii"ication. Thle Hague: Mou-
Gordstein, .a".,'iiffiil'iifiior*on
ton, PP.11-74'
,e, Jludith.
Irvine,
'..
{ir;.
rsa: Languase
i,idith. 1982. and Affect:
tunruuge anu nrrELr' """'- ;;;, lnterdisciplinary
9T:i.::::;:Y:Yi!,}?:tr ;kX
Dimen-

il#";;#i,'#ruYa::;:t*":'"!,"*:f&"fi:;;;:';;:";iri''
sibns. Washington,
Registering affect 127

connection with affect, even though the authors of that literature have
not always foregrounded this point. Whether recognized as such or not,
7. Registering affecfu heteroglossia the evidence has begun to accumulate that would support the incorpo-
in the linguistic expression ration of affect more centrally in models of communication (see ochs
1988 for a review and discussion of this literature).
of emotion Equally important to Sapir was the question of how much the com-
munication of emotion is culturally constructed and culturally variable.
Although lively debates on this subject.have efsted for decades, for
IUDITT+{. IRVINE
.h.-. much of that time empirical studies tended to focus on nonverbal, ex-
pressive.pestures and "paralanguage" (voice quality and dynamics).
Anthrolqffical studies arguing for cultural variability in the construc-
Introduction tion of ei*otion turned, instead, to the emotion lexicon (see Rosenberg,
this volume); but these labels for describing and reporting feelings have
"Linguists of all breeds seem to develop cold feet when it comes to dis- usually been studied at some remove from the emotional performance
cussion of the expression of emotion in language," remarks w. o. Bee- itself. sfudies on the use of a wider set of linguistic resources, during
Tan
in a recent paper (Beeman 19gg:9). The problem, he sugigests, is affective performance and in connection with cultural differences, are
that affectivity is seen as "soft" and "idiosyncratic," contrasting with only now beginning to appear (see e.g., Abu-Lughod 1986; Beeman 1988;
reference and cognition, and therefore with the rule-governed struc- Brenneis 1988; Irvine 1982; Ochs 1988).
tures on which linguistics as a discipline has principally focusedpllee- It is only recently, then, that scholars have begun to take an interest
man attributes these assumptions to the emergence of transformational- in extensively pursuing and substantiating a view once articulated by
generative theory in 19s7, with its explicit definition of linguistics as Sapir (1927, L930): that emotional expression is pervasive in linguistic
a
cognitive discipline; but similar ideas can be found in the traditions of structure, and that the communication of personality and emotional states
saussure and Bloomfield, and in many other disciplines as well. It is not is culturally organized in a speech communitSr.2In this chapter, I take
only Chomskyan Iinguists who have supposed that affectivity belongs up Sapir's theme of the conventionality of affective performance and
to an "expressive" (nonreferential) dimension of language, that it per- follow it in relation to the linguistic concept of "register" - a coherent
tains to individuals rather than speech communitier, ,rd tt ut it is largely complex of linguistic features linked to a situation of use. What I suggest
a property of parole rather than of langue. shweder (Lgg4) and others is that many of these "varieties according to use" (Halliday 1964) have
have argued that an emphasis on cognition and rationality and a mar- an affective dimension that organizes some of their linguistic character-
-
ginalization of affectivity - have increased in the social sciences gener- istics. The study of registers is a convenient way to look at the verbal
ally since World War II.1 aspects of affective display, because it suggests a set of complementary
Although acknowlfuging that some languages do provide phonolog- representations of feelings that are conventionalized among a commu-
ical and morphological devices marking emotional states in speakirig nity of speakers. Where my discussion differs from Sapils is in its greater
(Henry 1936; stankievmcz 1964), for the most part linguists have seen emphasis on an interactional approach and, more importantly, in fore-
"expressive phenomena" as marginal to the main bodlof the grammar, grounding the organization of diversity in a communit5z's conventions
Yet, two of linguistics'most stellar figures, ]akobson (1960) ind sapir of affective performance.
(1921',1927), thought otherwise: Affect, or emotion, according to The ethnographic case to be drawn on for this discussion concerns
them,
was a fundamental dimension of human life and a factor cro:ss-cutting styles of speaking in a rural Wolof community in Senegal - styles that
all levels of linguistic organization. More recently, a growing literaturi relate to the repreaentation of affectivity, on the one hand, and of social
on such diverse aspects of linguistic structure ae word order, ierb voice, rank, on the other. To emphasize the role of affect and personality in
quantifiers and deictice, right and left dlalocaHon, and eo on indicates theee lingulrtlc regloters is not only consonant with a polof cultural
a
726
..

128 Iudith T. Irvine Registering affect L29


scheme, but is also useful as a way to understand the registers' system- formal: Which dialect you use is determined by your geographical and
aticity and their relation to cultural images of persons. Those images, as social affiliation (especially your locus of origin); which register you use
we shall see, form an organized set, as do the linguistic forms that dis- is determined by the social activity you are engaged in (i.e., the proper-
play them. ties of your situation, whether it is something already given or some-
Indeed, the importance of complementarities - in affectivity, in social thing you are trying to create).
differentiation, and in linguistic form - in this case is one of the reasons In Halliday's view, this functional distinction has consequences for
for focusing on registers. The concept of register is inherently hetero- registers' (and dialects') formal properties: According to him, the regis-
glossic (i.e., suggesting a plurality ot "voices" in the Bakhtinian sense): ters of a language tend to differ from one another primarily in semantics,
By defifuion, there is always a set of registers, not just one. As opposed whereas the dialects tend to differ from one another in phonetics, phon-
to the'fiotion of a "langaage" with an assemblage of expressive devices ology, and "lexicogrammar" (Halliday 1978'35). Yet, even if such ten-
all equally available to each speaker as he or she goes through the full dencie;&an be identified in some cases - and they do not apply con-
gamut of feelings, "affective registers" suggest a set of different repre- venienffi the Wolof registers to be described - the fact remains that
sentations of emotion - a set that may be culturally defined and linked the differences among registers are not actually limited to semantics,
to cultural conceptions of the diversity of persons, personalities, and however broadly semantics is defined.s There are also grammatical,
situations. phonological, and phonetic differences.6
In short, the communication of feeling is not merely a property of the The questions, then, are what all the various linguistic devices making
individual, or a function of transient irrational impulses, or an unruly up a register have to do with one another; what governs their distribu-
force operating outside the realm of linguistic form.3 Instead, it is so- tion; and how tightly they cohere. Some sociolinguists have described
cially, culturally, and linguistically structured, and we cannot ade- the linkage of devices in a register in terms of "co-occurrence rules"
quately interpret individuals'behavior as emotional expression uffiil we (Ervin-Tripp 1972). But however useful such rules might be for descrip-
understand some of that framework. tive purposes, what is still needed is some sense of what motivates them
- in what sense a register may be systematic, its devices having some
Register, ?ialect, and the contextualization of emotion coherence, and how much coherence they must have.
My suggestion here is that one of the principles underlying a register's
Entering the vocabulary of American linguists only relatively recently, co-occurrence rules can be coherence of affective performance - a prin-
the term "registe{' is most closely associated with a British school of ciple governing prosody as well as vocabulary choice and sentence com-
linguistics, particularly the work of M. A. K. Halliday and his followers. position. The connection with affect is often explicit. In Javanese, for
It derives (Halliday 1978) from that school's interest in the ways speech example, speakers apparently conceive of the available language levels
is affected by the "context of situation," a notion prominent in the writ- as registers whose use depends on one's assessment of a situation, par-
ings of J. R. Firth and, earlier, in Malinowski's writings on language. As ticularly the relationships among its personnel and the appropriateness
defined by Hallidaydl964), a register is a linguistic "vaiety according to of displaying affect. The higher, more "polite" levels are considered de-
*a
use," in contrast to "dialect," which is a "vaiety according to users." personalized, regulated by an ethic of proper order, peace, and calm; in
Although American linguists (and sociolinguists) for a long time tended them ond "does not express one's own feelings" (Wolff and Poedjosoe-
to prefer the more general terms "linguistic variety" arrd "code," "teg- darmo 1982:4L). The lower, "coarser" levels, on the other hand, are the
ister" is now becoming widespread.a "language'. . one loses one's temper in" (Errington 1984:9).
In the usage of most linguists, registers are styles of speaking whose There is also a sense in which the Javanese language levels distinguish
grammatical configurations overlap; that is, they are styles that are lin- categories of speakers. People do not control all levels equally well. Geertz
guistically distinguishable, but only as varieties of one encompassing (1960) describes |avanese social ranks as differing significantly in the
"language." Of course, the same could be said for dialects. The defini- range they control within the total repertoire, the more "polite" levels
tional difference between dialect and register is functional, rather than being more arrocletcd with persons of traditional high rank. As images
Registering affect 131.
130 Judith T. Irvine
of "refinement'' and affective display, then, the language levels evoke feeling' so much as thoughts which are necessarily linked with social
both the situations characteristically evoking such resPonses and the situations and valued goals that give them moral force and direction."
But perhaps the most important reason the concept of register may be
persons characteristically manifesting them.
As the Javanese example illustrates, the distinction between dialect useful in the ethnographic study of the cultural construction of emotion
is the emphasis on the role of conventionality, linguistic and otherwise,
and register, although a useful analytical starting point, becomes more
complicated as soon as one looks more closely at a particular speech
in affective display. Such conventions, linguistically expressed, rePre-
sent a culfural construction of available emotions, personalities, and so
community. Although there are several reasons for this, some of the
most inqpgrtant ones have to do with the cultural structuring of "voices" on that are linked to other dimensions of culture and society. Indepen-
associglrid with social groups. Images of persons considered typical of
dent, in a way, of what a person may "actually" feel at a particular mo-
those grbups - and the personalities, moods, behavior, and settings ment, they nevertheless represent the resources the person has to draw
characteristically associated with them - are rationalized and organized
on fog. ffective display, the terms in which his or her behavior will be
in a cultural system, and become available as a frame of reference for interpffi by others, and the framework of interpretation for the ex-
one,S own performance and for interpreting the performances of others.
perienc6i as well. Moreover, as Abu-Lughod (1986:24L) has noted, the
This system informs the style switching that all speakers engage in. expression of feelings has relevance not only for the person expressing
Thus our verbal performances do not simply represent our own social them but also for the audience and the furtherance of social relation-
identity, our own feelings, and the social occasion here and now. They ships: "sentiments can be used to symbolize something about the per-
are full of allusions to the behavior of others and to other times and son expressing them." Communicative forms are not merely the "ex-
places. To put this another way: One of the many methods people have
pressive" result of psyctrological processes intemal to the communicator;
for differentiating situations and marking their moods is to draw6n (or they are also forms of social action.s
carefully avoid) the "voices" of others, or what they assume those "voices"
to be. The concept of register, then, although initially defined in termg Wolof registers and affectivity
of situation rather than person or group, in fact draws on cultural im-
Let us turn now to a consideration of styles of speaking in a rural Wolof
ages of p6rsons as well as situations and activities.T There is no contra-
diction between the situational, or contextual, focus and a connection community.e The Wolof language has relatively little geographical dia-
lect differentiation, although urban usage has come to differ somewhat
with the cultural construction of emotion and personality.
Actually, the situational focus in the concept of register is part of what from rural 'true Wolof ' (wolof piir). Differing forms of talk do exist on the
makes it compatible with recent anthropological approaches to the study
rural scene, however. Informants in a Wolof village identified two styles
of emotion, which have often emphasized a contextual approach. Al- of speaking, u)axlt gewel and waxu gi6r, which they linked with differ-
ences in affectivity and in social rank. One of these (waxu gd€r), a laconic
though our own society tends to locate emotion firmly within the indi-
vidual (as a matter of internal psychobiology and biography), other cul' way of speaking, is conventionally associated with high-ranking nobles,
because these persons are assumed to be its typical usersi the other style
tural systems may l&ate it in social relations or in the relationship between
(waxu gewel), hyperbolic and high in affectivity, is conventionally asso-
a person and an event (Brenneis, this volumei Lttz 1982; Myers L979),
In such cases - or for the scholars similarly disposed to construe emotiofl ciated with the lower ranks, particularly the "griots" (praise singers,
npeech makers, and bards).
as socially shaped - the social situations in which emotions variously
arise or are displayed, and the social relations experienced by and partly
Why call these forms of speaking "registers"? Although waxu gewel
and waxu g€€r may look at first like social dialects - distinct linguistic
defining the self, become critical for emotional understanding. As Lutz
varieties associated with distinct social SrouPS, the nobles and griotslo -
and White (L986:419) note, in such views "actors understand emotiont
as mediating social action: they arise in social situations and carry impll'
It turns out that speakers from all social categories use both ways of
cations for future thought and action. Emotional understandings, then,
rpeaking, depending on the situation. In practice, therefore, these ways
are not seen as abstract, symbollc formulatlone - not 'thinking about of opeaking opef&te aa registers, resources on which any speaker may
Registering affect 133
132 ludith T. Irvine
draw to define the situation and the relationships at hand. Moreover, nization of talk among any interactants should their differences in rank
the wolof labels do not mark two absolutely distinct varieties so much become relevant.
as they point to the poles of a behavioral continuum or range of possi- A connection between differences in rank, behavior, and affectivity is
bilities. fundamental in this ideological formation. Central to it is the idea that
In what follows, I shall describe the linguistic phenomena involved, people are inherent$ unequal, having different constitutions that gov-
their relation to affectivity, and their role in social interaction. Before th"ir feelings and motivations and make them behave in different
"rn
ways. These different behavioral dispositions incline their bearers to dif-
doing so, however, it is useful to set the cultural scene. A Wolof ideol-
ogy of lq+guage and social rank rationalizes the differentiation of ways fering but complementary forms of ,action, whence arise the occuPa-
of spe$!!tg and their connection with rank and situation. Images of af- tional specializations of the caste system. Since disposition and person-
fectivity and impulsivity, manifested in verbal behavior as well as in ality arl to be inheritable, personality of differences follow
"thought
other forms of activity, are central to this cultural scheme. genealqgdqal lines, especially the lines of caste endogamy. One's consti-
Ltior, ffi'Uiotogical, emotional, and moral qualities one derives from
ancestral]inheritance and from childhood experience - is the source of
orre,s nalo,temperament' or 'capacity for emotionality'. Persons of the
Ethnographic background: a Wolof ideology of rank same rank are nawli 'castemates', who resemble one another in temper-
Like many other peoples in the western sahel and savanna region of ament and moral status.l3
Africa, the Wolof traditionally were organized in a complex system of These temperaments contrast, according to informants, between de-
social stratification, usually termed a "caste" system in the African eth- grees of affectivity and impulsivity, on the one hand, and restraint and
nographic literature. It is a system of ranked grouPs and subgrgups, iorpidity, on the other. Restraint, self-control, and sangfroid are norma-
occupationally specialized and strictly endogamous. Although under- tive characteristics of the nobility, who are supposed to be more solid,
mined by government policies and other factors, the caste system re- stable, and "heavy" than the typical lower-caste Person, who is "light-
tains considerable importance in rural communities, and even on the weight,,, volatile, even frenetic. On a symbolic level, the contrast is
urban Scerfe, according to some observers.ll The material presented here ,o-"ti-"r described in terms of weight, sometimes in terms of differ-
is drawn mainly from fieldwork in a Senegalese village of some 1,200 ences in the viscosity of body fluids, and sometimes in terms of the
people (in the 1970s, when most of the relevant fieldwork was done). elements - earth, aig,, fire, and water - as if the various castes consisted
In Wolof village society, differences in rank are EIn acknowledged value of these, or resembled these, in di{ferent degrees, balances, or em-
that organizes all sorts of social activities and interactions, ranging from phases. Thus the nobles' self-control is like the stability of earth, as com-
economic specializations and exchange to the regulation of marriage, pared with the other elements, supposedly more prominent in lower
and induding social contact and conversation. Of the low-ranking castes, castes: air, which (in this cosmological scheme) suggests a capacity for
the most numerous and conspicuous is the bardic caste, or griots, whose frenetic movement; fire, which symbolizes anger, the warmth of pas-
specializations are p&blic speech making, praise singing, music, and the sion, and a consuming greed; and water, which symbolizes fear (but
rhetorical and communicative arts in general. To some extent, however, also, in a more positive sense, the cooling of anger).14
the rhetorical specialty is a prerogative of all nonnoble castes, because Of the lower castes, the griots, in particular, have the image of high
all of them can substitute for griots or participate along with griots on affectivity and excitability; as volatile and theatrical personalities; and
occasion. Moreover, in a way, the image of the griot publicly eulogizing, especially as people who excite others. High-ranking nobles are conven-
persuading, and entertaining the noble is only the crystallizatton, on the tionally u$o"irt"d with self-control, but also with lethargy and bland-
level of caste differences, of a culfural convention associating certain kinde ness. Their restraint (kersa) may "make them reluctant to say bad things,"
of verbal performances and conspicuous verbal activity with rank differ' as some villagers said, but (they continued) it also makes them reluctant
ences small or great.12 The difference in verbal behavior associated with to Bay or do much of anything. It takes a griot to make life interesting
griots and nobles servee as a cultural prototype, I believe, for the orga' and attractlv€ and to keep the high nobles awake. Just as a rush of air

G
L34 ]udith T. Irvine Registering affect 135

stirs a fire, so the breath of the griots' utterance rouses passion and en- The difference in "weighf'between high and low ranks thus affects
ergy in the noble, to the point where creative action becomes possible. the consequences of emotionally expressive behavior, as well as its like-
The "hot" elements in the noble's temperament, ordinarily dormant, are lihood of occurrence. Although griots may easily express emotionality,
thus activated. their feelings do not bear the heavy social import that high nobles' feel-
This complementarity is illustrated in a frequent image in Wolof oral ings do. The high nobility are administratively and economically respon-
poetry and epic narrative, where a slumbering king must be wakened sible for large numbers of dependents; their actions, and the sentiments
by an orchestra of griots playing drums and iron clappers lest his royal leading to their actions, are seen culturally as more consequential. One
duties g.o.unfulfilled. (The same image is cited for villageJevel political of the griot's main roles, therefore, is to express the ideas and feelings
assenslids today, where a glouP of griots punctuate the political speeches of nobles - to communicate for them energetically and persuasively, es-
with drum rolls and cheers "because otherwise the chiefs would fall pecially in the public arena. Griots are expressive vehicles, and much of
asleep.") Once roused, the king may be moved to Sreat deeds - the what th{y express is not ultimately their own. It is, instead, supportive
greater because of the seriousness and weightiness of his personality - repetiuffipersuasively elaborated and charged with excitement. As a
but griots are needed to stir him to that point. friitoriariand repository of folklore, the griot is the repeater of the voices
If griots are thought of as people who easily exPress their own feelings of tradition and the ancestors; as a public spokesperson for a noble, he
and who stir the feelings of others, the affectivity of their behavior is or she repeats and elaborates an idea the noble may have whispered in
shaped by their rhetorical and aesthetic skills. Related, on the concep- the griot's ear onlY a moment ago.
tual level, to the "lighter" elements' capacity for motion, fluency of wotor villagers summarize these aspects of the griot's role with the
expression is at least as important in the image of the griot as is lack of term jottali,transmission" which must be understood as supportive,
inhibition, from which it is in fact inseparable. Not only verbal ffpres- charged, and elaborated repetition. It is essential to what the griot does,
sion is involved: Griot women are suPposed, for example, to be the best and iherefore to the nature of the "griotlike" style of speaking. The cul-
dancers, since high-caste women are "too stiff" and "too ashamed" to tural framework suggests that sentiments are most constructively, per-
perform in the sexually suggestive manner deemed the most skillful. suasively, and even effectively expressed when routed through a speaker
As for fhat'shame' (jom) and its experiential concomitantruus ('to feel of lower rank than their originator. The forms of talk considered to have
ashamed') - these can perhaps best be understood as deriving from the the greatest emotional charge thus include those performance genres in
interactional model of emotional experience that seems to be important which griots specialize and that exemplily "gtiot talk." Moreover, the
in the Wolof cultural scheme. As informants explain them, the actions process of repetition, whether it concems repetitive constructions within
that give rise to a feeling of shame are not necessarily inherent$ bad in individual speaker's tum at talk or repetition by other speakers, is
(although they may be); rather, they are witnessed actions inappropriate considered to heighten the emotional effect.
to one's rank. Thus the acts one would be ashamed to do include behav-
ior that might be quite proper in private or if directed toward superiors, Linguistic oarieties
but shameful if wiCI*essed by a public that includes lower-ranking per-
sons.ls Fluent, voluble public speaking is an example: When addressed The two labeled styles of speaking mentioned earlier, waxu gewel and
by gnot to noble, the interaction of their naw operates to constructive waxu gdir -'griot talk' and 'noble talk,' respectively - can now be con-
effect, overcoming the noble's torpor; if addressed in the other direction, sidered against this cultural background. The linguistic phenomena that
the interactive effect would be destructive to both. The noble would fail constitute these registers are found in all aspects of verbal performance,
to maintain restraint, and the griot's own self-control supposedly would from prosody, phonetics, morphology, and sentence structure to turn
be too fragile to withstand the force of the noble's breath. As a result, taking and the management of conversational discourse. We shall see
the griot would be imperiled - made vulnerable to self-destructive im- that the contrasting images of nobles and griots - in terms of affectivity,
pulses and to attack by witches and nefarious spirits - whereas the noble rhetorical fluency and aesthetics, and elaborated repetition - shape many
17
would be ashamed and might even suffer some loss of rank.16 characterlltlct of the registers.

-==G
:.,
i

Registering affect 137


L36 Judith T. Irvine
which they are all "noble talk"; there are only differences relative to one
Table 7.L. Prosodic contrasts another and established locally, in a particular situation. In all these pro-
sodic phenomena, the registers are mirror images of one another,
"Noblelike" speech "Griotlike" speech emerging (and diverging) when, and to the extent that, differences in
Pitch Low High rank emerge as important.
Voice Breathy Clear The most extreme versions of the "griotlike" prosody are found in
Volume Soft Loud
Dynamic and pitch nucleus those public situations where the greatest differences in social rank are
Contour Pitch nucleus last;
little dynamic range first; wide dynamic and most pertinent. Political meetings, villagewide celebrations, and the feasts
tsi-l\ pitch range given by prominent families to mark major life events are good ex-
Speed t'"-- Slow Fast
amples. The griot stands and shots as loudly as possible to the assem-
bled grqwd; he sways and jabs at the air with dramatic and forceful ges-
Despite their connection with caste images, in practice these varieties tures,ffinting at his addressees and holding up the money he has
are not associated exclusively with a speaker's absolute rank (griot or receive#in largesse. The veins in his forehead and neck stand out from
noble). Instead, they are used as a range of registers signaling relatiae the effort of the performance. His voice rises into falsetto pitch, and the
rank (of the speaker as comPared to the addressee or relevant other) and rate of his utterance increases to more than 300 syllables a minute.le
signaling types of situations - those in which differences of rank are to Meanwhile, the higher-rank persons present, especially the highest no-
be attended to. Except for the extreme ends of the register continuum, bles, remain seated, silent, and motionless, until moved to some action
all speakers use all the rest of the range, depending on their relation to such as handing the griot a gift or (at a political meeting) making a terse
a particular addressee and situation. A person of noble rank uses a comment.
"griotlike" style of speaking when addressing a noble kinsmarfffrom Even in conversation, prosody observably shifts along with changes
*hom he or she wants to ask a f.avor, for example. Thus the registers in topic, affective tone, and attention to rank. For instance, one evening
invoke a kind of metaphor of high and low ranks - and a model of affec- a certain noble was sitting in my house, along with a few other nobles,
tive relations - in order to define (in this case) an act of petitioning and chatting about politics. It was early in my fieldwork, and I was trying to
to make the petition Persuasive. elicit talk for the tape recorder, without yet having developed much sense
Let us now examine the linguistic characteristics of these registers in of the appropriateness of different kinds of talk for different speakers.
more detail. Following this description of the registers, some text and Unaware that village nobles do not normally engage in public oratory of
transcript examples will show them in actual use. any sort, especially on their own account, I asked him to give me an
example of a political speech. After some mild teasing by the others and
Prosody.In some respects, prosodic features provide the most striking general amusement at the idea that he might do such a thing, he iden-
differences between "griotliks't and "noblelike" registers. The differ- tified one of the few appropriate contexts for such talk, offering com-
ences are not only gpvious to an observer but salient to Wolof villagers
ments about what he would do if he were made an official representa-
themselves, who explicitly call attention to the pitch, speed, volume, tive of his village charged with welcoming important visitors at the next
and quantity of talk as defining characteristics of the registers. Voice political assembly. His voice dropped to a basso profundo, and the rate
quality (breathiness) and intonational contour observably differ, along of his utterance slowed to fewer than 60 syllables a minute. Attention
with the other characteristics just cited. Table 7 .L summarizes these fea- having been called to his rank, he spoke in the most "noblelike" pros-
tures. ody possible, short of refusing to speak at all.
As Table 7.1 shows, what informants are identifying as two forms of
Phonology.Informants do not seem to be as explicitly aware of phonolog-
talk are actually two poles of a continuumi no sPecific demarcation point
distinguishes one form from the other.18 For example, there is no abso- ical contraste in registers as they are of the prosodic ones; at least, they
lute decibel level above which all utterances are "griot talk" and below are not ae likely to deecribe them. Still, some phonological differences
L

138 Judith T. Irvine Registuing affect 139

are observable, and I have heard them parodied by teenagers imitating


Table 7.2. Phonological contrasts
the voices of older people or taking parts in dramatic skits. Moreover,
villagers suggested that a certain type of stammer (ddr), in which "the "Noblelike" speech "Griotlike" speech
tongue sticks in the mouth," often afflicts Persons of extremely high
rank, whose tongues are not agile and whose utterance is not rhetori- Feature contrasts
Contrasts in vowel length and con-
Contrasts in vowel length and con-
cally fluent.2o sonant length not clearlY main- sonant length clearlY maintained
Many of the phonological differences between the two registers seem tained
to be related to differences in the tempo of speaking. The articulation of Nonnasal stops affricated and/or Stops in stressed syllables, and all
prenasalized, e.g.: [P]'-+ [d,] "-fortis" stops, energeticallY ar-
soun{rftly be affected either by the haste of extremely rapid speech or ticulated
by thetlawl of very slow speech. Compared with the less extreme Por- lbl- [F], [mb], [mB]
"Breathy{r or "creaky" (laryngeal- Voicing contrasts in syllable-initial
tions of the register continuum, the most rapid "griot speech" perfor- izedlSfulation of voiced stoPs consonants, and all "foftrs"
mances tend to elide segments in unstressed syllables. In a few words, stops, clearly maintained
an unstressed syllable may be elided or omitted altogether. "Noble Elisions in unstressed syllables
speech," in contrast, drags out its vowels and tends to affricate its con- Stresses not clearly marked (little Stressed syllables clearly articu- -
sonants, even though the overall number of words is far fewer. When difference between stressed and lated; elisions in unstressed sYl-
unstressed syllables) lables: "Lenis" final stoP (in un-
the highest-ranking nobles in town speak in this style, the general effect stressed syllable) --+ A
is of drawling and low-pitched mumbling, an effect further exaggerated (especially if voiceless)
by practices such as holding the hand over the mouth while talking or Unstressed CY# - C#
turning and lowering the head so that the voice will be muffled by a Initial [k] in unstressed syllable ->
bulky neck scarf. [?]
To the Western observer accustomed to the notion that voice belongs Vowel height
to the powerful, there is something surprising about the Wolof high Some lowering of vowels? Some fronting and raising of
vowels, especiallY before Palatal
noble's rrxrmbled, muffled speech and, as one might call it, "conspicu- glides
ous disfluenry." Yet, there is a sense in which Power is indeed dis-
played in this form of talk - the power to command the audience's atten-
tion. The people who speak in the most extreme versions of this style phonetic alterations, if the speech of citation forms or even of calm con-
are those whose high status and authority are unambiguous. Although versation among castemates were taken as a standard, "griot speech"
they speak little, their right to the floor is unquestionablei other people apparently r"rr,uir,, more faithful to the underlying system of significant
seem always to reserve part of their attention for watching whether the arts. ior example, girrots uttering the extreme forms of "griot speech"
"orrt
highest-ranking one present is about to speak. When the high noble always maintained contrasts in vowel length (which are very important
does begin to say sgmething, an immediate hush falls, no matter what in Wtlof;, even if some of the short vowels had to disappear altogether.
the level of chatter was before. Everyone leans forward, straining to make In the most extreme "noble speech," however, distinctions of vowel length
out what is being said. This is quite different from audience behavior at tended to disappear in the mumbling drawl'
griot performances, where people of medium and low rank often con- A summary-o] the phonological differences I have observed in these
tinue to converse among themselves, and the griot's shouting is in some registers is outlined inTable7.2.
respects an effort to gain attention. the variations described in Table 7.2 seem to apply only to the most
The disfluenry of high nobles, and the effort necessary to understand extreme ends of the register continuum, although there are some
other
it, derives not only from slow speed or physical barriers like hands or forms of phonological variation (e.g., further differences involving vowel
scarves, but also from the phonological rules this register involves. Al- height) that eeem to be connected with age and gender more directly
though both "noble speech" and "griot speech" could be said to make thai with ffink or with these registers, and thus apply throughout the

.'.iffi.
."

1,40 Judith T. Irvine Registuing affect r4L

continuum. Clearly, therefore, much more detailed examination of pho- cial categories label "wrong" in elicitation sessions; there is a relatively
nological variation in Wolof in general, as well as in these registers, could heavy use of the bi class, which is semantically more neutral than others;
and should be done. and agreement rules are not consistently followed.
In contrast, griots' class markers, especially as used in public perfor-
Morphology and syntax. The differences in fluency and level of excitation mance, are generally assumed by others to be grammatically "correct"
that we have noted so far in prosody and phonology also show up in even when they actually deviate from historically attested forms (Irvine
the registers' morphology and syntax. "Griot speech," the hyperbolic, 1978).It is also assumed that they involve some aesthetic principle or
elaborated, verbose variety, is replete with emphatic devices, parallel- semantic subtlety, even when the audience cannot identify what that
isnlsiFhd ideophones. Its grammitical constructions are thus consistent semantic effect is.
witffttre emphatic intonational pattern and dramatic gestures, and with Another aspect of rhetorical fluency concerns the completeness of
the image of the griot as poetically skillful, affectively charged, and con- seg,rt{nce structures. Whereas the sentences of "griot speech" are well
spicuously talkative. "Noble speech," on the other hand, is terse and forffi) complete, and often elaborate, with relative and conjoined
bland. It is also less fluent in that sentences may be left incomplete and clauJEs, "noble speech" shows false starts, unfinished sentences, and a
agreement rules are not consistently followed. tendency to use simple, unmarked constructions with few modifiers.
Differences in fluency and grammatical "correctness" - as well as in Again, the Wolof kings of former times provide some illustrations of
aesthetics and vividness of speech - are manifested in several ways. For "noble speech" patterns. Consider, for instance, the speech of the king
example, these differences are relevant to a morphological variation that of Saloum, recorded about 1865 when he paid a formal visit to the Cath-
seems to differentiate the two registers.2l The forms concerned are var- olic mission at Saint-|oseph de Ngasobil, which was then headed by the
iations in noun class assignment - where one tends to assign;r noun, missionary linguist Mgr. Alois Kobds. The missionary archives (Abiven
among the classes bi, gi, si, ji, wi, mi, li, or ki, so labeled according to the n.d.) describe the visit as follows (my translation from the French; Ko-
form of the proximal determinant used with them. In principle, classes bEs's Wolof transcription is unaltered):
are assigned on the basis sometimes of semantic subtleties such as con- The king has a birth defect: he stammers a great deal ordinarily,
notations of an object's size, sometimes of consonant harmony (thus but it is to be noted that he stammered rather more in speaking to
sordngs sl 'the orange [near me]'), and sometimes simply of historical Monsignor. . . . [When he saw a statue of ]esus open and shut its
convention. "Noble speech" and "griot speech" differ not only in which eyes and move its arms,l the king said (stammering): "Li, li, li, ll,
class markers tend to be assigned, but also in the consistenry of class- ma gis . . . y6pir, du du dara,lef ld, lef l'angi" ("Everything I have
marking agreement rules, and in whether the determinants and relativ- seen until now is nothing, this is the thing" - of all the other things
izing constructions that would require overt class marking are present in the mission.)
at all.
Rendered in modern orthography,n the king's utterance looks like
My attention was first called to this matter by informants' reports that
this:
the Wolof kings aryl high nobility of former times used to use the "wrong,,
class markers. They had to make mistakes in minor points of grammar, Li, li, li, li ma gis - yepp, du du
because correctness would be an unnecessary frill, an emphasis on fluenry This this this this I see everything neg neg
of performance or performance for its own sake, which would not be This - this - this - this I see - everything, it's not -
appropriate - or perhaps even possible - for these highest of nobles.
(The same was true for religious ascetics, who were too concerned with dara, l6f la, l6f laanggi.
otherworldly matters, and communion with God and the prophets, to nothing thing obj. foc. orNCM? thing pron*presentative it's
have room in their minds for forms of talk associated with the rhetoric nothing, that thing, that is the thing.
of this world.) In practice, "noble speech" uses few class markers. When Although Koble'e translation may well represent the import of what
markers do appear, they are quite often those that informants of all so- the king mGant to my, hie actual eentence ie lees well formed, The etam-

+-E
..

L42 Iudith T. Irvine Registering affect 1.43

mer we see here looks much like the false starts and incomplete struc- Table 7.3. Contrasts in morphology and syntax
tures observable in the speech of a modern village chief (see the tran-
script in the later "Examples" section) and other high nobles. Of all these "Noblelike speech" "Griotlike speech"
high-ranking people it is said that their tongues lack agility, their bodies Emphatic deoices
lack liveliness (unless aroused by a griot), and their minds are Preoccu- Unmarked (subject-verb-obj ect) Left dislocations, cleft sentences;
pied with weighty matters. order of basic constituents; heavy use of focus markers
sparse use of focus markers (subject focus, object focus, and
Although stammering and other disfluencies may lead to some repe- "explicative" verbal auxiliary)
tition of initial elements, this should not be confused with the parallelis- Frequent use of spatial deictics,
Sparse use of spatial deictics and
tic consFiffctions found in "griot talk." The constructions of "griot talk" determinants especially their "emphatic"
,r" .r"Ve, syntactically disfluent. Quite distinguishable from false starts formsa
and incomplete constituents, they represent instead complex (multiple- Sparse U+qof modifiers Ideophones (intensifiers); more use
clause) sentence strucfures and regular processes of morphological re- 'W'
q
of verb-complement construc-
|;ronni which often con-
duplication. Reduplication is a word-formation process in Wolof that, veys details of sound and
for verbs, suggests continuous or repetitive) intensified action. For ex- motion
-,
ample, crab-crab6, a verb formed from a reduplication of the French loan Parallelisms
wotd crabe 'crab' , means 'continuously, repetitively, or strikingly move Little use of parallelism Repetitive and parallel construc-
sideways, crab fashion'. It is a word I have encountered only in bardic tions (e.9., parallel clauses)
performances, although Wolof audiences did not find it obscure. It is Few reduplicated forms, especially Frequent use of morphological re-
in verbs; no novel use of mor- duplication, especially in verbs,
interesting to contrast these reduplicative and repetitive forms with con- phological reduplication. including novel word formations
structions involving the suffix -aat (as in dellusiwaaf 'return here dlain',
Disfluencies
waxaat'speak agairr', etc.), which refer to a repeated action but appar- Noun classification system: choice " Correc{' class markers; principles
ently bear no connotations of register difference. of class marker "wrong" or se- of consonant harmony and se-
That these reduplicative verb forms involve connotations of intensity, mantically neutral; avoidance of mantic subtlety; more use of
markers when possible; incom- markers; consistent and com-
vividness, and continuous action makes their frequency in "griot speech" plete concord
plete or inconsistent concord
consistent with other features of the register: its frequent use of em- Well-formed sentence structures
Incomplete sentence structures;
phatic devices (including those that result in cleft sentences and other false starts
complex structures), of the "explicative" and continuative auxiliaries, of
ideophones (intensifiers), and of a verb-complement construction that
most often conveys vivid details of sound and motion. All these features pushing'); and by continuative and explicative verbal auxiliaries (di I i,
relate to a conception of the griot as much involved with excitement, def\ t d?t I daa). Similarly, emphatic particles such as de'indeed' and ddll
noise, and frenetic lctivity. 'rcally' are also frequent in this register and more rare in "noble sPeech."
Table 7.3 summaiizes the morphological and syntactic constructions Other relevant lexical choices include the use of vivid vocabulary, es-
characteristic of the two poles of the register continuum. pecially in the representation of motion (e.g., iillanyo'toss about'; mfuntali
'shaking out, as when shaking skirts to get the dust off'; nappaaie 'squash
Lexicon. The preceding discussion of morphology and syntax has neces- with the foot'i sa j6d-a-j6d, ak sa jdd'your stiff strutting and stamping');
sarily touched on several aspects of lexical choice that differentiate these the representation of noise (mu dox di mbat-mbatia 'she walks around
Wolof registers. Thus we have already seen that the affectively charged making a slishy-sloshy sound'; nd ngtrlp!'went crashing down WOMP!');
register, "gnot speech," is characterizedby frequent use of spatial deic- or the description of strong affect (saalit'tremble in a frenzy of fear';
tics, especially in their "emphatic" forms; by verbs formed from mor- s\nggdm fooy 'hunch over weeping'). Ideophones (sedd guii'extremely
phological reduplication (such as topp4opp'continually and intensely cold'; nyuul kuak'extremely black'; wex tdll 'extremely bitter'; etc.) and

;;EEtu.
a

1.44 ludith T. Irvine Registering affect 1,45

hyperbolic expressions such as ba dee'to death' (ma ree ba dee'I died The village chief, for example, almost never speaks aloud in a public
laughing') or ba reey'till killing' are also characteristic. gathering, such as a public political meeting or a religious event held in
Honorific and eulogistic expressions, especially in forms of address the central plaza of. town. As his brother commented to me, "He [the
(praise naming) and expressions of thanks, are part of the "griot speech" chiefl ought not to speak in the public plaza; he ought not to speak in
register too. Indeed, because of the ideological connection between af- front of many persons. He is a chief and a great marabout [religious leader],
fectivity and rank, it is somewhat arbitrary whether we call this a hyper- and he would be ashamed." lf a high-ranking Person needs to commu-
bolic, affectively charged register or an honorific register, whose use im- nicate something to a large groupr or to some other person of similarly
plies taking on a low-ranking role for oneself and enthusiastically imputing high rank, he or she must resort to an intermediary, someone of lower
a high"-t'HHking role to someone else. However, the "giot speech" reg- rank (often a griot), to perform the communicative task. The noble says
ister nffiy be employed to express anger as well as enthusiasm. In the something quietly to the intermediary, perhaps whispering in his or her
case of anger, of course, honorific expressions will not occur, except ear; thqintermediary then repeats the message loudly and more elabo-
sarcastically ("F- N- [praise name], you're just like an outhouse: you rately,\illrying it to its intended receiver'
never lie, you never tell tales"). Sometimes derogatory nicknames and This frttern is widespread in the western Sahel. It is attested as early
insulting labels will be used where, in a praise performance, praise ntrnes as 1506, when the Portuguese geograPher Fernandes described the use
and eulogizing expressions would have been employed. of speech intermediaries among the Wolof and Manding, in groups as
In contrast, the laconic register, "noble speech," tends toward a bland small as three and even when all parties were Present together (Fer-
and affectively neutral vocabulary. This does not necessarily mean a nandes 1940:77-a). Similar reports have been made by many observers
general semantic neutralization, however. Nouns labeling detailed clas- through the years; for example, David Gamble (1967:75) notes that when
sifications of the natural world, such as the botanical or zoological lexi- a daughter or wife of a chief comes on a visit, she is accompanied by a
con, and those classifying the social world in a formal sense (wfthout retinue including griot women; the high-ranking visitor makes only a
any particular attitudinal load), may be found just as frequently in this few formal remarks, the lighter parts of the conversation being con-
register as in the other. What differentiates the registers has to do with ducted by her companions "on her behalf."2s Since a high-ranking host
affect andpublic performance, not with one's knowledge about the world. or hostess has griot spokespersons too, a formal visit may involve dou-
ble intermediaries. On one such occasion I was present when a particu-
Discourse management and interactional deaices. The register differences so larly important chief received another chief in his house. Each had his
far described are systematic in their connection with affectivity and rank, griot mediator, and after the initial greetings the conversation was en-
and with the Wolof cultural ideas that interpret these in the form of the tirely taken over by the griots, who conversed with each other on their
images, or behavioral stereotypes, of griot and noble. The same ideas nobles' behalf. The chiefs themselves simply sat in silence.
apply,too, to the management of discourse: the conversational roles Not only chiefs, but high-ranking Wolof villagers in general, rely heavily
available to participants and the distribution of rights to the floor. I have on intermediaries, occasionally even double intermediaries, in situations
described some of tfu;se patterns of interaction in other papers.2a For the they consider formal or important, and when they must address a mes-
present, I would like to emphasize a few aspects of the interactional sage to the public, to a stranger, or to someone of equally high rank. For
system. One (already mentioned) is the use of repetition to express sup- instance, any public announcement (of a birth, a death, a religious cele-
portive agreement with one's interlocutor; the other is the frequent use bration, an upcoming meeting, etc.) must be relayed through a griot.
of conversational intermediaries and spokespersons to relay a message Similarly, when important visitors come to his household, a high-rank-
through third parties, even when the message source and its target are ing noble does not greet them directly but calls on an intermediary - a
physically copresent. Both of these patterns are connected with the im- griot, a slave, a son/ or a wife - to do so and to mediate the conversation;
age of the griot as repeater, supporter, and transmitter of messages from only after some time (if at all) will the noble directly address the visitors
higher-ranking persons whose dig"ity predudes publidy expressing their himself. Large-ecale or ceremonial occasions are not the only relevant
own sentiments or arguing on their own behalf. ones for verbal lntermediaries. For example, an elderly woman with an
:

L46 Judith T. Irvine Registering affect 147

eye problem - someone I knew quite well - came to me for medication, dressed, sits with downcast eyes, silent and motionless, demonstrating
bringing along her young nephew so that she would not have to petition her strength and self-control to the extent that she remains unmoved by
me directly. "She needs . . . this thing," announced the nephew loudly, the barrage of angq, noise and activity directed at her' And the high-
pointing to a tube of eye medicine, while his aunt stood by in silence. ranking cowives, whose anger is being expressed, usually also remain
More striking examples for a Western reader, perhaps, than interme- quietly in the background.E
diaries who relay requests, announce decisions, or conduct political ne- These texts are examples, then, of "giot speech'" Uttered at high
gotiations are cases where the intermediary's role is to express strong pitch and high volume, their rhythm and articulation are nevertheless
emotion. One afternoon, a group of women (some five nobles and two very precise, and their sentences are well formed. Notable as well is the
griots)"*€re gathered near a well on the edge of town when another .rr" of emphatic forms, insulting labels ('crazy person', 'leper', 'witch',
*otnffiBt ode over to the well and threw herself down it. All the women 'thief'), vivid description (including "quoted" wailing), and the redupli-
were shocked at the apparent suicide attempt, but the noble women cated jgdi-a-jeex'all thoroughly worn out'. In the free translation, I have
were shocked in silence. Only the griot women screamed, on behalf of tried tffilicate the emphases as much as possible by underlining em-
aLL.26 phasizefforms. I have not included the many repetitions of text lines.
A
Examples: the registers in use
Man du ma wujjee ak ab doff! Lee
I (emph) neg I be cowife with det. crazy now (elided form) she
Now that the linguistic features of the registers have been described, it I won't be cowife with a crazy person! Now she's whipping me
remains to consider some recorded instances of their use. I have chosen ma ritax aw yett! Sa-i baaY woo na la,
three examples: two that represent public, mixed-caste situations and me whip det. stick, cane your pl. father call you
one that comes from a rno." ir,ti-ute gathering of castemates. ThSublic with a cane! Your fathers summoned you,
cases illustrate broad contrasts between the behavior of high and low d€f6 xonqx, kii d€f6 nyuul
castes: All the talk occupying the public "stage" is spoken by low-caste this person (emph) expl. red this person (emph) expl' black
persons, qnd all is "griot speech"l high-caste sponsors and addressees tftis [child] is red, this [child] is black -
are completely silent. The types of affect - anger and enthusiasm - are - toskln ginaar laa ci xamoon! Ling66r
quite different in the two cases, however. In the more private context have hatchlings chicken obj. focus+I about knew queen
represented by the third example, the behavioral contrasts are less ex- I've only heard of that when chickens hatch! [Even] the queen
treme, but we can see both registers emerge even though everyone pre- di xaxaardoomam- di wax/
cont. insult child her cont' about say
sent is of the same caste.27 saying,
would insult her own child [if she behaved like that] -
waa waa waa! lu ma xam, laa wax! Kii
Example 1. The text materials that follow are excerpts from an insult ses'
waa waa waa what I know obj. focus+I about say this person (emph)
sion at a high-caste6redding (all come from the same wedding). At these
Waa waa waa! What I know, I'm going to tell! This person
sessions, low-caste women, especially griots, insult the bride and her
ma wujj! Ab sacc la! Bdjjoam
kin, expressing jealousy and anger on behalf of the bride's high-ranking not worthy I cowife det. thief focus womb her
cowives. Lines of text initially shouted out by a low-caste solo speaker isn't worth being my cowife! She's a thief! Her womb
are repeated many times by a chorus of low-caste women, who are even- neexul! Boo debb, doxatt! Xari,
tually joined by most of the audience (the assembled public). The meter displease when you pound (grain) fart [bride's name]
of the text lines is picked up and repeated by griot drummers, whose is disgusting! When you pound [grain], you fart! Xari [bride],
rhythmic energy is supposed to stir the crowd to ever-higher levels of ngga ber ngatan, ba sa-b sirr
angry shouting, creative verbal fluency, and aggressive, sexually you put aelde [child's] cot till your mattress
suggestive dancing. Meanwhile the bride, to whom this tirade is ad- you htvo put alide your childhood cot, now your [big] mattress

:ffs
:

148 Iudith T. Irvine Registering affect L49

jeex-a-jeex! di I i. The griot's quotation of a proverbial expression, jdmbaar xam na la


be totally finished xeex di faj, shows him speaking on behalf of noble forebears
- the source
is all worn out! of proverbial sayings - at the same time as he addresses the particular
B noble sponsor and his ancestors.
|abari faj-mer la - nun, ka nu b€ggoon Maissa Moxurija Ali, Moxurija Ndumbe ]eey, Mori Taaw Njaay Fall
wife of assuaging anger obj. focuswe (emph) that person we wanted [name, extended] [name, extended] [name, extended]
She's the consolation-pize wife - the person zpe wanted
dikkul! Xari C-, jabari faj-mer ngga - nyoom/ warees na l66n-a tudd
.&five not [bride's name] wife of assuaging anger you (obj. foc.) yes them (emph) ought hort. you pl. * obj. focus name
t"didn't show up! Xari C- [bride], you're the consolation-prize wife
- Yes, them, you really ought to name them
nun/ la nubdggoon dikkul! b.q,$aaxabaax! Ndax nyoom s66n maam
we (emph) obj. foc. we wanted arrive not effingly well, thoroughly because they (emph) their ancestor
- the one we wanted. didn't show up! in rfrost glorious detail! Because it is their ancestor
C la nyu dakkilin tabala! Sinniko soxo
Xari yow, doomi setub
gaana, d€me yi! obj. foc. we rapidly beat drum for praise hymn fire off bullets
[bride's name] you (emph) child of (pl.) leper grandchild of witches to whom we drum paeans of praise! Fire off the sacred bullets!
Xari [bride], you are the child of lepers, the grandchild of witches! bi yaram! Nyu n6 ko borom ginnaay! Nyu n6 jimbaar xam na
sacred they say him owner battlefield they say hero knows
Example 2. This example is an excerpt from a griot's public performance For he is called "Owner of the Field of Battle"! He is called
of praise oratory on the occasion of a feast sponsored by a weffthy no- la xeex di fajt Alpaaq la
ble. The noble sponsor remains secluded (thus, silent and invisible) in- what fight cont. cure
side his house, while griots and guests assemble in the compound yard. "Hero Who Brings the Battle to a Glorious Close"! Alpaaq [name]
A griot.orator declaims the genealogy, family history, and great deeds nyibbisi, Alpaaq s66n maam! Ba mu demee ca xar6 ba
of the sponsor and his family with high pitch, loud volume, and great here [name] your (pl.) ancestor when he went
return to war that
rapidity; from time to time, he pauses while a chorus of other griots is incamate in you, Alpaaq your forefather! When he went to war
sings a verse of more conventional praise for the sponsor's patrilineage. - bi mu metti ba metti - mui d66r rek
The more enthusiasm the griot can express, and the more he can stir the , which it hurt till hurt he cont. strike only
crowd to admiration of the sponsor, the more moved the sponsor will - where pain piled on pain - he had only to strike and [his
be and the more enhanced will be his rank (because of the recognition dee! Nyui
daanu! Nyu baribari di
by others, and because of the immanent presence of ancestors ad- they cont. die they cont. fall they multitudinous cont.
dressed and invofied by the gnot). What the noble is moved to is more enemies,l they would die! They would falMn multitudes they would
effective social action, including the greater display of material gener- |imbaar, mooi xam
osity. tremble in frenzied fear Hero he (subj. foc.) cont. know
The following text represents a few moments of soio oratory. It illus- tremble in a frenzy of fear! The hero, he knows
trates "griot speech" as the expression of intense enthusiasm. Note that la xeex di fajl
to use extended forms of personal names and to invoke ancestral names what fight cont. cure
what brings the battle to a glorious close!
- both of which are done here - is to express special enthusiasm and
praise. Other characteristics of the discourse include emphatic forms,
reduplicated forms (baaxabaax, metti ba metti, baribari) and parallelisms, Example 3. The following conversation is drawn from a much more pri-
and much description of motion, often with the continuative auxiliary vate, caEual altuatlon. The assembled company - a small group - are all

,*
."

150 Judith T. Irvine Registering affect 151

castemates (nobles),2e and nothing earthshaking transpires. Still, the Chief: li ca - li nyu-i ci - jog6 ba xam - [pause]
conversation illustrates, in a small way, some of the differences between what at what they about set out till know
registers and how the registers can be drawn on according to topic, af- what - about what - get to know - [pause]
fectivity, participants, and other emergent properties of a situation. li ci de wonee -
The setting is the village chief's courtyard one summer evening. Sev- what about emph was shown
eral villagers of noble rank, the chief himself, and I are sitting around what was indeed shown -
chatting; there is no special reason for the gathering other than sociabil- MT: A! A- a- am nggoonent, da nyu fi musa am
ity. After some trucks are heard in the distance, the conversation turns an afternoon expl they here once have
to the.&grrying skills of drivers. MF, a young man whose family belongs Ah! w- w- One afternoon, they once had here
to a fuyal patriclan, claims that drivers who work for the nearby phos- benn, benn, benn kii bu fu - bu ko musa riix ci
phate mine are better trained than others. In support of this allegation, one. one one machine rel. where rel it once stick in
*rfulachinewhere - that got stuck in a
MI, a man whose family, although of noble caste, is a dient of the chief's,
entertains the group with a short narative. Thus MT, perhaps the lowest-
g"ffi kekk. [pause] kauytopp ko, toPP-topP-toPP-toPP-
one(aug) mudhole come push it push continually
ranking man present, takes on a griotlike speaking role with his narra-
abigmudhok [They] come and push it, pushing pushing
tive as he attempts to lend persuasive support to MF: [A note on tran-
topp-topP - mu nyingg - mu nYim6
script conventions: I I : oveflapped transition; large brackets : simul-
taneous utterance.]
pushing it suddenly active it manage
pushing - it suddenly moves - it manages to
MT: Xamxam ci li nyu xam n6, nyoonYu! tabasiku!
knowledge of what they know those people (emphatic) suddenly disengage
The knotoledge they have, those people! burst out!
MF: E - Ting66j [? indistinct] KN: A bonnak -
[name of town, also known as Rufisque] well thus
Etr-- [the ones in] Rufisque. [?] Well then -
MT: Xamxam, xamxam ddll la nyu ko am, ba m' ko doyl6 MT: Bon nak, bu nyu jubul6, demal ci biir nger/
knowledge knowledge really they it have till it suffices well thus when they straighten go caus. in middle highway
They really have enough knowledge Well then, when they straighten up, they make it go in the
ci digginte, digginte, aall middle of the highway,
between between kii bu nyu-i law!
machine rel they cont. touch
[to drive] between - ah -
the machine they were handling!
Chief: // [indistinguishable] Mariama, y6w fi
p UTI] youhere MT's narrative uses a number of emphatic devices: the particle drill
Mariama []TIl, you 'really' , the numerals benn and genn'one' (outside any context of actual
de woon sax. [pause] Fekk fu kenn du nyu-i xam fu - counting), an augmentative noun class marker (gi, incorporated in the
emph were even find where noone they know where numeral genn) for lcelck'mudhole', syntactic focusing devices, repetitions
were here then. [pause] Turns out nobody knows where - and the reduplicative form topp-topp 'continually and energetically push-
MT: digginte Ting66j ak Ndar. ing', and vivid vocabulary. He repeats MF's tingddiand KN's bon nak;
between [place] and [place] and he uses a dramatic Prosody, with wide-ranging pitch and stress con-
between Rufisque and Saint-Louis. tours. In contragt/ the utterances of the chief - apparently the main tar-
- display
Bet of MT'B rhetorical effort, the person who is to be persuaded
a flat intonaHonal contour and what might be considered conspicuous
diefluenry, wlth hlr mumblint and incomPlete constructions.3o
..

152 Judith T. Irvine Registering affect 153

Even on a quite ordinary social occasion, then, and when the speaker
is not himself a griot, the attempt to persuade and entertain one's inter-
Conclusion: complementarities and dialogic relations
locutors carries implications of rank and bears overtones of the griot's The system of sociolinguistic registers I have outlined in the preceding
affectively charged, theatrical manner. To put this another way, the at- pages is based on cultural assumptions that link affectivity inextricably
tempt at persuasion is an effort to bring about affective involvement in with social structure - and therefore with the discourses in which society
one's listeners. Since this is a griot specialty, even the noble speaker enacts, constructs, and reproduces its relationships. For Wolof villagers,
echoes a griot voice. social differentiation, conceived as a differentiation of ranks, is funda-
mentally based on the differentiation of affective modes and the control,
t.r" expression, and channeling of affectivity. Although rank has many di-
Sumrp-ri1
mensions for the Wolof, including differences in wealth, political au-
In sum, I have suggested that the linguistic phenomena comprising two thorityr $[:nealogy, and moral stature, all of these are connected, in the
labeled Wolof registers cohere in terms of affectivity and ideas about the ideologffithe system, with affect, impulsivity, and self-control. Like
nature of high and low ranks. These phenomena concern all aspects of their neigtbors the Fulani, described by Paul Riesman (1983), Wolof no-
the linguistic system, although some asPects are more salient to Wolof bles rationalize their claim to superiority over lower ranks in terms of
villagers than others; and the principles governing the differences be- sangfroid and restraint.3l
tween these registers also reach beyond them to organize patterns of Despite the moral weight nobles attach to differences in emotionality
social contact and conversation. There are no absolute boundaries be- and its expression - and their view on these moral implications is not
tween these registers and no linguistic feature that, in isolation, categor- wholly shared by the lower ranks - no Wolof villager suggested that the
ically identifies a stretch of talk as belonging to one variety or the other. high ranks could be self-sufficient or that society could function if every-
Yet, the varieties are systematic: They reflect consistent patternsltf con- one behaved like the highest nobles. As members of this Wolof com-
trast that emerge in the course of interaction. munity see it, society rs differentiation, and high rank cannot be ex-
Thus the characteristics that distinguish "giot speech" from "noble pressed if there is no one to take on "lowering" tasks. As one high-
speech"all involve contrasts in affectivity (hyperbolic versus restrained) ranking elder argued, only a balance and an interaction between the
and rhetorical elaboration. They are rooted in images of persons, whose high and the low, the calm and the impulsive, can make social action
voices a speaker echoes when he or she echoes their social roles. These possible. Comparing the impulsive energy (fl, residing in the liver) that
images are coherent, even if they are sometimes attenuated, echoes of different personalities and ranks possess and control in different de-
another's voice. The co-occurrence of linguistic characteristics in a reg- grees to the engine of a car, he asked me: "Which would you rather
ister, and the association of register with social role, are not arbitrary. have - a car that does not move, or a car that can only go a thousand
Instead, they are products of a Wolof cultural association between social kilometers an hour?" (I didn't know.) "Neither one. One would crash,
ranks and affectivity. and the other would never start. Either way, you would never get to
Whether this kipd of connection would apply to the sociolinguistic Dakar." What you need, he concluded, is both a strong engine and a
organization of registers in other cultural systems is not clear, and space good set of brakes. The two are complemenfary in that both are neces-
does not permit the examination of this issue. Some interesting cases for sary for the well-being of the whole and its potential for constructive
comparison might include |avanese language levels (mentioned earlier); action.
Samoan registers (Ochs 1988); registers in Fijian Hindi (Brenneis 1988); In Wolof discourse, this complementarity has several facets. The first
and perhaps, from European material, the linguistic expression of the of these is the complementarity of "personalities" or affective images
"royal rage," or heroic anger/ if information is available. These cases associated with different social ranks, the high and the low. I have ar-
make convenient comparisons with Wolof because all involve relations gued that the registers described here represent the voices attached to
between affect, speaking, and social rank - although the particulars of those images, voices that a speaker takes on in different social situations
the connection can be expected to differ. and for differcnt purposes, whatever his or her own rank. The second
,"

154 Judith T. Irvine Registering affect 155

host,s behalf. Again, although the discourse is charged with high


affect,
facet is that these voices necessarily interact: They are mirror images of
the speaker's olin feelings are not the ones principally being expressed.
one another, and like mirror images they cannot exist in isolation. Al-
though there can be conversations in which register differences are min- t have called all these relations - between contrasting registers, be-
imized, there is no Wolof discourse in which everyone present uses, tween the contrasting social personae they represent, between the con-
say, extremely "griot-like" speech; even if all the audible talk that oc- trasting roles organizing a social situation, and between the expressor
curred were in the waxu gewel regSster, the silence of the audience would and th-e owner of a sentiment - "complementarities." This term ex-
presses a Wolof sense of mutual necessity about these contrasts,
but it
represent the waxu gdir side of the coin. It is the contrast between them
is not otherwise ideal, for it may suggest that there is no difference in
that organizes Wolof discourse, and organizes it to some effect (as, for
instal&,lrwhen the griot animates a noble and moves him or her to ac- power or value between the contrasting forms' Moreover' it may sug-
gest a certain social completeness and harmony, a sense that all
of soci-
tion,disimply when a speaker takes on griotlike rhetoric in an effort at
persuasion). eWhasbq:ncomfortablyaccountedfor,thatmaysuittheviewsofWolof
As the prototype of the verbal intermediary - the message bearer - .,obt", Ufun.ot suit everyone else. In particular, the jaam- the former
here.
the griot faces in two directions: toward the message's soutce and toward slaves - ulfr poorty accommodated in the set of registers described
Aligned with griots in many respects because of their low rank' aligned
its destination. It is the same for the griot's emotional expressiveness
and for hyperbolic speaking that takes on the griot's "voice." The griot witl nobles in other respects,32 but unlike either in having been unfree,
the jaamin some sense perhaps have no "voice" in the wolof
system at
is both excited by others and exciting to them: There is a double comple-
mentarity. On the one hand, we have the relation to the addressee, in all.
,,Heteroglossia,, and ,,forms of dialogicality,, might be better terms to
utterances that attribute some emotion or personality to him or her and
describe the discourse relations in this system. Although they
do no
stimulate the addressee and/or the audience to some feeling and action.
better than
,,complementarities" in attending to differences in power or
The griots' praise oratory is a good example of discourse orientedtn this
value, at least they do not seem to imply that such differences do not
direction. Extravaganfly praising an addressee, it arouses feelings of family
What is
pride. To experience those feelings intensely supposedly "strengthens" exist. or that structured coexistence is necessarily harmonious.
useful about them, too, is that they call attention to the fact that dis-
the addressee and moves him or her to praiseworthy acts (such as dis-
course is a matter of relations - a matter of social relations and
social
tributing largesse). The audience, too, is moved and persuaded o^ the that
respectability of the person being praised. Although the discourse is re- situations. In doing so, they invoke an anthropological literature
calls into question monologic models of language and their
analogues in
plete with signals of high affect, it is not really the griot's own feelings
models of society (Hill 1985). It is clear that monologic models of lan-
that are particularly at issue (see also Appadurai, this volume).
guage, for all their usefulness in other arenas, cannot comfortably ac-
On the other hand, we have the relation between a speaker and the
we have
source of the message: the person whose ideas and sentiments the in- comhodate an organization of registers and their uses such as

termediary expresses. In the Wolof communicative system, the dis- seen in Wolof.
player of affect (ora"the person who expresses an idea) need not be the The affective registers I have described in this chapter concern the
same person who supposedly possesses it. A griot may display emotion organizationofverbalmeansthroughwhichWolofspeakersdisplay
dis-
on behalf of a noble, to whom the emotion is attributed but who sits by emotion and in terms of which wolof audiences interpret emotional
impassively. If a high-ranking woman remains silent while griots ex- play.IhavenotmuchtreatedquestionsaboutwhatemotionsWolof
press her anger at her husband's having taken a new wife, the sentiment ,p"ut"rt "really feel" or to what extent such questions' which assume
of its cultural
is nevertheless attributed to her, not to the griots themselves. The songs some level of affectivity independent of its expression and
construction, make sense. I would like to conclude by pointing
out'
and wailing of the griots who serve as public moumers when an impor-
dialogical
tant person dies also express the family's feelings in this way; so too however, that seeing the discourse of emotion as essentially
an im-
does the intermediary who loudly exPresses cordiality to a visitor on a may have some implications about emotion itself - particularly

;ia:
;

156 Judith T. Irvine Registering affect 157

level of emotional universals and temperamental tendencies, he argued


plication that a view of emotion may need to be relational as well. An nevertheless that both the expression of feelings and the notion of what
approach that locates emotion only within the individual, rather than in feelings and personalities there are to be expressed are subject to the con-
the relationships between persons or between persons and events, sug- ventions of a cultural system and the organization of linguistic form. It is
only in terms of these conventions that the behavior of individuals can fully
gests a monologic discourse and will not apply well to the case at hand.
make sense.
As proponents of a contextual approach to emotion have suggested, 3. Lutz (1986) argues that these assumptions, and the contrast between emo-
the feelings one experiences in a given situation are not just primally tion and thought, are deeply embedded in the Westem world view of the
"thete," but are experienced in relation to something - indeed, to many person, of social life, and of morality. Although she wltes from the per-
spective of cultural anthropology, psychology, and philosophy, the rele-
things, including the possibilities of expression, the feelings one attri- vance to linguistics is obvious.
butes t&.irthers with whom one interacts, and the feelings one attributes 4. The terms nstyle" and "language level" are also relevant, although not
to otlfurs with whom one may be associated. Moreover, the possibilities identical to "register." "Language level" is used principally for cases like
]avangse, which has a labeled set of linguistic varieties ranked according to
of expression include not only one's own performance, but the perfor- refi'n&ent and respect conveyed for the addressee and/or the referent.
mances of others who serve as one's expressive vehicle. Actually, we "Strffiii a term thlt has .t utty ,rtuget, including situational variety, but
know little about the subjective dimension of these relayed perfor- als5 irftluding the notion of an individual's style (his or her personal speech
mances. If these possibilities of expression have different social conse- characteristici) and even (in the work of Labov) degrees of self-conscious
attention to talk.
quences, can the feelings they express be the same? 5. When Halliday (1978:35) says that registers tend to differ in semantics, he
To speak like a griot, if (for example) you are not one, is to forego the adds "and hence in lexicogrammar, and sometimes phonology, as realiza-
option of having your feelings expressed by a specialist whose perfor- tion of this." But he does not tell us how this works or why phonological
differences should be seen as the "realization" of semantic differences.
mance skills, you may believe, exceed your own. Instead, you take on 6. Consider, for example, the "language levels" of some Indonesian lan-
the mantle of the griot's supposed emotionality, and you contrast your- guages, such as )avanese. Named linguistic varieties conveying degrees.of
self with some more restrained interlocutor. Whether or not yousreally foliteness, refinement, and respect for the addressee and/or referent, the
feel" the particular emotion you display, your subjective experience pre- iavanese language levels are perhaps an especially formalized instance of
register diffeiencts. Linguistically, the language levels havebeen described
sumably includes knowing that you sound like a griot (about whose as differing mainly in lexicon (including sets of leical substitutions) aldin
emotionality you have certain beliefs). Your attitude toward griots, and some special affix-es. But recent work has pointed out that they also differ
in prosody and morphophonemics, although these aspects have been little
toward being for the moment associated with them, must color your
studied (Errington 1984:9).
feelings toward other aspects of the situation. If you are a giot, an equally 7. Although register is defined in terms of situational context, the connection
complex contextualization applies, including your attitudes toward the with affect and the cultural construction of the person has not been utterly
person whose feelings you perform and toward the person you arouse, absent from linguistic discussion of what registers are. Halliday writes, for
example, abouithe "degree of emotional charge" in participants' relation-
as well as your attitude toward the griot status you are for the moment ship is a situational factor contributing to registq choice (1978:33); and
typifying. Firth speaks of "treating personality and language in society as a sort of
With Wolof villagers, it would be difficult even to approach a discus- basis for linguistics with a sociological component" (L957 [1950]:189).
sion about the subfgctive experience of emotion without bearing some
8. See also the discussions of "sincerity" and affect in Irvine (1982) and Abu-
Lughod's discussion (op. cit.) of "hypocrisy."
of these relations in mind. It seems to me, however, that the point is not 9. The largest ethnic group in Senegal, the Wolof population now stands at
limited to this ethnographic case. One cannot experience a situation only over 2 million - urban as well as rural and elite as well as Peasant' Before
the French colonial conquest in the late nineteenth century, they were po-
in itself, without also being influenced by feelings about the social voices
litically organized in a iet of kingdoms whose history dates back to the
talking in it and about it, and by feelings about the roads not taken. thirte6nthientury, when a Wolof state fitst emerged within the Empire of
Mali.
Most of the fieldwork on which this chapter is based was carried out in
lgT};t,7975, and1977 ina village in the Pr6fecture de Tivaouane, Senegal.
Notes I also returned there briefly in 1984. Thanks are due to the National Insti-
7. See also Ochs 1988. tute of Mentel Health, the National Science Foundation, and Brandeis Uni-
2. Although Sapir did not exclude the possibility of some biologically based vereity for flnancial support of that research; to the Institut Fondamental

=re
Registering affect 159
158 Judith T. Irvine
cussed in detail in Irvine (1978) and is related to historical changes in the
d'Afrique Noire and the Centre de Linguistique Appliqu6e de.Dakar for linguistic system. Although the discussion,in that paper is framed largely
i*tit"ho"ut sponsorship; and to my viliige h6sts for their hospitality and in lerms of social categories (age, caste, and gender), part of the argument
intellectual assistance. rests on the idea that each speaker controls more than one variant of many
10. fnu *ofof hbels even derive from these groups: Literally, waxu gaoel means of the nouns concerned. The variation can therefore be thought of in rela-
'griot tatk' andwaxu 966r means 'noble talk'' tion to registers that carry different connotations of social identity and in-
11. See, e.g., Silla 1966. tention - Ihat represent different "voices" in the community a and are dif-
1;. N-rt"-ifi"i p;ople can also be ranked within a caste according to genealo8l- ferentially drawn on by speakers according to their own social situation.
cai puritv'rtta s"t i"tity, age, gender, wealth, and other factors'
although
need not be obvious' 22. The orthography system devel-
hy used in-this chapter is based on the official s,ystem dev
thebutc6me of all these criteria oped bv
ooed by the Centre Linsuistique Appliqu6e de Dakar and adopted by
lentre de Linguistique
13. eitf,r"ift nawlt tsualTy refers to similarity of,co-religionary'.
caste of the
Tu"k' 9":11te
For a discus- Seneeal in
the Republic of,f Senegal 1971. It is a phonemic system
in1971.It svstem whose phonetic
*dnur if,gr"aient it is s6medmes translated as
A few may require some explanation:
valuesire fairly transparent. symbols
tsicm of iaw and nnwl6, and of a wolof metaphysics that-links
perso_nality
and. ijare
c and oalatal affricates; among
are palatal 6:1e1, e:lrl,
amons the vowels, 6:le), e:1e1, c:l+f,
c:l+1, a=l,tl,
a=1,t1,
of the liquid elements of the body, see Marone
and status to the composition d:lal, o:Itl, d:[o].
(7e6e). 23. fhk End some other features of "griot speech" seem to occur less often in
so that
L4. ifi"r6 ry-Uolic connections also underlie caste-linked occupations, erioffiformal performance of historical narrative. There is some intersec-
tt affinityfor earth supposedly makes them esp-ecially suitable for tion d[
f;o.r resister, therefore, that needs to be more care-
genre flatures with register,
8t'n"n."
" "oui"r, hnd'administratio^ri, wheieas the griots'affinity for air is con'
irr*i"gi"a fully examined.
nected"with their verbal activities. Similarly, blacksmiths supposedly have 24. See, e.g., Irvine (1974) on Wolof greetings, Irvine (1980) on directives, and
," ,ifmty for fire, as do tanners (who work-w-ith.fiey and potters'
39iqt) Irvine (1979) on the organization of talk in political meetings.
(noble) ranks,
whereas iishermen have an affinity for water. Within the high 25. According to Gamble,-women of slave stafus may serve as intermediaries
hand,
distinctions are sometimes drawn between clerics (s6tin), o1 lhe one too. For s-ome other comments on griots as intermediaries, see Diop (1981)
;i;;;;, u.a on the other. In addition to their role as land
"o***ers, and Camara (1976).
ua#ttitt utors, the princ es igarmi) and, in somewhat different ways' the
26. See Camara (1976) for a useful discussion of intermediaries among the Mal-
commoners and military slaies (ceddo\ whom they command also engage
,,hoi" compared with that of th6 "cool" inke and the vicarious expression of emotion, particularly anger.
in warfare; thei. temf"iament i; 27. Because of problems of presentation, the transcriptions I provide_here do
clerics. not show piosodic featuies or phonetic detail. Their illustration of register
15. Also acts that cannot appropriately be witnessed by one's castemates or differences is thus limited to morphology, syntax, lexicon, and (to a small
supenors. extent) discourse management.
76. Ndte the expression nit kii dokna sama naw'this person has curdled my nau)' , 28. There is a distinction here between the bride's actual cowives (women al-
relftine to interaction considered emotionally destructive' ready married to the bride's husband) and,her,classificatory cowives (other
77. i,l;";ll"po1srHe reelings or expressive devicis.are caugh.t in this Particular women married into the husband's household - e'g., his brother's wives).
the ones .oristit tl.tg ilnage.s of nobles and griots and thus rele' The classificatory cowives may take part in the performance, usually as part
""t, ""fy
t" in" t*o registers ,tu*"i after"them' In an earliei-papgr ({y.rne. !42)' of the crowd who eventually repeat lines of text.
I outlined some alpects of the linguistic expression of afleCt in Wolof. O-nly
"u"t
29. Because I myself was classified as high caste, the only situations I observed
il;ddiilrsion 6f prosody ur,d'discor.ie management, however, did I being de' that I can characterize as completely caste internal are those where every-
ufi"i" to the registeis of ""griot spee-ch" -an{ "noble -sPeech" one else was high caste too. Nb situations in which I was present to record
,.JU"a f,"r"- Atlough som; of the other devices mentioned in that paper and take notes can be considered wholly low caste.
ire included here, no-t all of them clearly sort out into these twolegisters. 30. It is perhaps a little unusual that the chief speaks as much as he does,
Pitch is a partial exception. In male speech one can distinguish falsetto
from
18.
with extrime renderings of disflrient oi not. Notice, howevet, that he addresses me, not the other vil-
;;;il;5t." nnJi"ir"it. is linked exclusively
;'triot t rr.-;; fr-ich distinctions also apply wiihin lagers, whom he more clearly outranks.
l,t1TI-.T.*-1,,h::"L::' 31. Ahong the Fulani studied by Riesman, the lower ranks consist of former
,ffi;;-i"".Lved in the less extreme veisions of the "griotlike" register.
slaves;"among rural Wolof, there is a broader range of lower ranks, includ-
in"*, ulso applv within normal voice for female speakers, although one ing slaves as well as other grouPs. Wolof discussions of emotionality usu-
forri of femitb ,rtt".un." representing extreme emotionality is ululation. ally focus on griots rather than on slaves.
t9. if," nignert rate I have recorded anJdone a_syllable count for came out to 32. A-jaam is a person whose ancestots were once higher-ranking but were
386 sviiables per minute. These rates are maintained over several
minutesi
not included in dehoted, either because they were captured in battle or because they sold
;il;y il;l"d; iurru. for breath intake. Were such Pauses themselves for material goods. Should any iaam manage to buy back their
the count, the rates would be higher. freedom - theoretically, atthough seldom practically, a possibility - t!"y
20. See later comments on the stammer of the king of Saloum' would ascend to their ancestors'free rank' Since slavery has been illegal in
27. i ."r,
,,ru"*, to,, because it is somewhat difficult, for statistical reasons, to
Senegal for a considerable time, these principles represent the conventional
i"1 ;h"t# the particular type of variation primarily characterizes.situa' calculation of nocial rank, not a legal status.
iio"r oi social grbups. Variition in Wolof n6un clas's morphology is dis'
160 Iudith T. Irvine Registering affect L5L

Lutz, Catherine. L982. The Domain of Emotion words on Ifaluk. American Eth-
References nologist 9:113-28.
1986. Emotion, Thought, and Estrangement: Emotion as a Cultural Category'
Abiven, Pdre O., compiler. n.d. Annales religieuses de saint-loseph de Ngasobil, Cultural Anthropolo gy 1':287 -309.
1849-1929. Dakar: Archives, Archev6ch6 de Dakar'
Lutz, Catherine, and Gebffrey white. 1986. The Anthropology of Emotions. Azr-
Abu-Luehod, Lila. 1985. Veiled sentimenfs. Berkeley: University of California Press' nual Reoiew of Anthropology 15:405-35'
r"e*u"l william o. 1988. Affectivity in Persian Language use. culture, Medicine Marone, Oumar.'1959. Essaiiur les fondements de l'6ducation s6n6galaise i la
and Psychiatry 12:9-30.
lumiEre des m6taphores aqueuses de la langue wolof.. Bulletin de l'lnstitut
Brenneis, 6onald. 1988. Shared and Solitary Sentiments. Worklng Papers and Pro'
--- itriirgt Fondnmental d' Afrique Noire 37, s6t. 8:787-852.
of the Center Psychosocial Stidies 23. Chicago: Center for Psycho-
for Myers, Fred. 1979. Emotions and the self: A Theory_of Personhood and Political
- Order
social Studies. Among Pintupi Aborigines. EthosT:34b-70.
Camar&Fory .1976. Les Gens de la Parole. The Hague: Mouton'
La Soci,t€Wolof : Tradition et Changemenf . Paris: Kar-
Ochs, Elinor. 1988. Cul[ure and Language Deaelopment. Cambridge: Cambridge
oiop*{6aorilaye-Bara. 198'1,.
University Press.
thala.
language and social Change in Jaaa. Athens: ohio Uni-
Riesman,f,aui. tgsg. on the Irrelevance of Child Rearing Practices for the For-
Errington, Joseph. 1984. ma-Hi* pf Personality. Culture, Medicine, and Psychiatry 7:10T29.
v:ersity Center for International Studies.
Sapir, Effitd .1921,.laiguage. New York: Harcou* Brace.
Ervin-Tripp, Susan. 1972. On Sociolinguistic_ Rules: Alternation and Co-Occur- ieZZ. Sib'ecfr as a Persdnality Trait. American lournal of Socioloy 32892-905.
."".i.'i" f. Gumperz and D. Hym1s, eds., Directions -in Sociolinguistics: pp.
The
1930. The Cultural Approach to the Study of Personality. !up"l delivered at
Ethnography of Communicafion. New York: Holt, Rinehart, & winston,
the Hanover Confer6nce, social science Research Council. To be published
21150. in The Collected Works of Edwatd Sapir. Berlin: de Gruyter.
Fernandes, valentim. 1940. o Manuscrito "valentim Fetnandes." Lisboa: Acade- Shweder, Richard. 1984. Anthropology's Romantic Rebellion Against the En-
mia Portugesa da Historia. (From a ryo-rk dated 1506-10') lightenment; or, There's More to Thinking Than Reason and Evidence. In
Firth, J. R. 1957. (1950). Personality and Language in society. The sociological Rl Schweder and R. LeVine, eds., Culture Theory: Essays on Mind, Self , and
Reaiew 42(2):8-1'4.
Emotion. Cambridge: Cambridge University Press, pp. 27-66.
in Linguistics 1934-L951. London: oxford University
1957. Papers Press.
silla, ousmane. 1966. La persistance des castes dans la soci6t6 wolof contem-
CamUte, David. tgdZ. flrc WoIot' of Senegambia,2nd ed. London: Inte,6Srationaf
poraine. Bulletin de l'Institut Fondamental d'Afrique Noire 28, s6t.^8:731-69.
African Institute. Staniciewicz, Edward. 1964. Problems of Emotive Language. In T. Sebeok, A. S.
Geertz, Clifford. 1960. The Religion of laoa. Chicago: university of Chica8o Press. Hayes, and M. C. Bateson, eds., Approaches to Semiotics. Transactions of the
C"*plrr, John. 1968. The Spiech Community. International Encyclopedia of the Indiana University Conference on Paralinguistics and Kinesics. The Hague:
Social Sciences. New York: Macmillan.
Mouton, pp.239-54.
--
Halliday] Michael A. K. 7964. The Users and Uses of Langt'agle' In M' A' K'
Wolff, |ohn Li., and S. Poedjosoedarmo. 1982. Communicatiae Codes in Central
Hrifauy, A. Mclntosh, and P. Strevens, eds., The Linguistic Sciences and Lan-
guaxeTeaching. London: Longmans, pp. 7!-110' lapa.Ithaca, NY: Cornell University Department of Asian Studies, Southeast
Asia Program, Data Paper 116.
DVS."Languageis Social Semioti. London: Edward Amold'
H;;r, Jute?. r'sgo. The Linguistic Expression of Emotion. American Anthropolo'
gist 38:250-6.
'i."*-u..1985. The Grammar of Consciousness and the Consciousness of
Hi[,'1ane.
American Ethnologist 12:725-37 '
Irvine, Judith T. 1974. strateg;el of status Manipulation in the
wolof Greeting.
- il it. Bur*u. and ]. Shelzer, eds., Explorations in the Ethnography ot' Speaking.
London: Cambeidge University Press, pp. 167-91"
1978. Wolof NouffCiassification: The Social Setting of Divergent Change. Lan'
guage in SocietY 7:37-64.
tg7i. "Formality"and Informality in Communicative Events. American Anthro'
pologist 81':773-90.
fSbO. tlow Not to Ask a Favor in Wolof. Papers in Linguistics 13:3-50.
iOai. f-i"g"rge and Affect: Some Cross-iultural Isiues. In H..Bymes, ed.,
Contemp"orafi perceptions of Language: lnterdisciplinary D.imensions. -George'
town Univeisig Rbund tiUte on Langr4gel and Linguistics. Washington,
DC: Georgetown University Press, pp. 3147 '
yakobson, Rorfian. 1950. Linguiitics and Poe_tics. In T. Sebeok, ed., Siyle in Lan'
guage. Cambridge, MA: MIT Press, pp. 35f77.

?G
Language in the discourse of the emotions 1,63

chological anthropology, is also a period in which they have cham-


pioned a revamped picture of culture, not silently univocal but audibly
8. Language in the discourse multivocal. Accordingly, the contributors to this volume assign "the
emotions," or a particular rhetorical construal of self and emotion, to
of the emotions one or another class of speakers in particular discursive contexts. All of
them, more or less explicitly, locate their efforts on a certain trajectory,
DANIEL V. ROSENBERG in which a transformation is said to be under way in the role language
must play as foundation and datum for cultural accounts.
i,
,s!+
rJ In what follows, I look at some claims made for and uses made of
;a,--
language on the part of those who seek a "discourse of the emotions."
In exarpi4ing some signal contributions of the last decade, I argue that
the tum\#discourse" risks preserving much of what it aims to supersede.
"Discourse" and "the emotions" I claim th?t these contributions represent cooperative native and ethno-
"The emotions" are a lexical domain in oui Anglophone consciousness graphic reflections of lexical reference, and that such reflections are lin-
of the person, a higher status label for that family of objects known in guistic data, to be sure, but not discourse. I suggest that referential ide-
the folk ethnopsychology as "feelings." "Discourse" is the consensugl ologies of these kinds must be located in a field in which sharper
best name for what the human sciences today make of their perennially distinctions are made among semantics, reference, pragmatics, andideology.
most reliable object - human symbolic behavior, especially as encoded
in language - when this is conceived not as a fixed superstructure of Knowledge and Passion
collective meaning and order but as a fluid field of interested,lndeed
contentious and factious, social activity. In this chapter I consider what Michelle Z. Rosaldo's Knowledge and Passion (1980) prefigured the kind
anthropologists have discerned and might be able to discern at the in- of account of "the emotions" in terms of "discourse" that characterizes
tersection of "discourse" and "[the] emotions." many of today's efforts. In that text, Rosaldo offered a sustained meta-
A conception of culture that anthropologists came about a decade and commentary on the kind of ethnographic datum language must be. Lan-
a half ago to reconsider, and then largely to reject, had its foundations gu.age, she claimed, does something more than just "demarcate" the
in a much-distilled inheritance of certain Genevan and Praguean under- world into "classificatory gids"; "meaning," she insisted, "is bound up
standings of linguistic structure, especially as that structure might be with use." She urged us to rely instead on "sentences and style," "com-
understood to emerge in the parallel segmentations of language's sen- mon discourse," "habitual ways of talking," "datly life and talk," and
sible medium - sound - and of meaning. In turning from that conceP- the like. In taking what she called this Wittgensteinian turn, Rosaldo
tion, anthropologists ceased to render culture as a set of freeze-dried intended to find "the emotions" in the language of ordinary life and not
crystallizations of gymbols and meanings, and went on to reconstitute it in their famously problematic ritual embodiments. In addition, she ar-
as a field of oblique communicative vectors - discourses - truer to life, it gued that "the emotions" could not be encompassed by the methods of
seemed, even if threatening to undermine the neat order that the earlier the so-called ethnographic semantics, and its successor cognitive an-
ethnographic practice had been able to impute. These anthropologists thropology, of the American anthropology of the 1960s and 1970s.1 The
intended a rediscovery of interested human agency, as a feature of will' practitioners of these methods, as many have noted, reasoned that a
ful individuals and as a feature of social segments - genders, classes, culture could most parsimoniously be captured by selecting from its lan-
castes, and so on - or of corporate institutions, such as the state, the guage, in grammar-free and context-free fashion, classes of nouns (oc-
men's house, or the women's domestic grouP. casionally, adjectives), which in durn represented object-classes that cor-
The period during which anthropologists have settled on "the emo- responded with domains in the experiential world. The promotion of
tions" as a domain of analysis, as virtually the domain of choice in psy' such a lexical gloaeary from datum to culture was effected by passage

162
L64 Daniel V. Rosenberg l^anguage in the discourse of the emotions 165

thr'ough an intermediate step, a cognitive "model" or "representation," without single quotation marks; or of an English gloss, with or without
whictr typically took the form of a spatial arrangement of the lexical items. single quotation marks) in sentences in English grammar. There is, in
When Knowledge and Passion was published, that style of analysis was my view, more than the usual ethnographic ambiguify about whether
falling from favor under several influences, including the turn to prag- each sentence is to be understood as a kind of quotative, with an implied
matics and context on the part of linguistic anthropology, the discipline "Ilongot say that . . ." atits beginning, or whether some kind of author-
that ostensibly had bred it; cultural anthropology's passage from or- ial abstraction has been made.
dered, timeless knowledge to disorderly macro- and microhistorical pro- The preponderance, grammatically and letcally, of English carries with
cess; and cognitive psychology's (lately, "cognitive science" 's) replace- it a radical infiltration not just of the world view encoded in the ethnog-
men;6Iiirodal treelike representations of passively held knowledge (again, rapher's language, but of the ethnopsychology of the ethnographer's
alsoTilgely lexical and spatial) with real-time processual schemata for culture. In Rosaldo's account of Ilongot emotional life, 'liget' and its
quotidian engagement. Ilongot.ftllows inhabit the same world and the same Srammar inhab-
I(nowledge nnd Passion finds "the emotions" embedded in ordinary ited bffiftr ry i' " distractiort," " col7rage," " itritation," "confusion, "
process.2 In place of the rigorous methodological abstraction from the "separaffon," and"ioy." \tVhat is more, the Ilongot names/things accePt
ordinary on the part of ethnographic semantics, Rosaldo practiced the same range of predicates and modifiers that the English terms do.
"interpretation," which amounted in this case to a rich narrative about some of these modifiers follow the inherent tendency of English to spa-
how certain Ilongot "emotional terms" - especially 'liget' ,loosely trans- tialize, quantify, and generally'physicahze' all nouns. Others seem to
latable as 'anger', also as the titular 'passion' - were instanced in dis- derive more directly from the tenets of American ethnopsychology, with
course about the self and its orientation to important institutions, partic- its idioms of causation and motivation nested in near-animate forces
ularly kin relations, labor and economy, politics, and headhunt{tg. subject to hydraulic Pressures. Thus, 'liget' and its kind have 'lweight";
But in spite of the claims for a focus on everyday talk, for the fficourse their "energy" can\be "diffused" or "concentrated"; if they are weighty
of daily life, Rosaldo retums repeatedly to a family of "emotional terms," enough, or if their "wildness" can be "controlled" and "focused," they
"affective terms," or "focal analytical terms." These are mainly nouns. are "loosed"; they may otherwise be suppressed; "derived" from or
Her abstraction of such nominal foci from everyday talk is no less radical "bred" by other inner states or public circumstances, they may in turn
than the earlier ethnographic semantics abstraction, howevet much it be "creative" and "transcendant," "beating fruits" and "stirring,"
may be less mechanical and conspicuous. And, despite a number of claims "breeding," and "spurring" their own outcomes; capable of being "in-
for a focus on functions of language beyond the referential, the PurPose tense" and "disruptive," they need to be "dealt with"; and so on. Dis-
of the abstracted lexical foci is to stand for things. 'LIGET' and 'BEYA' tinctions between use and mention, between the native's 'energy' and
are the essences of which 'liget' and'beya' ('knowledge') are the names the ethnographer's energy, are suspended. 'Liget'is embedded in a web
or labels. To be sure, the kinds of things to which 'liget' and'beya' seem of social-scientific idioms: It is "labor and reproductive" force marshaled
to refer are not the denotata demarcated in the earlier picture. One of in the "interest of collective life"i it can be "bred" by "inequaltty" } and
the challenges injgeading Rosaldo's work is to reconcile one's sense of so on. Finally, there is a stunning breadth of consequences brought about
'tiget' and 'beya' as quiddities with the author's arguments against our by 'liget', a vast range of contexts in which it is "expressed" and "dis-
conceiving of them as denoted objects. played."3
Rosaldo,s reasons for wanting to reject the notion that meaning is an Rosaldo's announced foundation is language in context, but Knowl-
ostensive relation between word and object are fine. Yet 'liget', 'beya', edge and Passion is an effort that abstracts from language and then reco'n-
and their kin seem to come to life, first as essences and then as seeming texfuatzes the abstracted product, an effort we must see as distinct from
psychological forces, in a species of bilingual ethnographic writing in an understanding of language in context. ln lfuowledge and Passion,lan-
which the "focal terms," having been glossed early in the text, become guage is a symbolic medium fiom which certain things - words - are
so familiar to the reader that they can be used throughout the remainder abstracted and made to stand for certain other things. What separates
of the account (in the form of the indigenou8 word in italics, with or Roealdo,s work from that of her cognitive anthropological contemPo-

E ffiF
1,66 Daniel V. Rosenberg lnnguage in the discourse of the emotions 1,67

raries is the status of these other things. The abstracted product, the "Ethnopsychological vocabulary," according to Kirkpatrick and White
"focal terms," achieve the stafus of leitmotifs, or master metaphors, in (1985:16), "is used to talk about personal and social experience, not to
terms of which the author can interpret - this being the recontextuali- identify or classify'objects' separable from their cultural contexts." The
zation - a vast range of beliefs and practices. contributors to their volume (White and Kirkpatrick L985)
Rosaldo makes it clear that headhunting, the life cycle, marriage, and generally agree that the vocabulary in question does not function
so on among the Ilongot can be described in sentences in which 'ligef' primarily to "classify" social reality in terms of taxonomic schemata
figures as master metaphor. As such, the account echoes a very old consisting of categories related through contrast and inclusion.
Westem tradition in which certain congeries of nouns/things, usually Rather, by focusing on frequently used words that point to salient
"paspi8iis" or "sentiments" (but compare also "virtues" and "vices"), cultural notions, such as Bimin-Kuskusmin finiik .. . , Marquesan
are sEEi, to be in collusion or competition with each other, with all of ka'oha .. . , or Baining aknmbain. . . , the ethnographer is able to
these, in furn, in more or less of an equilibrium with "teason." The trq${Jrferential pathways through webs of associative knowledge
Ilongot rendered in terms of 'liget'is a story that is resonant with an old *ilffi [sic] the terms'various meanings and uses' (1985:19)
genre in which all the peculiarities that seemed to characterize the French ""a"dy
could be woven together under the slogan "vanity," and just so the It is not hard to see in this dismissal of an earlier picture the survival
Spaniards under "pide," the Zuni under "Apollonian," ar.d the Kwal+ of many of that view's foundations. The analysis is anchored by a small
iutl under "Dionysian." selection of items from a vocabulary; those items stand for, and in the
The idea that "the emotions" for a given people amount to a glossary ethnographic account just are, locally important notions or concepts; the
of nouns/essences continues this old tradition. Most modern approaches concepts or notions hang together in some structured way; the structure
to "the emotions" in alien settings make this glossary assumptiort; Ros' of concepts is something that natives know and share; as such, that
aldo's is the position most faithful to the tradition, since it acc#nts for structure "underlies" everyday practice, in particular the practice of us-
a vast range of practices in terms of one or two enshrined entries from ing words to talk about things, with meaning-to-use understood as an
such a glossary. The move from enshrined noun/essence to narrative instance of the relation of knowledge-to-action generally.
leitmotif and the further move from motif to culture, fosters a kind of To be sure, the "net of senses" or "web of meanings" discerned by
cultural naturalism in which a small set of Ur-principles lurk behind Rosaldo, in which words (concepts, images, symbols) "suggest," "hint
everything as first causes. 'Liget' and its kind assume the status of nat- at," "eyoke.," are "associated with" or "related to" each other, had for
ural, even animate, forces: ideals toward which any Ilongot behavior can her no such internal location. The "senses" and "meanings" were in-
be seen as striving, principles of which any practice becomes an exPres' tended as just public symbols. Many contributions of the mid-1980s, in
sion. To that extent, the roles that'liget' and its kind play belie the au- which Rosaldo's influence nonetheless is evident, do locate these struc-
thor's many claims against the mentalism and functionalism of the old- tures internally and mentally. They render them in a harder form, in the
style culture and personality: They function as internal(ized), motiva- process very nearly resurrecting (even when disavowing) the earlier pic-
tional forces, encqiled and expressed in the grarunar and idioms of En- ture, as "associative knowledge," "organized knowledge structures,"
glish ascriptive talk. "inferential pathways," and the like. The focus has shifted away from
the mechanics of constructing the arrays, but the elements of the array
(lexemes), its form (quasi-spatial), and its status (cognitive) are the same.
Ethnographic metasemantics Consider, by way of example, ]ohn Kirkpatrick's "Some Marquesan
In the anthropology of "the emotions," our arrival today at "discourse" understandings of Action and Identity" (1985). Kirkpatrick (1985:85) notes
in no small way rests on Rosaldo's articulation of an intellectual Sener- that Marquesan "ethnopsychology . . . is realized in everyday dis-
ation's discontent with its predecessors' understandings of language. By course, not in specialists' accodnts of the world or strongly demarcated
the mid-1980s, most anthropological commentaries on "the emotions" speech situationg." He takes up three "lerms," 'ka'oha', 'haka'ikn', and
echoed Rosaldo's attack on denotation, reference, and classification. 'keitani', which he glosses as 'concern,' 'shame', and 'envy', resPec-
'.

168 Daniel V. Rosenberg Language in the discourse of the emotions 1'69

tively. Each is (in his words) a"term" and, at the same time, a "state" reflected on by both native and ethnographer. Further, the glossing rou-
or "capabihty." tine rests on a transitivity of reference, such that the routine works when
Situations in which the terms are (again in his words) "mentioned" stll +. . . s[n] provide a concrete referent - ideally approximating what
and "used" are identified, and informants' "accounts of " the "terrns"l we might dub an "ostension-cartoon," since the nearer it comes to
"states" are cited. We see, for example, something we could draw a picture of, the better it works - with 'that'
referring as anaphor to the picture, and finally, with the lexical focus as
how the sense of ka'oha as 'gift' and mea ka'oha as'object of con- equivalent (via 'to be') to 'that'. By such procedures, an initial glossary
cern, compassion' can be closely linked to a view of ka'oha as a that maps words to scenes can be uncoveredi as Kirkpatrick says, the
pliJt"ss integrating perception, emotion, and action: "terms ire treated as mapping realities," and the analysis rests on the
+--"1'- native and the ethnographer sharing the assumption that this is how a
poor. I don't come to your place' You recognize [this]; you
come to my place, bringing money [and] say, 'Take. You, your metasg4pntic exercise should proceed.
wife, your children will eat" You iust give, I didn't ask' That's If thffinrographic semantic strategy - words (now "key vernacular
ka'oha." (1985:88; brackets and quoting as in the original) terms") 3s hbels for things - persists, where in the turn to "natural
discourse,, (Kirkpatrick and white 1985:23) does the innovation lie? It
From a glossary uncovered in this way, Kirkpatrick (L985:88) moves out- lies, first, in the nature of the referents. In part because "the emotions"
ward to exemplify how all three "terms"l"states" "involve complex at- famously fail as tangible objects for ostensive reference, "emotion words"
titudes toward life in a social world'" instead are said to refer to or "point to" "emotion concepts," which in
Kirkpatrick takes the "tetms"l"states" to be exemplary (as "tokens") turn are said to encode knowledge about social action' Thus:
ol "types" of "general modes of engagement" or "personal processes" The authors in this volume repeatedly analyze emotion words ''
-Mirquesan '
u-or,g Marquesans. The model of etttnopsycholo$ is an for what they say, implicitly or explicitly, about interactive situa-
"array" of four "process"-t5lpes/ rePresenting all four permutations tions. . . . we view emotion words . . as guideposts to cultural
(+l+, +10, Ol+, and 0/0) of the distinctive features "boundary" and knowledge about social and affective experience. . . . '[L]oneliness',
"self-concept." The model "maps the assumptions Marquesans draw on aiai,io, and awumbuk differ largely in the ways the situation of an
to define'concern', 'lust', and the like" (1985:98). Kirkpatrick shows that isolated person is understood to be problematic in American, Mar-
the four "process"-types, and their key (as well as lesser) lexical-cate- quesan, and Baining culture'a
gorical "tokens," can serve to describe Marquesan expectations about
events and predicaments in everyday life, especially as these expecta- The "key [emotion] terms" serve as labels for chunks of native theory -
,,webs of associative knowledge" about the individual as social actor,
tions change through the life cycle. -
Note that the analysis depends on the ethnographer and the native as evaluator of the moral propriety, and so on of surrounding events.
focusing their attention on the same segments of speech. The infor- They help the native describe social interaction; they provide the eth-
mant's expressioq6translated as "That's ka'oha" requires that Marque- nographe, with a name for part of the native theory about social inter-
san speakers be able to provide the equivalent of the English glossing action. As "terms" they name, and as "concepts" they are, "prototypi-
routine cal event scenarios," "schema-" or "script-"like microsocial guidelines'
Such a word-to-concept-to-scene mapping, however, is just reference
SIU+. . . S[n]+ of a different kind, with such "prototypical scenarios" replacing plants
S[n+ 1] (Demonstrative* [to be] +NP) and animals. But in the rendering of such maps of models, we find as
in which the Demonstrative in s[n+ L] is a discourse anaphor for all of well a commentary about how the "key terms" may be "used'" The
Stll + . . . S[n], the verb 'to be' in S[n + L] has a metasemantic sense, and miCrOSOCial SCenario turns out'to be, for a "key tetr\," both its referent
the NP in S[n+1.] is what is being glossed. The formula must devolve and the arena for its appropriate instancing. so, the second innovation
on a salient linguistic focus, typically lexical and nominal, that can be seems to lie in An aecount of usage; an understanding of "natural dis-

h .=gE
i

L7L
170 Daniel V. Rosenberg Language in the discourse of the emotions

course" takes the form of a menu that tells you when a term may be recontextualization reduces situated discourse to discourse about situa-
pulled out of its glossary and deployed. For "key emotion words," this tions, and discourse about situations to word reference only. Interested
kind of account aims to capture the native calculus about the conditions human interaction appears as one of the mapped realities. Instead of
that must be met before a certain interpersonal state of affairs - an insult exemplifying contextualized languaSe use, investigators take word ref-
or a death, for example - warrants *le "use" of certain words. The analysis erence to be what language can be used for'
then shows how a plurality of these calculi is a local model of and for In recontextualizing language, ethnographers attribute agency to words.
interpersonal life. words ,,say,, things: Recall Kirkpatrick and white's observation that
Certainly, we find in this view a richer account of reference than the emotion words are analyze d "fot what they say, implicitly or explicitly,
onepi&ided by the earlier ethnographic semantics. Because the condi- about interactive situations."s The ethnographer gives the native eth-
Uontlor the use of "emotion words" engage the most intricate local nopsychology the shape of one of his or her own metapragmatic meta-
presuppositions about the person as social (e.g., moral, political) actor, phtrsr.,*r eihnopsychology is a list of words, each of which Possesses
those words may be the best, or at least the most parsimonious, meta- u., ug"ffi" po*L, to refer, to "tell a story'" The theory is packaged in
phors of native provenance for any "person-centered ethnography," and *o.Jr, fia tn" words describe social life'
perhaps for ethnography generally. Alongside the many expressions in our current literature in which the
Yet, for all the benefits that accrue to our specifying the conditions for key terms are the subjects for verbs of speaking such as 'to tell', 'to say',
the use of "key [ascriptive] terms," our doing so does not amount to a ,rd ,to talk (about)' are expressions in which the terms/concepts take
proper account of "natural discourse." For all the counsel that we attend English psychological predicates. Kirkpatrick's claims that'haka'ikn' "im-
to "everyday talk," the anthropolory of "the emotions" in the mid-1980s p"ir" ur,d-"inhibits" 11eas:sl;, and that 'ka'oha'can be "experienced"
and,,enacted,, (1985:89), are among many examples.6 On occasion, local
l;:.'T[',:l#l'#::,"ffl :1,j:";':]?i,ff ff ::"*ilffi:1lffi,'JJi; notions about motivation are appealed to, but this appeal seldom takes
the speech situation. They are, as many things informants say are, me- the most reasonable tack, which would be to examine the verbal predi-
tasemantic acts of revelation; they reveal a "concept" or "category," arrd cdtes that the nouns take in the original language, among other facts
they exernplify the correct usage of the word that names that "concept" about their grammatical contexts. Most often these ethnographic predi-
or "category," thereby telling us something about the native theory of cations slide ambiguously between that sense in which they would fig-
the word's contextual appropriateness in instances of reference. Such a ure in any anthropological account about the enactment of an "undet-
metasemantics of appropriate lexical usage contains a commentary about lying model,, and that sense in which emotion "states" would manifest
everyday life of a sort that yesterday's ethnographic semantics did not tii"rir"lrr", in a local ethnopsychology about the enactment of internal
provide. It appears to be an account of "natural discourse" because it states.
tells us how words refer to everyday life and tells us when they might Just as our historical practice of enshrining
noun families is well estab-
be useful in that life. As such, it is a familiar species of metapragmatics: lished in the domain of "the emotions," thus making our metalinguistic
a reflection on wt6t you can do with words. habits especially tenacious there, so there is another fact about "the
Once again, a metasemantics of this sort is not a contextualization but emotionsT' chez nous that infiltrates our accounts. When we ask that the
a recontextualization of the linguistic datum. From the relatively safe metasemantic glossary not only refer to the social world but also em-
terrain on which native and ethnographer cooperate in a lexical metase- body native cognitionabout it, we risk smuggling into the account what
mantics, the ethnographer imposes on the glossary so obtained the de- we elsewhere assert is peculiar to our own local ethnopsychology.
mand that it serve at once as a description of situated discourse and as In the current picture, the key words have unique conceptual quali-
the native theory about persons in social interaction. The first demand ties, and these in turn constitute, or at least populate, the native ethno-
is met when words that refer to scenes are shown to be instanced in psychologicaltheory.Sotransfigured,theybecomethecontentsofan
discourse. The second is met when a collocation of the conceptual es- ;,underlylng model." Our ethnopsychology intrudes when the under-
sences of a family of terms is said to be a model of social life. This dual lying cognidve 8tatu8 of the ethnopsychology-as-model - recall Kirk-

,#=
;

172 Daniel V. Rosenberg Language in the discourse of the emotions 173

patrick's claim that "the model maps the assumptions Marquesans draw sus "context-free reference" and "action refetence" versus "abstract trait
on to define 'concern', 'lust', and the like" - is confounded with our reference," with the latter glossed also as "does" versus "is" "refer-
own ethnopsychological understanding of "the emotions" as underly- ence." The result about differences in "modes of thinking" rests on
ing forces. The special risk of an "underlying model" account of "the statistical differences between Oriyas and Americans alons these two
emotions" is that we already conceive of "the emotions" as "underly- dimensions. The dimensions, in turn, represent aggregate sums of judges'
ing" out lives; we have trouble sustaining the distinction between the more specific ratings of fragments of informants' responses'
model's governance of life, as a scientific abstraction, and "the emo- shweder and Bourne's coding scheme captures a number of overt lin-
tions' "-g.overnance, which as a local theory is ours and may or may not guistic features of their informants' resPonses, but these coded features
be thqir.H do not stand convincingly for nontrivial kinds of cultural differences in
ascriptive practices. The coding categories capture the presence versus
absenqeJ[f some simple English grammatical categories, but these are at
Grammar and indexical pragmatics, minimally best obhffi indices of the cultural construction of person concepts. To
see ho*this might be the case, we need to retrace the iourney from
To the question that is the title of their 1982 essay, "Does the Concept
linguistic data to cultural account'
of the Person Vary Cross-Culturally?," R. A. Shweder and E. J. Bourne
Oriyas are said to have a "context-dependent person concept"; their
answer, emphatically, "yes." Their concern is "with other people's con-
"attention is directed towards the behavioral context in which particular
ceptions of the person and ideas about the self." They aim to "interpret
behavioral instances occurred." Oriya "context dependence" iS a glosS
an alien mode of social thought," bearing the global description "con-
on the statistical preponderance among Oriyas' resPonses of five kinds
crete, undifferentiated, context-specific, or occasion-bound thinking,"
of ,,contexfual qualifications." The units of informants' resPonses coded
and {ry
as either "qualified" or not were "clauses," each of which "contained
culminating in the view that specific situations determine the moral no more than one subject-predicate-object sequence" (1982:112)' When
character of a particular action, that the individual person per se is we examine "clauses" coded as "qualified,"7 we can see that "context
neither an object of importance nor inherently worthy of respect, dependence" is a conclusion based on the Presence of what are stan-
that the individual as moral agent ought not be distinguished from aaiaty called (a) prepositional phrases or (b) time and place adverbials.
the social status s(he) occupies; a view that, indeed, the individual By contrast, the "context-independence" of Americans rests on the pau-
as an abstract ethical and normatiae cate1ory is not to be acknowl- .ity of these constituents. So, among these examples that the authors
edged. (L982:97; emphases in the original) prbvide,
Although the authors' focus is "the concept of the context-dependent (a(i)) 'He is honest with others'
person," I take the issues they raise about ascriptions of personalityand (b(i)) 'Last year he did favors frequently'
character broadly speaking to be essentially the same as the ones we (c(i)) 'She is stubborn'
might raise about dfect more narrowly. (d(i)) 'He is verballY abusive'
The authors conducted interviews with seventeen American and sev- the first two are "context dependent" and the latter two are "context
enfy Indian (Orissan) informants. The American informants came from independent."
three groups, in each of which informants were acquainted with one Consider these sentences semantically, leaving (b(i)) to the side for
anotheri each informant was asked, with reference to another member now. As speakers of English, we will intuit that
of the group, "How would you characterize so-and-so's personahty?"
(a(ii)) 'He is honest'
Indian informants described "up to three friends, neighbors, or work'
(c(ii)) 'She is stubborn with others'
mates" in response to the request translated as "Tell me in depth about
(d(ii)) 'He ie verbally abusive with others'
so-and-so's character, nature, and behavior."
The responses of Shweder and Bourne'e informants to these queries hardly dlffer from the members of the first set. The complement phrases
were coded along two dimensione: "context-dependent reference" ver- that the conttluctlona ln (a), (c), and (d), namely

.5G:
175
174 Daniel V. Rosenberg Language in the discourse of the emotions
overt
(e) (Personal Pronoun) + (to be [present]) + for certain ascriptive adiectives, 'with others' just makes lexically
(Adjective of ascription) . . . what is entailed, semantically, by the adjective'
about
When considering (a(i)) igainst (a(ii)), we might ask questions
allow are all of the form will proba-
the discourse contexts itt *t i"f, they occur. In doing so, we
(f) . . .* (Preposition) + (Obiect NP) bly want to consider how such contrast pairs index the greater versus
le,se, authority of the speaker to ascribe character to the third party,
where all such Object NPs must be animate and sentient,s that is, mini-
relevant evidence im-
mally capable of perceiving the ascribed quality. '[W]ith others' makes and in particuia, how the speaker's access to the
time how im-
overt tfre unmarked case of such Object NPs, as animate, sentient, and parts that authority. To see how messy, and at the same
he did
portant, these questions are, let us return to (b(i))' 'Last year
&rw
humwp'rh distinction to the marked
(g) 'He is verbally abusive towards his dog' favors
- frequentlY'.
wirtu; ,p"ut"r, of English need to know in order to know whether
and, additionally, makes overt the unmarked case of Subiect NPs, as (b(i)) is\Waontextually qrulifi"d' athibution? ln particular do they
animate, intentional (or at least responsible), and human, ruling out the ,.1".* in orderio judge it "mole qualified" than either of the
marked """i
following?
(h(i)) 'He's ('She's'?) stubborn' (b(ii)) 'He does favors'
where 'he' is anaphoric for a lawn mower that will not start, and the (b(iii)) 'He did favors'e
like, since Minimally,theyneedtoknowthreethings:(1)thekindofstressthe
year as a unit
(h(ii)) 'He's stubbom with others' sentenceieceived when uttered; (2) the ratio between last
can Pre-
of time and the period of time that the speaker and addressee
cannot be used with inanimate or nonresponsible antecedent su$ects. to evidence about the third party's behav-
suppose the speaker's access
At least in this sample of sentences, then, the distributional facts, the
facts about the contribution of structure to meaning, concerning 'with
ior;and(3)thelengthoftimefromtheendoflastyeartothemoment
of speaking.lo
others' tell us so far that its overt instancing picks out referentially the
unmarked case of human Subject and Object for sentences in which cer-
Llt us consider a simplified example,
tain adjectives of ascription appear. We ought to expect a similar Pattern (b(iv)) 'Last Year he did favors'
in the
for constructions that contain other ascriptive adjectives that require the which by all accounts is also codable as "contextualty qualified"
referential disambiguation of human versus nonhuman subjects and ob- shwedei and Bourne study. The adverbial last year', in (b(iv)) is un-
is indexed by the selection
jects. stressed when it merely contributes to what
of the
'[To be] 'honest' [with]', of course, seems to require human subject' of the past tense ol'to do" namely, the location of the referent
can imagine
In addition, for most speakers, who find verb phrase in time vis-d-vis the moment of speaking' We
just asks when the favors were done'
(a(iii)) (?) 'H€h honest with his dog' a context in which an interlocutor
sense' as (a kind of
But'to do' in (b(iv)) can have a relatively aspectual
unacceptable, it requires human object as well. In this case, the refer- insinuation of) nondurative and completive aspect' io ootlT contexts'
ential contribution of 'with others' rests only on its contrast with such
suchasinresponseto,Hasheeaerdonefavors('..inhislife)?'Presum-
constructions as
ably this sense will be signaled in the speaker's resPonse
by stress on
placed' the speaker indexes
(a(iv)) 'He's honest with me' the adverbial 'last year'. With stress so
evidence for
for just those discourse contexts in which the speaker has already estab- something about thl ,"op" of his or her claims to having
same time
lished a contrast between 'others' and 'me'. In the absence of such a the predilation. By delimiting the evidentiary claims at the
the
context, 'with others' once again merely makes overt the unmarked case that he or the aaserts that the third party does indeed do favors,
implicit'He has never done favors'
of human complement obiects. On such grounds, we can conclude that speater falrlfleo the interlocutor's

h
1l :

L76 Daniel V. Rosenberg Language in the discourse of the emotions 177

while at the same time cooperates with the addressee by suggesting that tation of the verb phrase; one of these cases we could gloss as "contex-
the speaker has not witnessed the third party doing many favors. In tually nonqualified" and the other as "contextually qualified." These
other words, the speaker licenses the addressee to take the speaker's also happen to be, with good reason, the cases in which adverbial stress
delimited exposure to the relevant evidence as evidence for a delimited does not matter. In the vast terrain between the two extremes, the ad-
ascription - the one the addressee seems to want - without the speak- dressee must locate the expression in some cell of a matrix that maps (1)
er's having to make such a predication referentially. The adverbial stress stress, against the ratio in (2), against the length of time in (3). The ad-
forces a pragmatic pun that rests on the sustained ambiguity between dressee does so just because the utterance in every one of these cells
an (inde$cal) evidentiary delimitation (of how much the speaker claims sustains an ambigUity between the evidentia4z and the referential senses
to knoly)rffid a referential one (of how many favors there have been and of the ascription in its verb phrase.
of how"hurative, or completive, or iterative is the third party's tendency A proper analysis of the relations among tense, asPect, and explicit
to perform them). Discursively, the former delimitation insinuates the adverbdl phrases, as grammatical phenomena the distribution of which
latter. both cffiibutes to reference and indexes a variety of speech situations,
Fortunately, there are facts about the speech situation that the addres- lies bey6"nd the purposes of this chapter.l2 To simplify, let us retum to
see can invoke in disambiguating the stressed-adverbial pun form of the particular context defined by the queries Shweder and Bourne put
(b(iv)). These are the facts I noted earlier in (2) and (3). Where the ratio to their informants. There is every reason to think that informants' re-
in (2) approaches 1 (the speaker and addressee PresuPPose that the sponses are simultaneously evidentiary and more strictly referential, with
speaker knew the third party for one year, and that year was last year), certain important constructions - indeed, many of the ones coded as
and where the conversation takes place in January (or in the early sum- "qualifications" - seeming to foreground the former function. In partic-
mer, in U.S. academic contexts), there the adverbial imparts a maximally ular (and leaving translation issues to the side for the moment), the re-
durative and noncompletive sense to the verb phrase. In this cas?i, the quest to the Oriyas, "Tell me in depth about so-and-so's character, na-
pun can be disambiguated confidently in the direction of what, if we ture, and behavior," would seem to enjoin a maximal expansion for
have to, we can call a "contextually nonqualified" attribution or refer- evidentiary purposes of every kind of predication, from complements
ence. There is no reason for the addressee to interpret the speaker as that merely spell out what is referentially redundant (' . . . honest with
wanting to delimit the scope of his or her claims to the evidence; equiv- others'), to complements that are simply required by the verb (consider
alently, we can say that the indexical contribution of adverbial stress is 'he makes fun . . . ', which is awkward in English without some explic-
minimal in this case. itly named object), to adverbials of time and place.
At another extreme, in cases where the ratio in (2) approaches zero By giving 'with others' and other prepositional phrases and 'fre-
(participants presuppose the speaker's many years of knowledge about quently' and other adverbials the notation "contextual qualification,"
the third parry) and where the speech situation occtrs in December (APril, and then taking the latter to be "context dependence," Shweder and
say, in the academy), the addressee's appeal to such facts yields a dif- Bourne rename a grammatical phenomenon a cognitive one, and this in
ferent kind of disansiguation. Here the addressee will know that a cer- turn an anthropological one. The renaming results on our metaPrag-
tain kind of speech event has taken place, one bearing the folk label matic intuition that any overt expansion of a minimal subject-verb-obiect
'damning with faint praise'. In this case as well, stress on the time ad- string must be in the service of "better" or "more precise" reference, of
verbial cannot be taken to index an evidentiary delimitation. It has in- a sharper picking out of objects in the world. We intuit this even when,
stead the curious indexical function of signaling to addressee that the by virtue of our lexical (in relations of entailment) and grammatical (in
action referred to in the verb phrase is, in fact, the opposite of what it the governance of complement phrases) semantics, we can be said al-
appears to be; that it is nondurative, noniterative, or completive - in ready safely to "know" much of what our overt expansion will seem to
short, rare.ll specify.
The two extreme cases I have pointed to seem to be the only safe If the Shweder and Bourne coding scheme cannot distinguish a verb
grounds on which the addressee can manage a Purely referential delimi' phrase expanded for pure referential delimitation from one expanded

:4*t==
..

L78 Daniel V. Rosenberg Language in the discourse of the emotions 179

as II, and IV has just one ascriptive clause, which must be "nonquali-
just because the grammar governs its expansion, from one expanded to
index informants' evidentiary authority, from one expanded for other hed."
indexical reasons - having to do, for example, with the etiquette of dis- Yet, our native speaker intuition that 'A' has predicated essentially
closure (it being in certain contexts ProPer and in others improper to let the same thing(s) of 'him' in all of I-IV is sound. with a view to the
on that you have seen the evidence) - and cannot unpackage analytically kinds of conversations people actually have when they ascribe things to
the standard case in which these functions coexist, then the ascriptive persons, our analysis ought to be able to capture all of these cross-clause,
utterances they have collected cannot supPort the set of condusions about cross-sentence, and cross-turn-at-speech continuities. These will ex-
"world view" and "mode of thought" that the authors ProPose. In par- ploit the structural facts of anaphora and ellipsis (among others); these,
ticuhp h-cause overtly minimal ascriptive phrases can be expanded on ir, tr*, will interact with such features of conversational pragmatics as
gru.r,rfri-ti"ul and/or indexical grounds, and because such expansions are the structure of query-and-response routines, especially as these engage
in many cases less referentially delimiting than unexpanded phrases, local eXnectations about the proper relations between abstraction and
the statistical prevalence of these expansions among Oriyas cannot set evidenfurd these latter relations will have a specific shape for specific
us on a path that leads to "occasion-bound thinking" or to the "ethical kinds oihascription, against the background of what the speaker and
and normative" irrelevance of the "abstract" "individual." addressee presuppose of each other's knowledge of the object of ascrip-
tion.
None of this can be discerned in the Shweder and Bourne abstraction
Conversation, minimally of ,,descriptive phrases." Note that IV must, by any account, be seen as
the maximally referentially delimiting ascription. Supposing such a con-
In the foregoing, I have looked at Shweder and Bourne's analysis of versation to te coded in the current study, you would not know this
what they call "descriptive phrases." Now I argue that this absQpction
fact. Its first turn would be analyzed as "nonqualihed," and its last two
from language is as radical as the ethnographic semantic abstraction of
furns could not be analYzed.l3
lexemes and the ethnographic metasemantic abstraction of "key terms."
The shweder and Bourne study not only abstracts from connected
If the semantic and referential (and then cultural) consequences of the
discourse in its analysis but also forecloses it in its initial task design.
surface eipansion of clause constituents may be trivial, so too can the
American informants provided twenty written "descriptive phrases or
fact of a clause's nonexpansion (what we might call the "nonqualified"
sentences"; Oriyas responded orally "in as many or as few ways as they
American model, in the context of this study) be trivial. The reason is
chose.,,14 So, a two-sentence string such as III is possible for Oriyas
as
that referential delimitations can be made across clause boundaries, and an answer in connected oral discourse to "Tell me in depth about . . ,"
as well across sentence and turn-at-speech boundaries. The subtleties of sentence
but would seem not to be possible for Americans: The second
distinguishing referential from evidentiary and other indexical delimi- is a poor candidate for a list of twenty distinct written answers to "How
tations remain. Consider:
*ortd you characterize so-and-so's personality?" Il in utterances n+ . . .
I. A: 'He is yery loving toward his children" in natural discourse the speaker will adduce "concrete" evidence for an
II. A: 'He isVery loving, the kind of guy who's always showing "abstraction" i\ n (or if ihe speaker concludes "abstractly" in n what
affection, you know, like for his children.' was prediiated "concretely" in m, m+ . ' '), the current study would
III. A: 'He is very loving. You should see the things he does for seem to be able to capture such sequences for Oriyas but not for Amer-
his children.' icans.
IV. A: lHe is very loving.'
B: 'I've never felt that way about him.'
Semantice, reference, pragmatics, and ideology
A: 'Have you ever seen him with his children on weekends?'
picture
For Shweder and Boume, I. must be a "qualified" clause, II is two clauses, By 1980, it eeemed that if an abstract, ahistorical, and desiccated
oi euch obfaete li "culture," "knowledge," and "text," along with an
one "qualified" and the other "nonqualified," III Sets the same analysis

,-==
;

180 Daniel V. Rosenberg Language in the discourse of the emotions 181

understanding of such objects as "persory" "sell," "at)thor," and "readel' own grammar, and that grammar encodes an ethnopsychology in some
as nonagentive, disinterested, and unwitting, had been founded on a measure. Second, a pervasive metalinguistic metaphor of our own as-
certain semiotic of "reference," "signirticafion," and "dassification," then criptive practices - indeed, of our own culfure of "concepts," "emo-
it stood to reason that a human science of the conctete, processual, and tional" and not - maps unique nouns to unique mental or underlying
interested would require a new semiotic. In "discourse" we found both quiddities; much of our own ethnopsychology emerges in reflection on
a reconfigured object, a real, palpable manifestation of "culture," and the referential power and appropriateness of isolable foci.
so on, and a new semiotic, "discourse" seeming in the first instance to It was Whorf who proposed that facts of these two kinds are analyti-
name some kind of linguistic object' cally distinct but not independent. In grammar, analogy of distribution
Yet, the h&vantages for the human sciences that might lie in the ap- entails analogy of meaning; t}:re semantics of any diskibutional class, such
parent ,&rt of "discourse" as situated activity to "discourse" as as a class of nouns, emerges in the scope of its possible locations in a
"tt
(semiotic foundation for) analytic abstraction have rarely been reatzed. grammar,gf sentences. Equivalently, any class is a class only by virtue
For outside of linguistics ProPer, "discourse" has scarcely been con- of its eqrl$Sency of distribution. By contrast, an ideology of reference is
ceived of as a semiotic datum at all. It has served as a metalinguistic an obliqudfeflection on the semantics of apparent classes of salient seg-
metaphor whose use indexes its author's cominitment to the idea that ments of speech, understood not as it emerges in grammar but as it
what is cultural is not timeless and disembodied artifact, but instead is might denote objects in the world.
like something you might say to me, on a certain day for a certain rea- For the "emotions" or any other proposed class, then, we need to see
son. But when everything cultural is like everyday speech, then every- that it is only in an ideology of reference that "key terms" will select
day speech itself is a datum of uncertain status. unique quiddities or concepts. A mapping of this kind will be a true
In this chapter, I have argued that in the first generation of anthro- ethnopsychological datum for those cultures, such as the culfure of
pological approaches to ascriptive discourse, a passage was made ft8rn speakers of American English, in which it is an attested practice. Where
abstracted parts of speech to ideation and culture at large that bypassed ethnopsychological utterances take the form of metalinguistic ideologies
the scene that is their true intersection - everyday discourse - and that of referential function - everywhere, according to the Whorfian hypoth-
this scene ral,as bypassed even by investigators who named it as their esis - we need to examine them in their local forms.
place of work. I have argued that the ordinary context of everyday life It is not possible that the "key" segments on which natives reflect will
emerged not in contextualized speech but in two kinds of recontextual- everywhere be nouns. Nor is it possible that, to the extent that the gram-
ization: one in which abstracted nouns were glossed in such a way as to matical segmentations are similar, the content of native reflection on
establish that they referred to, and hence could be instanced in, every- those segments will be similar. This is so in part because the glosses of
day life (Rosaldo, Kirkpatrick); and another in which the statistical dis- those segments will be unique, as the work of the past decade has estab-
tribution of abstracted phenomena of surface grammar was said to es- lished, but more centrally because the grammar that surrounds those
tablish the distribution across cultures of theories of the Person as more segments is also unique, and it is in that grammar that meaning (equiv-
versus less embeddedf,l everyday predicament, with these theories cor- alently, the possibility of glossing) is to be found. This semantic level of
responding to the liteial propositional content of the abstracted phe- analysis must be seen as distinct from local (and transcultural) practices
nomena (shweder and Bourne). I have said that such recontextualiza- of lexical glossing; it is foundational of the latter, meaningful in its own
tions are metasemantic and metapragmatic exercises, with a long history right, and only obliquely accessible to native reflection.
not least in the ascriptive domain, and against whose potential for fos- As Silverstein (1984) has established, the (Whorfian) analogies or (Ja-
tering rich ethnographic description we must counterpose the risks of kobsonian) equivalencies that inhere in a grammar emerge in the real
infiltration by the ethnographer's ethnopsychology. time of conversation in the maintenance of topic-comment cohesion -
The risks arise for two reasons. First, when we discover only isolable, of, reference we can say. One of the curiosities of the discourse function-

nominal foci in our ethnographic metasemantic practices, we commit alism of the last decade is that it has seen itself as finding functions that
ourselves in our ethnographic writing to embedding these foci in our would replace or overrule "reference," an impossible task should every-
lnnguage in the discourse of the emotions
183
182 Daniel V. Rosenberg
day speech be our object. But instead of recovering "reference" in ide- hitherto had been; and (3) take the local articulation of rationales for
we will
otoges of lexical appropriateness, we need to look more closely at actual speaking ascriptively as a datum in its own right' For this last'
conversations, from which we will be able to uncover foci of native in- need tolkeep in mind that the native metapragmatics of nonreferential
terest in the parallelisms of referential cohesion. functions is as expandable as ours is, and hence will not map uniquely
why do this? The contributors to this volume have discovered, in the onto "discourses" of unique kinds'
turn to "discourse," that persons in other cultural settings provide mul-
tiplex homes for "the emotions" and many neighbors - such as somat-
ici, characterology, behavioral predilection, belief, cognition, and rights
of neigh- Notes
- and^tftir investigations have uncovered important forms o{
borly &gagement. But as with other anthropological domains - con- 1. White (1978) and Lutz (1982) are two efforts to apprehend t-T;l:1I8y?Ce
ioots in ethnographic semantiis, although the object of
of expand- ascriqtion that have
sidei "kinship" - so with "the emotions" has the excitement ili;ffiJ; .riti.irr"-ir the wider eiirnbgraphic semantics-<ognitive anthro-
ing the domain (into "morality," "politics," "the bodyi'and elsewhere)
beln accompanied by debates and wories over its continuing integrity. z. Ei',tffik"ijti'if'," cognitive models. about ordinary proc-ess: Rosardo's
"emotional t"t*r"1."-fi'ttended as public symbols'-by appeal to
(Rylean-
Moreover, the tension between a focal and ah expanding domain is mir- of the public.. Aside from
cum_) Geertziun and wittgensteinian deJenles
rored in a tension between "the emotions" as anthropological construct coenitive anthropol,ogy, an"other object of Rosaldo's criticisms is structural-
and "the 'emotions'" as Euro-American quotative, again a cus- mental-objects from
;',fi;';;;;;i;;';;', lif!i;. and too oiderlv abstraction ofstructuralism must be
history-of Sut'ssut"un
tomary anthropological tension, and an issue on which I have taken a p"Ufi. fifu. Although the entire
mentalism,
understood as an eifort to isolate u p,r[rI. semiotic from a private
side typographically in this chapter. one way to resolve these tensions the coincidence in the 1.960s and 1970s of American anthropology's
cognitive
is to examine how the requirements of referential maintenance in con- trr" u ae.ideJ."u"tutist turn in Levi-strauss's structuralism, coupled
with ""a
tree.s
versation foster, but at the same time for any language delimit, thd'bcope the iconic likenesses on paper between cognitive ltt\ttqgt^"^g/s
made-it possible by 1980 to attack
urra tf," hke and Levi-strauss'i siructures,
of the neighboring semantic terrain into which a speaker may turn the trr" ,"pp.r"a *""iurir*r, with attendant methodologies, of both schools
conversation. For an introduced topic, the semantics of grammar not only with tlie same stroke of the Pen'
allows but circumscribes and even unifies what counts as coherent com- 3. Aii;i th" quoted terms in' the foregoing paragraph come from Rosaldo
ment. Although such language- and culture-specific considerations may (1980:47, 49; 56-7, 161, 188,217').
4. A[ th."u citations are from Kirkpatrick andand White (1985:!!,
not everlrr.rzhere yield "emotions," they should yield true local domains.
-. ..-
(1985:20): "In sum, the
5. see also N. Quinn, quoted in Kirkpatrick aboutwhite
At the same time, the turn to "discourse" and especially to instead- word,commitment'iells a complei story American marriage."
6. "Impels" and like i".-t ut" prehicates rypicallyihatassigned to "th.e emotions"'
of-,,denotational" functions of speech risks an unprincipled expansion others may disapproveof
of ,,the emotions,, (and any other) domain by the lights of the ethnog-
U"i'[" a r".ogttitiott
"fr.;Fi;kJili";olies
. . .;;;;ut" o. ,.ti"ity'; 1loc. cit';, i'hith usto^es that the term must be
rapher's so-called speech-act theoretical classification by intents and not onlv "state" but also conceptual microcosm'
giy91,
prrpot"t. There can be, I submit, no tactical use of the "speech-act" sort Z. eiif.*in ,"""rriotf,". kinds of coders' "qualification..categorigs-*9
as defined on p' 116, which
tt not included in the aggregate category,
ui'"
furagir,g, negotiatiJg8, complimenting, ostracizing, sympathizin$,
ex- iiefe.ence"io a" specifiiinhividual, often denoted by a
"r"
irr"frra"" these' (ai
cusir,g, issuaging, fleading, flirting, self-presenting, and so on) that the noun 1e.g.,
,he gbts anerv with his father')"; (b) "refer-
proper or common
passions will not be found to subserve, anyr,rrhere. To locate apragmatics ence to a speclnc ;;;'i;lh"., tE.g.,.,t""',makes tun of his family')";.(c)
of ascriptive discourse, the most we can do is (1) investigate indexicals "reference to p"opi"-oiothers in gene-ral (e'gr' 'he is honest with others')";
(;t';i;;p".rl'lqultficationl: stat6ment of vihen or how frequently the.at-
of a sort that we define typologically as "emotive";rs (2) examine how tfute ol..rrc
;i".iy"ar he did favors freq-uen:!y'\':iand (e) "[qualifi-
G'.g-,
the conversational deployment of shifting topic-comment foci not only cation bvl locale: itatement of where or in what location the attribute
occurs

maintains reference but also shifts the evidentiary requirements and ex- (e.g., 'At-school he puts on a front')" (1982:113)'
Amonq Orivas, cliuses rated as ''contextually qualified" outnumber those
pectations of ascriptions, at the same time turning the conversation in a ,"iJ';"3 quiification,, by Z.S3 to 1 (given as 3 fo 1 in the text); Americans
direction that is locally construed as more versus less polite/rude, know- displav a 1 to I ratio on this measure'
8. ili[o,igh-1;;5;ui;i1ou;u"t NP)], e'g', 'He is honest about financial mat-
able/ineffable, sufficient/insufficient, and so on, compared with where it

€Ee,
Language in the discourse of the emotions 185
LU Daniel V. Rosenberg
the emotion,, (1982:34), and offers an interesting analysis in the_sapir-P_ra-
ters', and relative clauses, e.g., 'He is honest where money is concemed"
gue tradition of the latter as "modes of affective e-xpression" in.Wolof. But
present excePtions. f,rere is no reason to presume that the "emotion[s]" on both sides of this
by con-
9. bhweder und Bor.n"'s (b(i)) appears to be "contextually qrralified" distinction are conspecific beasts, and that expressive,indedcals are to the
the other of which is the "subject-
t urt *itn either (b(ii)) ,]. tiitif)i, one or
of lexicon and native rbflections thereon as one "channel" or "means" to an-
r:redicate-obi"ct" stfig .odlri'prest'-ably have access t-o' as speakers
;,"o"q,iunfi"d,, other of "expressing" the same underlying qua-lity (Irvine -1982:32; al-
il;;h.;; t'h; ieferential hinimum that lies behind (b(i)).
though Irvin6 prediits that the two levels of analysis are only obliquely
o.ittt" irtft of tf," ui"*ples provided, it seems that either would do; no
relatid, I argue'that we should start with the working assumption that they
is given i" *f1i.fi t"e.l specification of tense counts as "contextual
"*u*pt" are not relaied at all, executing the two levels of analysis at once only for
qualification. , , ,. ---,-^^ things
, too what turn out to be duplex indeical-cum-semantic phenomena)'
10. A fourth fact, the ratio of (2) to (3), is also relevant but makes
i(cmplicated for our purposes here'
.
ends;
11*"--ft die greatest u*tr"ir",la"erbial stress becomes redundant for these
i
i consider
,r. p"."-u". of his first term in office he frequently did good References
work', where stress on the time adverbials (along-with *lttry' arched eye-
i
HffiSF" Dell H. L962. Linguistic Aspects of Cross-Cultural Personality Study. In
brows,etc.)canonlybeinterpretedasindexingthespeaker,slowestima- 's kaphn, ed.., Stuffig Persorutity Cross-Culturally. Evanston, IL: Row, Pe-
tion of the addressee's pragmatic comPetence' terson, pp. 31!59.
12. By now, linguistically a'sttrie readers *rill hurre recognized that.my -discus- Irvine, Judirir'T. 1982. Language and Affect: some Cross-Culturallssues. In H.
ri'." fruJU"".r a native grammarian's ad hoc intentionalist car-
"i"-rripf"r and Byines, ed., Contempirary Perceptions of Innguage: Intudisliplinary Dimen-
icature of what *", rt.i.tfy speakin[, structural facts about derivation ton, DC: Georgetown
Washinston,
si-ons. Washington,
sibzs. umversrty rress/
G-eorqetown University pp' Jr-
Press, pp- 31-47.
distribution. Some Mirquesan
Marquesan Understandings
UnderstanctmSs of
oJ Action
ACUon ano loen-
and Iden-
Kirkpatrick, John. 1985.
13. I ieave aside here the classic statistician's problem about the independence I. G.
b. tU. SeIf, and Lxqen
Kiikpatrick, eds., Person, 5eq,
White and J. Kirkpatrick,
one, tity. In
htv. M. Wnitu Experience:
oi.ur"r, the shoals on which so many attempts, including the current summed Eiploring Pacific Ethnopsychologies.-Berkeley: university of California Press,
io aiu*'pry.t ofo6cal" sociological, ahd cultural conclusions from
op.
pp. 80-120.
o..,rr."rri.d, of abitracied lang[age segments founder' Grammar' a1d any- firtiitrict, fohn, and Geoffrey M. White. 1985. Exploring Ethnopsychologies.
ilG; ;uld want to call;disiourJe," conspire, if they do,grything' to
in C. tU. Wfrite and ]. Kirkiatrick, eds., Person, Self , and Experience: Exp^lorin?
*ut"" ,p"".r, segments nonindependent cases. Readers probably willversus have
Pacific Ethnopsychologies. B6rkeley: U:riversity of California Press, pp. 3-32'
alreidv that what is'wrong with the analysiq of'-say' I'
Emotion'words on Ifaluk. American Eth-
citrrerine. 1t"ssz. fhe Domain of Emotion'Words
Lutz,. Catherine. Etl
iII ir";;T;Jt tnu't ru is less "quatifie-d,, than I., but also that it sums to
"..oented
nologist 9:113-28.
twice as many cases as I. Lutz, Ca"therine, and Geoffrey M. white. 1986. The Anthropology of Emotions.
14. di;;;;;B;;"; forthcoming about the nontandem design oJ their
Annual Reztieut of Anthropology 1'5:405-36-
IndianandAmericanstudies,althou"ghlthinktheyunderestimatetheim- Rosaldo, Michelle Z. lg8}. knoiiedge and Passion: llongot Notions
of Selt' and Social
of the differences between tfie two. They deal fully with the
ques-
t#;f *i;ai.;tiii".".r."r
oortance
in education or caste would account for differ- Lit'e. Cambidge: Cambridge University Press.
Sapir,'Edward. L{27. Speech as a Personality Trait. American lournal of Sociology
u^o"g t"bfi;Pt;i Indian informants' but they dismiss too easily
""."t
ir," p*rilirfty tn!i-"i".uti"n, and especially the American within-the-
32:892-905.
Shweder, Richard A., and Edmund |. Bourne. 1982. Does the Concept of the
;;dilt aisiourse setting _-_ih" ,q.-u.ican informantsandwere counseling
nur?gry.sclogl Person Vary Cross-Culturally? In A. J, Ivlarsella and G' M' White, eds',
pry.t ort girtr, U;;;ity 3r cni.ugo undergraduates,
Cultural Coiceptions of Menta[ Heatth and Therapy. Dordrecht: D. Reidel, pp.
teichers,"all three of *fii.h g.o.rp"r *e ou{h,t to expect to traffic in
"ab-
,,context-free" asciiptions, especially when responding to qu€r- 97-137.
stract,, and
silverstein, Michael. 1984. on the Pragmatic 'Poetry', of Prose: Parallelism, Rep'
i", ir.^ Ugiversity of chi"ugo'r"ruarchers - mignt account for the differ- etition, and Cohesive Structure in the Time Course of Dyadic Conversation.
ences across cultures'
In D. schiffrin, ed., Meaning, Form, anil Llse in Context: Linguistic Applications.
15. I have said little about two converSent traditions (with Hymesto 1962be
estab-
the-key Washington, DC: Georgetown University Press, pp. 181-99'
lishing their convergence) that do"not take lexical dlscovery
White, Geofirey M. 1978. embiguity and Ambivalence in A'ara Personality De-
to the",,languug" of"e-o[ion.,, one is modeled.on sapir's work
(1927
-es-
both identifylargely nonlexical index- scriptors. American Ethnologist 5:334-50. -
p".irirvl. uf;a tfi" other is-praguean;
"emo- White, Geoffrey M., and |ohn Kirkpatrick, eds. 1985. Person, 941,-!!d Experience:
i.uls of .p"e.h, usually called;expressive devices," that subserve an Exploring Pacific Ethnopsychologies. Berkeley: university of California Press.
funition.- fi ir important to see that such efforts rest explicitly on
tive
unirrersal_typological aslumptions as defined in advance by. the observer.
Thev analvzi indefcals buthot semantics and reference as localphenom-
taken up in
l* , airr*""t species of analysis from the one I have way
"ir,'tt
it i, "y I*il; J;;;" a similar diitinction between "the people
,,what upon to impute
"nr[rt"r.
;diil iur.rfi emotion,, and they actually rely

*
Llntouchability and the fear of death L87

lage, and had asked her in advance if she would agree to be recorded.
She had obviously given some thought to the proposal, and perhaps
9. UntouchabilitY and the fear of had even'rehearsed the song in advance, when Kuppucdmi returned to
her house some weeks later with my portable cassette tape recorder. The
death in a Tamil song recording session was quite formal, including a brief, audience-directed
introductory dialogue between the performer and the "M.C," as though
for a radio program. C€vi chose in particular this song as one she would
MARGARET TRAWICK
sing to a Western audience.
&" CEvi sang the hymn in a strong, clear voice. Her performance took
+-- nearly an hour. The name "singamma" formed a constant refrain, re-
peated;{t the end of every line in the song. The- name as C€vi sang it
was quiet, low in pitch, persistent, like a moan. Sometimes
Introduction in "singamma" quavered heavily, ending in a fall to a hang-
la- ing tone, like a sob. It was as though C€vi were feeling for Singammd.
This is a chapter about a song sung by C€vi, a landless agricultural
District in Tamil Nadu. CEvi But C€vi's voice never broke. Her performance was highly expressive
borer living in tt"r" northern part of Madurai
castes but also strongly controlled - between the abandonment of wept singing
belongs to-the Palaiyar jati, one of the largest of the untouchable
and the expressionlessness of chant. In its precisely measured, trem-
in south India. Paraiyars are untouchable (tludd) to higher castes be- bling passion, it somewhat resembled classical south Indian music.
cause their traditional duties, still performed by them in this
part of Tamil
Nadu, include the handling and processing of human and animal feces
these duties matter-
and corpses. The Palaiyars of Ccvi,s village perform The story told by the song
of-factly and conscientiously, and in general are firm in their
adherence
of Tamil villagers, even as they some- The story of Singamma, as tecounted in C€vi's song, is stark and ellipt-
to the Lstablished social values
rank and their economic situation ical.s As it opens, we learn that Singammd is kept confined in the house
times militantly seek to improve their
woman by her brothers' wives. But one day, as she sees her brothers' wives
vis-i-vis higher castes and iandowners'l Ccvi is a young malned
is also coming, she "puts on different clothes" and slips off to the market to
of twenty-iive and is the mother of four children, all sons.2 She
sell trinkets (a traditional occupation of Kugavars). Her brothers weep
considerld to be one of the best singers in the large village where
she
performer' She is one of the when they come home and find her missing. They go to the market,
lives, although she is not a professional
from, the vast and varie- where Singammd is playing cymbals (cinkipAddu) and selling songs at the
many thousJnds who contribute to, and draw
dried-fish store. Her lame older brother (noudi a44an) meets her there
gatei corpus of unwritten verbal art in Tamil Nadu'
her to but does not recognize her. Still in town at dusk, Singamma goes to a
The work of C€vi's to be considered here is a hymn sung by
un- wedding feast, gathers up the leftover rice from discarded leaves there,
and for the local Kr.&var goddess Singammd!' Kulavars are another
the Palaiyars and untoucha- and brings it back home in jars (Kusavars traditionally get the polluted,
touchable caste in Tamil Nadu, lower than
Marathi leftover food discarded at weddings). For punishment, her brothers send
ble even to them.3 Kulavars live as migrant scavengers' claim
her to a house in the forest and make her stay there, "grinding and
ancestry, and distinguish themselves from the Tamil people surround-
I shall pounding with a mortar." All four brothers go into the house, pull the
ing them by noticeable features of dress, speech, and way of life'
doors shut and lock them, and the mortar stone pounds. Then Sin-
say more about them PresentlY.
-ce.ri,, gammd says: "The sun has set on our good caste; we are excluded from
performance of the hymn to singammal was recorded by my
in the caste (j dtil€y um tallua aSi)."
assistant, Kuppucdmi, a twenty-year-old male kinsman of Ccvi's,
The lame older brother's wedding day arrives. Singammi is to cook
,pri.,g oi tg4.n Kuppucdmi hid explained to Cevi that I was interested
and sen e at thc wedding feast, but the pots of rice and milk won't boil
in "nonmovie, nonwritten" songs that were sung by people of that vil-
185

h, tft
L88 Margaret Trawick Untouchability and the fear of death 189

for her - the sun will not allow it. (The Tamil festival of the sun, cele- self." Then, as the song goes, Singammd, having emerged from within
brated around winter solstice time, is marked by the boiling of a rice- the house, "rising up high, speaking with unsheathed energy/ wearing
milk dish called ponknl.If the pot boils over, it means that the sun will pearls," addresses her lame older brother, telling him, "You are the one
bless that family with abundance in the coming year. If the pot fails to who killed me, who saw my sin, who undid me. That which entered
boil over, it is an evil sign.) the house truly was. Tell me to rise up outside. Now I will stand up
Singamma is again locked in the house, "weeping in her heart." But straight and show you." She leaves the house and goes outside; they
at dusk she declares , "If my honor (patti) is destroyed, let the doors of raise her up. And the final stanza of the song says, "As soon as they
the house stay shut, but if my honor is undestroyed, let them open." raised you up, Singammd, your building, too, stood up tall."
the doqr$dpen, and Singamm6 flees. Her brothers come home and ask
where Sfigamma is, and their wives say, "Ahusband has married her
and taken her away." The brothers then dig holes in the floor inside the Pollutio;&and fear
house. The holes are to be Singamma's Srave' w*
A multiflrrlle of questions might be asked of this song. Here I will be
Singammd lies down to sleep, with her head on her mother's lap. Her
mother tells her that because she is excluded from the caste and pots asking three general categories of question. First, what is C€vi doing when
will not boil for her, her brothers are going to kill her. As she lies with she sings this song to my tape recorder? If the song is, among other
her head in her mother's lap, she takes a louse comb in her hand and things, an act of self-expression, what is being expressed? In what does
assumes the form of louse eggs. While she is in this form, her brothers this self-expression consist? Why does it have to be done?
pound rice for her mouth (rice is put in the mouth of a corpse before Second, of all the songs that C€vi might have chosen to put on the
burial) and drop chicken and crow meat in the holes in which she is to tape recorder, why does she choose this one?
be entombed. While the lame older brother weeps, they kill ut d*dit- Third, given that the song is by an untouchable wom€ul about a woman
member (kulaicukum) her human body and bury it in the holes. Only the who is untouchable to her, and since it is a song about death, about
heart of the mother (atta) is not pleased by this act, says the song. being defiled and polluted, can the song tell us anything about the
From Singammd's grave, a poisonous red oleander plant (ceattarali) meaning of such pollution in South India, about how it feels to be be-
springs up. The lame older brother comes to the place where the olean- yond the pale?
der has grown. A poisonous earthworm (punaknm) emerges from the Much has been written on the topic of "ritual pollution' in South Asia.
oleander and tells the brother to build a palace there for Singammd. The The entire caste system, as well as the internal organization of castes, is
brothers crumble the house and build a foundation on that site; when saicl by some to be founded on this concept.6 The pollution avoidance
they do so, an oleander bud blossoms for Singammd. Then the lame that separates caste from caste is often described by ethnographers as
older brother touches (tindu) the earthworm, which, in retum, tells him though it were a mechanical process with little affective content. A Brah-
a story. It says that it once lived in the house. It says that Singammi's man is said to cleanse himself of "ritual pollution" (after a death in the
brothers' wives saw [[ and told Singamma to catch it and take it awayi family, for instance, or after having touched a polluted person) by pour-
both it and Singamma left the house, "weeping and sobbing." The ing water over his head; he does not really scrub himself, as he would if
earthworm concludes its story by saying, "Even if you build palaces, he were sprayed by a skunk or fell into a sewer. And yet, compulsive
you may not stay. Only the lame older brother may stay." cleanliness often does go along with extreme orthopraxy among Hin-
All the brothers and their wives then come to that place to burn lime dus. Similarly, one searches in vain through the ethnographic literature
for mortar. As soon as they have burned and moistened the lime, build- for testimonies on the part of high-caste people concerning the gut re-
ings grow up in that spot. A woman comes to live there, and when she vulsion they feel toward low-caste people. And yet if one asks, one finds
does so, the building in which she lives "leans over with its founda- that revulsion is there. It is much stronger in some individuals than in
tion." Singammi then appears to the woman. The woman asks her where others, for no two people even in the same family have the same emo-
she comes from. Singammd replies, "I came from within the house it- tional makeupr but for all its variability, revulsion toward pollution and

.
Untouchability and the fear of death T9L
L90 Margaret Trawick
thepollutedisawell-articulated,culturallysanctionedfeelinginTamil Palaiyars and Kuravars
Nadu.
is called in Tamil Tamil Palaiyars are in a difficult situation, for in many South Indian
A response of revulsion to someone or something
payam|u*o'dthatweoftentranslateintoEnglishaS,fear.',foritmeans village communities, they are tlis fence. They are untouchable to others
even with the substances of death, all those substances that
inlr, ,oo. An ugly Person is said to be 'fearsome" payanknramLkn)woman
because they deal
iy friends who inow that the personis harmless' I have heard one are shunned by others in order that they may feel live and whole. Ac-
ar- cording to the overwhelming consensus of the society of which they are
,ly of another that she por,""tt 'fearsome beauty' (payanknrymanathat
the ancient Tamil idea part, Palaiyars' own living bodiesbelong among those substances of death.
aku\ -astatement that perhaps harks back to
i, dur,g"'o"s; she mav literallv take a man's breath They are filth, merged with what is excreted and left behind. Hence they
;;:*ilii,';;;;" -are Tamil thought. Ex-
;;;r;;;"ry ""a death closely associated in are avoided , feared, and despised.
treme attractiveness and extreme repulsiveness
have the same danger- One o$ective way that Palaiyars have found of dealing with this ab-
ject staffib to deny it by passing it on to others, placing these others
ousemotionalpowerandareoftentreatedastransformsofeachother'
Both are to be avoided. betweenlhemselves and death. Traditionally, the scapegoats for the
the obiect of payam' Pagaiyars have been Kulavars, who, because of their way of life, admi-
Anything that a Person avoids may be said to be
A man concerned for his reputation says that he fears
(payppafluttcn) rably fill this symbolic role. Although most Pagaiyars are landless, hence
said tobe fear inducil,S impoverished and symbolically disembodied, Kuravars are even more
nothing but shame (payi).Thiiving little.boys are
on a plate of meat calls conspicuously without a place. They have no fixed homes, no mud huts,
(payartiaramdkn). Au"g"iu'iu" ryi-tootilgdown
ii e signt that he seeJ and his feeling when he sees it fearsome (paya- but only rag tents and a territory through which they wander, a territory
is part of its nature' A that in no sense belongs to them. Their regular haunts are train stations,
mam|lOfa timid child, it is said thatiear (payam)
payam' bus stops, markets, and festivals - crossroads of every kind. The women
child afflicted with the evil eye is said to experience -A*voman
death of a child refers to her feeling in their daily rounds pick through garbage bins and gather rice off the
suffering chronic anxiety after the
he may become sick after- eccil-ilai, the 'spittle leaves', thrown away by others after meals and feasts.
as payam.If a person is suddenly frightened'
in the dark forest may This is not the desperate strategy of a starving few but the normal and
ward as-a consequenc e of payam., Demons lurking
them and inducing payam highly visible way of life of the Kuravar community.
kill their unwary victims by suddenly surprising
to be what kills the The word kur-atsar,'hill dweller', is in modern times a polite name for
in them. This payam, the ieeting of iear itself, is said
a foraging community elsewhere called kuruoikkarankn. The latter name
demon afflicted.
Tamil payam,like English "fear," isa complex term
whose.meanings derives from the word for small fowl (kuruai), which men of that com-
are varied and diffuse. io attempt to define
it precisely would be a mis- munity hunt. The commonest kuruoi of present-day Tamil Nadu is the
take. However, one can go so far as to say that
Tamil payltn' much more aggressively greedy, noisy, and messy English sParrow. I have heard
In the eyes of most one Tamil woman say, on observing two English sParrows copulating,
than English "fear," is a"thing itself to be avoided'
Tamil people, pay**napPears ti be not so much
a healthy' life-preserving "You shameless thing; you will do it even to your own daughter." One
instinci as a dangerous, life-corroding disease' implication of the name kuruaikkarar is that its bearers, like the birds they
is surely
If there is one feeling that goes with pollution, that feeling hunt, are lawless, dirty, and of no consequence.
payam. To submit to piyam islo lose control'
to come under another's In particular, however, Kuravars are known for hunting and eating
Payam is the feeling scavenging beasts and birds, such as jackals and crows. In modem times,
power, to let part o. il of one's life flow away'
to happen' To survive in Kuravar men may still be seen wandering with their rifles through fields
evoked by death, and it can cause more death
around it or around and wastelands, with crows their principal quarry. Kuravars are also
the face of it, one must Put some kind of fence
oneself. known for keeping pigs, which feed on human excrement. Other castes
also keep pigs, but Kutavars are the only people who will eat crows, for
crows are, above all others, the birds of death. Crows, like some other

t
192 Margaret Trawick Untouchability and the fear of death 193

birds, eat carrion. In ancient times, human corpses used to be offered value to ethnographers interested in the relations between gender, lan.
by Tamils specifically to crows (Tirumantiram verse 191), and crows are gaage, and the construction of the self.
still treated as the embodiment of dead ancestors and given offerings of The basic premises of Kristeva's work might be summarized ae fol.
rice on the festival day of ponkal, which marks the end of winter, the lows:
season of the crow, and the return of the lifegiving sun. In Tamil the cry Any human language consists of a set of conventions encoding ontol-
of the crow, ka kfi, is a cry for protection (knppu) and evokes images of ogy - how the world is, how it is supposed to be. These conventions are
f.eces (knk*a), darkness, and the wandering wind (kattu). largely (in Saussure's terms) arbitrary and (in Peirce's terms) legislated,
Palaiyars are despised by other castes for eating the meat of cattle, the Such conventions join together to form a whole, a coherent system; at
purest, most lifegiving of animals, whose destruction, therefore, is the least they present the appearance of doing so. This seemingly whole and
gravest of sins. But as eaters of crows, Kuravars eat the eaters of death. coherent system may be called the "symbolic world." It is the world of
Finally, Kuravars also are thought to be sexually lawless. The term meaning and relations, of totality and law. It is also called (by lacquee
cirtki in Tamil means both 'Kuravar woman' and 'whore' (the name Sin- Lacan) the "Law of the Fathe{' (Le Nom du Pdre) because it is legislated,
gammd, or cinknmmd according to the system of transcription used in so it seems, by a distant and inaccessible authority. It is already there
this paper, would be synonymous in Tamil with 'Kuravar womad. Singh before one is born. The individual self has no say in its decrees.
is a North Indian surname meaning'lion', whence the Tamil word for Entry into this phallic, totalizing, slrmbolic world, the world of mean.
hon, cifiknm). The notion that Kulavar women sell their bodies in the ing created by the ancestors, begins with the subject-object distinction,
market together with their trinkets is helped along by their different The self, in order to exist and have meaning, must be delimited and
style of dress; they wear a kind of skirt and blouse instead of a sari. In defined. Self and not-self, subject and object, are defined against each
songs sung by Pagaiyars, Kuravars are stereotypically and mockingly other, as each other's opposites, each other's negations. Inasmuch ag
'w
accused of brother-sister incest. "there exist only differences" (Saussure), the self, the subject, can only
For this reason, CCvi's song came as a surprise to me. Rather than be defined negatively, as not this or that. I am Jane; I am not Joan or
mocking or rejecting the Kuravar heroine, CEvi's song shows deep sym- John. I am little; I am not big. I am child, not mother or father. I am part
pathy for her. Rather than pushing the Singammi song to the margins of the world; I am not the world. I follow the rules; I do not make them.
of her repertory or giving it short shrift, C€vi places it at center stage The speaking subject, therefore, the self that knows language and ite
and honors it with a style of performance both formal and moving. IA/hy? own name/ is by its very inception not whole. It feels itself to be incom-
plete and fragmented. In this sense, it is castrated. It desires its opposite
to make it whole again. It desires to embrace, to own and to be the
symbolic-system-as-totality. It desires to become the father, to appro.
An approach to the study of fear and pollution priate the code in its perfect wholeness. But it never achieves this end,
In seeking an answer to this question, I have turned to a recent book by because the perfect wholeness of the symbolic world is only an illusion
the french literary critic ]ulia Kristeva, entitled Powers of Horror: An Essay and, just for this reason, stays always out of reach.
on Abjection, which discusses Indian thought (as represented by French Prior, however, to the subject--object distinction, prior to the relation
ethnographers) at some length.T Kristeva's approach to literature, lan- between self as part and the Law of the Father as whole, one dwelle in
guiage, and culfure is feminist, psychoanalytic, and semiotic.s In partic- another state, in a relation between what might be called the "proto.
ular, she argues that language is, before all else, a sexual act: The for- self" and the body of the mother from which one has not yet been fully
mation of the speaking subject and its entry into the symbolic world are separated. The body of the mother is different from the Law of the Fa.
founded upon the division between masculine and feminine. I would ther. The Law of the Father is an abstract code that governs the self
not claim that the psychoanalytic-cum-semiotic approach elaborated by through definition. The mother's body is a physical presence that con.
Kristeva is the only one that might throw light on the significance(s) of trols the child through physical pressures.
Cevi's song, but I would suggest that this approach may be of great In the preaymbollc atate, before the Law of the Father takes over, one

*fte:
;

Llntouchability and the fear of death 195


194 Margaret Trawick
There are, perhaps, some people who live their whole lives in this
has just been born. one has just been through "the immemorial vio-
twilight, presymbolic state, a state not of radical, once-and-for-all sepa-
lence with which a body becomes separated from another body in order
ration of ielf from other, but rather a graded continuum of always partial
to be,, (Kristeva L982:10). From here on out, one will keep oneself intact
separations. In South Asian society, argues Kristeva, an absolute sepa-
only by constantly repeating such extrications. Birth is the first abjec-
,utio. of self from other, self from mother, is never established. Instead
tion, the first casting out. what has been abjected and left behind is the
of this single, radical separation, there occurs a series of multiple, graded,
body of the mother, which, however, will still threaten, as long as one
and partial separations - rituals of purification, acts of defilement, the
has not fully separated from it, not just to take away a part of one's being
caste hierarchy. The feminine is not foreign, alien, and other: It remains
(castra$on) but to engulf one's whole being entirely.
part of one's own. Marriage is endogamous; the boundary between in-
The B$hV of the mother, therefore, is prior to and more powerful than
cestuous and proper unions is shifting; and the horror of incest, of being
ttte l5* oi thu Father. The encompassing precondition of meaning, it
swallow-qs up again by the mother, is great. self is divided from other,
threatens meaning itself.e If meaning is the totality composed of arbi-
not absffipely, as subject from object, but relatively, as inside from out-
trary, legislated distinctions, there is still this presymbolic state that points
side. Selffiw"Ui"g i" ihe world of nontotal separation is defined in terms
to the void upon which meaning is founded. Hence the detotalizing
of place (the physical pressures against the body) more than in terms of
effect of the outcast, of that which has been expelled from the self at
name (the words used to define the person). It wanders over a territory,
birth, but not.quite completely. "The iettisoned object . . . lies outside,
never fully detached from this territory, the soil that nurtures it, the
beyond the set, and does not seem to agree to the rules of the game.
earth from which it sprung. The Law of the Father - all totalizing social
And yet, from its place of banishment, the abject does not cease chal-
codes and conventions, all arrangements of the superego, above all, lan-
lenging its master" (L982:2).
guage - never quite displaces the authority of the mother. Language
The abject is that which is losu the object is that which is lot yet
obtained. The object is the Law of the Faiher. The abiect, whi[h pre-
gr"i u form to and encloses the horror of death, as speech replaces the
mother's breast in the mouth - but never quite completely'
cedes, encompasses, and threatens to overthrow the object, and with it
The abject is what is left over, what is remaindered. It is what each
all of significance, is first of all the body of the mother. At the other end
self leaves behind in order to be whole and live. But in,south Asian
of life, taking the place of the mother's body, is the corpse, from which
thought the remainder is also, often enough, the source of new life.
also one is incompletely separated and which also defies meaning:
"Here is perhaps the essential point," Kristeva concludes:
The corpse does not sxgnlfu death. . . . Without make-up or masks,
The remainder appears to be co-extensive with the entire architec-
refuse and corpses show me what I permanently thrust aside in or- ' ture of non-totalizing thought. In its view there is nothing that is
der to live. These body fluids, this defilement, this shit is what life
everything; nothing is exhaustive, there is a residue in every sys-
withstands, hardly and with difficulty, on the part of death' There,
tem - in cosmogony, food ritual, and even sacrifice, which depos-
I am at the border of my condition as a living being' My body
its, through ashes, for instance, ambivalent remains. A challenge
extricates itself, as being alive, from that border. such wastes drop
to our monotheistic and monological universes, such a mode of
so that I migAt live until, from loss to loss, nothing remains in me
thinking apparently needs the ambivalence of remainder if it is not
and my entire body falls beyond the limit - cadere, cadaver'
to become enclosed within one single-level symbolics, and thus al-
If dung signifies the other side of the border, the place where I
ways posits a non-object as polluting as it is reviving - defilement
am not and which permits me to be, the corpse, the most sickening
and genesis. That is why the poet of the Atharoa veda extols the
of wastes, is a border that has encroached uPon everything' It is
defiling and rejuvenating remainder (uchista) as precondition for all
no longer "1" who expel. "I" is expelled.
forms. "IJpon remainder the name and the form are founded, upon
. . . The corpse is the utmost of abjection. It is death infecting
remainder the world is founded. . ' . Being and non-being, both
life. It is something rejected from which one does not Part.
are in the remainder, death, vigor." (L982:76)
(1982:34)

=t:
.

196 Margaret Trawick Llntouchability andthefear of death 197

My analysis of C€vi's song is built on the thoughts of Kristeva just and (I mean no disrespect here) like that portion of himself that a baby
outlined. One of the things I wish to show is that C€vi, being always first learns to part with, ccvi's tape-recorded song is her loving gift to
close both to the corpse (because she is a Papiyar) and to the body of the world beyond her ken.ll Cevi's song exPresses the state of incom-
the mother (because she is female and Indian),1o is not fully governed plete separation, the remaindered state, that engenders antitotalizing
by Language or the Father in the way that we are: She breaks their rules, visions. As she sings she becomes, like the goddess she sings of , a fla-
she makes her own, and she does so systematically. I would suggest grant rule breaker, rejecting and reconstructing languages, making room
that many aspects of CEvi's style point to an actively detotalizing artistic for something new to grow.
technique. By calling this technique "detotalizing," I mean that it not
only rffiIses to be bound by any single, internally consistent, authorita- Detotalizing strategies in C€vi's art
tivellhegislated, holistic set of rules for linguistic behavior, but that it
actively sabotages those rules, like a bomber selecting key structures to We mav speak of events on at least two levels here. On one level are the
explode. ,*io,rr.#foes and events of the story, what the story is about' We might
,,level of the explicit." On another level is the way the story
C€vi as an untouchable, a handler of feces and corpses, is more threat- call this ffi
ened than others by those very powers of horror, the engulfing powers is told, what the story expresses. we might call this the "level of the
of death and the fear of death, that have been thrust upon her by those implicit."
others in ordei that they might live, in order that they might be whole, On the level of what the story is about, there are numerous discom-
clean, and intact. She must face this darkness directly and somehow forting reminders of the nonwholeness of existence, as the song dwells
give form to it if she is to give form to herself. But in order to do this, on the themes of death, exile, twilight times, leftover food, and remain-
she must shatter the perfect image of the other that the other has created ders. These themes are mingled with a whole series of violations of Tamil
for himself. She must shatter the code. Her voice is not a revoltitonary social code: Singammi breaks out of the house against the will of her
voice, for it does not seek to replace one order with another. Rather, her brothers; she sells her songs at the dried-fish store (fish being a conven-
voice upholds the twilight world, the world of no one order, of no one tional symbol for female genitals); she gathers up polluted rice to take
clear way of seeing, the world of mdya, of multiplicity, of relative rather back homei her brothers raPe, murder, and dismember her; they bury
than absolute differences, of countless gradations, of remainders. her in the floor of the house; they build one house on the remains of
In performing her hymn to Singammd, C€vi confronts her own abjec- another.
tion, and as she does so, her voice becomes the voice of abjection that But I would like to consider here not so much the "what" of Ccvi's
,,how,,, precisely those aspects of a tale that we do not foots
Kristeva describes, "from its place of banishment . . . challenging its story as the
master." Singammd is to C€vi the horror that threatens to swallow her, on when we hear it, because they point away from themselves - those
the corpse from which she, even more than others, cannot part. The aspects that we call "sFyrle," that are all the more powerfulbecause they
goddess of this hymn is, like so many Indian goddesses, the mother of are not always subject to the full light of consciousness'
death as well as of $rth. In CEvr-s song, her being as corpse is especially
strongly emphasized. She is also the excluded, cast out, abject in that Anomalies of inclusion and exclusion
sense. She is to C€vi what CEvi is to the rest of Indian society, and she
is at the same time what all people must exclude from themselves in In C€vi's performance, on the level of explicit content, the level of focus,
order to be whole. The terrible dilemma of the untouchable: self-defini- , p."o..,rpution with problems of inclusion and exclusion is manifest in
tion. She is not whole, she is fragmented. She is defined in terms of the repeaied themes of confinement and exile. Simultaneously on the
place, above all in terms of within and without, and she is also without level o] nonfocus, on the level of mmns ofexpression, states of exclusion
a place, wandering. In inscribing the song of Singammi on the tape and inclusion are repeatedly remarked, stumbled over, defied, and com-
recorder, C€vi may grve a place to her fears - not only name and form, mingled.
but an objective locality to her abjection. Like the scription of the author, n key example of this scrambling of the categories of exclusion and

:#E
..

L98 Margaret Trawick llntouchability and the fear of death L99

inclusion comes in the section describing the caste exclusion


of Kuta-
Singammi is driven from our caste, or our caste will be driven from among the body
vars. At the beginning of this section, the song tells how
market. of castes.
to the
punished by he.-r brottiers for leaving the house and going Most interesting are the details of person and case in the last few lines.
they lock her in the house alone in the forest (stanzas 336):
The sun sets "in the idti"; the brothers and their sister are "excluded in
Leaving you within the house, Singammd the jati." There are ways to exPress "olnt of" and "away from" in Tamil'
All four doors, Singamma, Singamma but CEvi chooses not to use them here. And in the phrases "in our good
They closed and locked and came, Singammi' Singamma' jati,, and "ue ate excluded," the exclusitse (excluding the listener) rather
A$r{our men, Singamma,
than the inclusitse forms of the first person plural are used. If a sister is
ftllfour men, Singamma, Singammd, addressing her brothers regarding "ou(' caste, it would only make sense
Went inside the house, Singammd, Singammd' for her t8- use the inclusive form. Although the topic of the passage is
And as they went inside the house, SingammS' exclusibt*lhs pronoun "we" shouldbe inclusioe. As though to underline
The mortar stone pounds, doesn't it, Singamma' Singamma' this anoffii1z, two stanzas later, the brothers speaking to the sister use
trne inclusiai form of "we" to refer to precisely that body from which
the
The mortar stone pounds, doesn't it, Singammi'
sister has now been excluded (stanza 40):
And as the mortar stone pounds, Singamma'
All four doors, Singammi, Singammd, Because you acted wrongly, Singamma,
They pulled shut and locked, Singammd, Singammi' Excluded from our (incl.) jati, Singammi, Singammi,
To them, to all four of them, Singamma, Excluded from the idti, Singammi, Singamma'
Singamma could give an answer, couldn't she' Singammd'
In our (excl.) good jdti, Singammd, '{'s Here it seems as though the ingroup-outEroup ambiguity is striking
In our (excl.) good jdti, Singamma, Singamma' at the heart of grammar.
The sun (sg.) has fallen and gone (pl'), One more deviation from the code deserves mention here, for it also
Singamma, SingammS, we (excl') points to the link between ambiguities of inclusion and ambiguities of
m if,e tali . . . in the iati are excluded, marriage rules. The deviation is actually a slip - a Freudian slip in the
Singamma, Singamma. strictest sense. Having taken a breath, cevi begins a line, "tdlilEyum
implies gang
. .,, she then stops abruptly, returns to the beginning of the melodic
The first part of this passage, although indirect' strongly phrase, and sings, "iitil€yum talluvadi Singammd Singammi'" Evi-
of incestuous multi-
rape of Siniamma by her four brothers. The image
dently she has said. tali where she meant to say iati. The frill is the mar-
outside (are the
ptl ,up" ls iinked to ambiguities concerning inside and riage necklace that a husband ties around his wife's neck on their wed-
the house where singamma is confined, or outside it?),
Lrothers inside ding day. It is a powerful symbol for Tamils of the bound and confined
Singammd'
concerning coming and going (are the brothers approaching state of married womanhood. The term jdti means'caste'. The whole
desire and repulsion (do
or are they"withdr&wing from her?), concerning line could be translate d, "we are excluded from the tdli, (I mean) from
oPen and closed
the brothers act out of iove or hatred?), and concerning the jati, singamma." However, in this line, both tali and idti are in the
(what is happening to Singammd's body?)' locative case, so that a more literal translation would be, "We are ex-
Asawtrote,thepassageSeemstopointtothelinkbetweenquestions cluded in the tali, in the jdti, singamm6." As I have mentioned, there
of incest and questions of caste inclusion. (/rilti literally means "birth are precise ways of expressing the concept "out of" in Tamil. But in the
group." It is generally glossed in English as "caste"') The dilemma
is
of the preceding stantza, CEvi, for whatever reasons, chooses not to use them'
Iirr,pi", To sta:y within tle caste (as well as to abide by the rules io ,uy "in" when she means "otJt of-" AII such evidence considered
caste system as a whole), we must marry within our
grouP' But if we
toget-her suggests that questions of inclusion and exclusion, problems
marry ioo closely within our grouP - or in other words' if we stress
de?ning wnit is inside and what is outside, run deep in C€vi's mind.
inclusion too heavily - we wilibe radically excluded: Either we will be
On tiia Bame level, the level of nonfocus, there is also a scrambling of

t =' Hl
i

Untouchability and the fear of ileath 20L


200 Margaret Trawick
er's directly addressing the subject, calling out the name "Singammd"
the categories of singular and plural. This is yet another way in which
the concept of totality may be subtly fragmented or, conversely, the con-
in every line; by the frequent use of the second person (e.9., stanza9,
"You are going to the M€lur Market to sell baubles, Singamm6"); and
cept of fragmentation may be expressed. So Singammd's body becomes
plural - plural louse eggs and then plural pieces buried in plural Sraves.
by the equally frequent use of tag questions (e.9., stanza 24, "On the
stony ground of MEhfr, a wedding is taking place, isn't it, Singammd?").
The sun - by whose masculine authority (referred to as such in the song)
But these are intermingled with references to Singammi in the third
Singammd is expelled from the caste - also receives a plural marker. The
person, as though she were absent.
plurality of Singammd's brothers is pluralized; throughout the song, their
Then there is the device we call "enjambement," which frequently
numberq.keep inexplicably changing: sometimes they are four, some-
occurs in all varieties of Tamil verbal art and which C€vi employs very
tlmes fpil,isometimes seven. And finally, the truth that Singammi speaks
heavily in the hymn to Singammi. Enjambement appears when a metric
when she emerges from her grave is called not the truth but "truths,"
line (oq 4usical phrase, or set of words uttered in a breath) does not
plural.
correspfufuth a grammatical sentence or phrase, so that the first word
of a newBentence appears at the end of a metric line and the sentence
Remainders and incomplete separations is completed on the next line. What results is a kind of verbal "remain-
The detotalizing vision of the Person who is subject to abjection, argues
dering." In the hymn to Singammi, almost always the word that is so
remaindered is ni,'you', so that many lines end Singammd ni . . . , "Sit't-
Kristeva, is founded upon incomplete separations of all kinds. It is
gamma, you . . ." What is predicated of "Singammi you" is left to the
founded upon remainders. No matter how much you try to separate
next line. Meanwhile this "you" hangs in limbo.
thing A from thing B, other from self, there is always a little bit left over,
that sticks on. No system, especially no living system, is herme$cally
sealed. No single vision can account for everything. Every paradigm has The way the name "Singammi" wanders
anomalies. All grammars leak.
However, this detotalizing vision, the principle of the remainder, man-
CEvi's song expresses, above all else, this state of incomplete separa-
ifests most strongly not through poetic devices like enjambement, which
tion betwEen self and other, of fertile remainders proliferating new vi-
themselves by now have become almost like grammatical rules, pat-
sions and (we would hope) new cultures. The stumbling over the inclu-
terned and predictable, but rather through a more troubling, though
sive versus the exclusive "we" is a case in point. "Are you and I separate,
seemingly simple, thing: the repeated refrain "Singamma," the name
or are we together?" this stumbling asks. "Are brother and sister one,
cried out once, twice, three or four times at the end of each of the more
or are they not?" A similar relation mediates CEvi, the singer, the un-
than 300 lines of the song. The hymn is to and about Singammd; the
touchable, and Singammd, the goddess, the mother, the corpse, un-
name cried out is always the same and always (ot almost always) in the
touchable to the untouchable, other to a self who is herself other. These
same place - the end of the line. For she is the mother of death. Her
two, also, are partly but not fully separated.
musical presence is constant. And yet, from the point of view of gram'
In the song, Sing#nmi is "you," CCvi is "l." C€vi is the singer, Sin-
mar, Singamma's place in the song is ill-defined. The cry of her name
gammd the topic of her song. But this topic and this singer, this "you"
floats through the song like a ghost, its place in the total order of each
and this '"1," are never very far apart. A sense of immediary runs through
stanza, its relation to the whole of each sentence, shifting, inconstant,
every line. This feeling of closeness is produced, first of all, by Ccvi's
and therefore from point to point difficult to determine. This indeter-
preferential use of the present tense. In this song of 101 stanzas, 41" ate
minacy of the grammatical status of the name "Singammd" affects the
in the present tense, 31 are indefinite or unspecified as to tense, and29
narrative structure at crucial spots and fills it with doubt.
are in the past tense. The tenses are all mingled. Sometimes we are Pre-
As we follow the journey of Singamma, of the name "Singammi,"
sent, other times inexplicably past.
through the song, we become lost - I became lost - because the degree
This feeling of partial presence, of. relatiae nonseparation between the
to which thie name ie bound to the narrative, and the ways in which it
singer and the subject of the song, is further strengthened by the sing-

;€iEEl=
202 Margaret Trawick llntouchability and the fear of death 203

is bound, both vary widely. In the first two stanzas it is clear: singamma
the refrain, it seems that she addresses her as "mother, Singamma" -
is being addressed by the singer, who is going to narrate to singammd the same pair of terms by which she herself is addressed by her own
her own story. "I am going to tell the story of your birth, Singammd," mother subsequentlY.
CEvi sings. But this clear division between singer, topic, and addressee,
In yet another place (stanza 94), Singamma addresses her lame elder
or more generally between narrative and narrative context, very soon brother and accuses him, saYing:
becomes blurred. For instance, when one of the characters speaks in the You are the one who killed me, Singammd,
song, that is, when there is embedded speech, then the refrain "Sin- The one who saw my sin, Singamma, Singamma,
gamma,,_becomes ambiguously incorporated into that embedded speech. The one who undid me, Singammd ' . .
Then i*-iFho longer direct address, no longer the singer's own word to
the gofiIess, but something removed from both of them' In stanzas containing dialogue between two charaeters, neither of whom
In other places, Singammd herself speaks, and then the cry of her is Singar4ma, the refrain apPears completely free-floating, as when the
name seems not, as when Singammd is addressed, embedded in the brothei${$ives say to the brothers (stanza 51):
*
rB

narrative and thus distanced from the context, but rather the reverse - "Your younger sister has been taken away
dislocated from the narrative, an intrusion from the context. For ex- by a husband, Singammd, Singamma . . '
ample, consider stanzas 43-44:
Finally, there are stanzas in which the name "Singammd" becomes
Six o'clock has come, Singammi. split into two forms, one of them tightly bound to the syntax of that
Six o'clock has come, SingammS, Singamma' purti",rt* stanza, the other either attached as a vocative or divorced from
If my honor is destroyed, Singammd, I too, ihe text and free-floating, in stanza 51. For instance (stanzas 33 and 37):
If my honor is destroyed, Singammi, I too 4ry

Must stay inside the house, Singammd, Singamma, I am All four brothers having spoken, Singamma,
A woman of perfect honor, Singamma, Singammi, They sent Singamma away, Singamma, Singammd '''
A wgman of perfect honor, Singammd, SingammS, these
To them, to all four of them Singammd,
Four doors must open up, Singamma, SingammS.
Singammi could give an answer, couldn't she, Singammd?
Here as though the singer is addressing singammd
it sounds at first
directly in her own voice, the voice of ccvi. she seems to be saying, "I, In sum, as one listens to stanza after stanza of the hymn to Singamma,
too, must stay confined, Singammd. I, too, am a woman of perfect honor' one hears the name "singammd" sometimes closely incorporated into
I, too, am like you." Later it seems that the singer must have intended the syntactic and narrative structure of the song, sometimes loosely in-
the words "1, too," to be attributed to Singammi herself, the "l" to refer corporated, sometimes unincorporated, and sometimes split into two
to her, whereas the cry "singamma, Singammi" is not included in the forms, one incorporated and the other not. In the in-between, loosely
refrain, incorporated areas, there occur faint, ambiguous, partial identifications
speech "I am a woqmn of perfect honor" but is simply the song's
put there to keep time. Yet this echo of identification between the singer between Singamma and others, notably her mother, her brother, and
and the heroine of the song remains strong on the level of music, that the singer. I would suggest that this way of handling the name in the
is, on the level of what one is made to feel, if not on the level of what refrain is not just the consequence of random fluctuations in parole, but
makes logical sense, throughout the performance. The words "sin- that it bears meaning. I think that it may say something about the indis-
gammi, I too," are stressed in the song; they sound as though they are tinct boundaries between Persons in Cevi's world, about the absence of
part of the same pronouncement, uttered by the same voice; and they total separation between self and the substrate of self. It gives us a view
linger that way in the mind of the hearer. from the far side of wholeness, from one who is placeless and always
Elsewhere, singammi is speaking, addressing her mother, and with remaindered.
llntouchability and the fear of death 205
204 Margaret Trawick

Conclusion Notes
1. There is an unresolved debate concerning the degree to which South In-
I have tried to show in this chapter how the Tamil untouchable singer dian untouchables accept the principle of caste hierarchy. Berreman (1971),
Cevi, not only in her choice of a song to sing but in the very way that Mencher (1974), and Gough (1973) argue that untouchables subscribe to an
she sings it, fragments visions of wholeness, which, like the one brightly essentially egalitarian ideblogy. Moffatt (1979) points-out that Tamil un-
touchables re-plicate the caste hierarchy among themselves; therefore, they
burning and unchallenged sun in the hot South Indian sky, govern the
must subscribe to its principles. In previous publications (1986, 1988), I
world Cevi lives in, but never quite completely. C€vi's share of her world have tried to show thal the verbal arf of Tamil untouchables does covertly
is part of that never-quite-completely, the in-between space of defiling challenge some of the principles of caste h,frarchy.-Many. of the songs of
though.f&file remainders, the substrate from which the soul seeking untoucfiables show strong ambivalence both toward the singers' status in
the community and toward their own status as Persons.
freedonhurd puity tries to extract itself, but never quite can. C6vi speaki 2. For want of space, I am not able to include in this chapter the details of
from that place of defilement, reminding us in her own way that our Cevi's*autobiography, which she provided some time after she performed
wholeness is partial, being bought, as it is, at her expense. the Sqfu"pf Sirigainrira for me. In inother publication (Trawick in press), I
discuHthis auiobiography, which manifeists many of the same preoccu-
But this is not the end. To do us the service of humbling us is not pations that are expressed in the song.
CCvi's only aim in singing this song, surely. She also is seeking her own J. See Moffatt (1979) and McGilvray (1983) on the relationship between Ku-
wholeness, I would surmise, for who wants to stay forever in the shit? ravars, Kuruvikkaraflka, and Palaiyars in Tamil and Sri Lankan society.
If she defies the code that excludes her from the best portion of life, this 4. Cevi,s song was recorded as part of a project on Tamil articulations of love
(anpQ and related sentiments. The project included -open-ended,
tape-
is partly, perhaps, in order that she may create a more huppy place for recbrded interviews with about 100 people in Cevi's village, together with
herself. recordings of songs sung by those same people.
C€vi also wants wholeness, we might guess, and so in this song,she The absence of any obvious figurative language in the hymn to Singamma
is notable, since most Tamil poetry, both oral and written, is florid with
tries to create it for herself. But she creates her wholeness in a diffS}ent metaphor, internal parallelism, and other forms of ornamentation.
way from the way prescribed in ancient Indian law books - not by ex- 6. For eiample, Dumont (1970), Harper (1964), Khare (1976).
cluding the other, but by embracing her. Singamma the KuSatti, abject 7. Kristeva iites as her ethnological and Indological sources Georges Bataille,
Mary Douglas, V. S. Naipaul Louis Dumont, Charles Malamoud, K. Mad-
beyond CCvi's own abjection, is this other whom Cevi embraces.
doc(, Katfrleen Gough, Murray Emeneau, S. Lindenbaum, and C6lestin
So the song is completed, not in the death and disembodiment that Bougl6. In her discussion of purity and pollution in India, she relies most
have been its themes throughout, but in the new and fully embodied heavily on Dumont (1970), Bougl6 (L969), and Douglas (1969)- -
life of this embrace. It ends with singammd resurrected and having a As a f6minist, Kristeva is in theichool of Simone de Beauvoir; her psycho-
analytic training comes from |acques Lacan; she was trained in semiotic
place, a palace, a home of her own. It ends with her standing up straight and literary studies by Roland Barthes.
and beautiful in the daylight. 9. Ashis Nandy (1983) rirakes a very similar point conceming the relation be-
To form a word around death is to get a hold on it. To cry out,,m6,, tween colonizer and colonized.
10. Much has been written about how Indian men teel (or might be surmised
is to put the lost mother back in the mouth. I think that for C€vi, to give to feel) about their mothers, but not much has been written about how
singammi a song anffo put this song on the tape recorder is rather like Indian Ttomen feel about their mothers. So it is hard to cite authorities on
this matter. Regarding the great stress in South Asia on tlre body of the
Sving her a building, giving her a shrine, giving her a sound and strong mother (or woman in general) as the origin of life and death, we may cite
body, so that she is not a corpse or a shade any longer. And singing this the work of Carstairs 095n, Kakar (1981), Yalman (1963), and Hart (1973).
song for her, making her story public, is rather like setting her free, Regarding the sense of connectedness between mother and_daughter in
letting her stand up straight again in the light and the open air, where Ind-ia, soire information is given by Beck (1982), Shulman (1980), and Tra-
everyone can see and admire her. wick (1986, 1990).
11. See Derrida (1974) on scription as spoor.

r,€E.
206 Margaret Trawick

References
Beck, Brenda . 1982. The Three Twins: The Telling of a South lndian Folk Eplc. Bloom-
Index
ington: Indiana UniversitY Press.
Berrenian, Gerald. 1971. Ttre ilrahmanical View of Caste. Conttibutions to Indian
Sociology (n. s.) 5: 1G23.
Bougl6, C6l6stin. 1969. Essai sur le regime des castes. Paris: Presses Universitaires
de France.
Carstairs, G. Morris. 1957. The Twice-Botn. London: Hogarth Press.
Derrida, Iaegues. 1974. Of Grammatology. Gayatri $pivak, trans. Baltimore: ]ohns
Hool&Ii$ Universiw Press.
DoughslMary.1959. iurity and Danger' London: Routled8^e & Kegan Paul.
Dum"ont, Louis. 1970. Hoio Hicrarciicus: An Essay on the Caste System. Chicago: J
University of Chicago Press' A'ara:'o-1 ".
and social differentiation, 153
Gough, Kathli:en. 1973.-Haiians in Thanjavu.. I" 5: Gough and H.?.. Sharma, discr-rffie practice as theraPY, 14 Aggression, 5
"eds.,
lmperialism and Reoolution in south Asia. New York Monthly Review speak6?s, 49, sz-l, s2,66 and sex differences, 81
Press, pp. 22245. Abiven, P.,'1.4'1. studies in emotion of, 78
Harper, Edviard. 1964. Ritual Pollution as an Integrator of Caste and Religion. Abu-Lughod, L., 6, 15,24-M, 65,72, Ambiguity, 75-6
lournal of Asian Studies 2:151-97. 88-9, 92, 1.10, 115-18, 127, 131, and-conflicts and disentangling, 63
Hart, Georg6. 1973. Women and the Sacred in Ancient Tamilnad. lournal of Asian L57 and disentangling,5T
Studies 32:233-50. Aesthetics of emotion talk, 51-2
Kakar, sudhir. 1981. The lnner world: A Psycho-Analytic study of Childhood and in Bhatgaon among Hindi-sPeakinP and evil eye, 100
Society in lndin. Delhi: Oxford University Press. Indians, 17,117-19 Ambivalence, 36
I(rare, Ra"vindra S. 1976. The Hindu Hearth and Home. New Delhi: Vikas Publish- Hindu, and emotion, 16-17, 107- American(s), 40, 48,53, 128
ing House. 10 emotion discourse and gender, 16,
Kristeig, lulia. 1982. Powers of Horror: An Essay on Abiection. Leon Roudiez, trans. "Rasa" as master theory of Hindu, 19,69-90
New York: Columbia University Press. 106-110 ethnopsychology of, 69-90, 1'65-6,
McGilway, Dennis. 1983. Paraiyar Drummers of sri Lanka: Consensus and Con- Tamil poetics, 18, 187, 195,201''5 t71.-3, 180-1, 184
straint in an Untouchable Caste. American Ethnologist 10:97-115. transaction, 1.02-5 American children, gender stereotyP-
Mencher, loan. 1974. The Caste system Upside Down, or the Not-so-Mysterious see also Poetics; Social aesthetics ing among, 81
East. Current Anthropology 15:469-93. Affect Amity
Moffatt, Michael. |979.'Anflntouchable Community in south lnilia: structure and among Chewong in MalaYsia, 4 emotion of, in Bhatgaon, 119'20,
Consensus. Princeton: Princeton University Press. , and cultural construction of per- t23
Nandy, Ashis. 1983. The lntimate Enemy: Loss anil Recooery of self under Colonial' son,757 Anger
l'srn. Oelhi: Oxford University Press. and emotion, 126-7 ii A'ara view (di'a tagna), 48-66
Shulman, David. 198O. Tamil Timple Myths: Sacrifice and Dioine Marriage in the and gender, 86 in Bhatgaon (gussa) ll9-23
South lndian Saioa Traditioa. Princeton: Princeton University Press' and Wolof linguistic exPression, in Engli"sh lan-guage, 48, 72, 81, 88
Trawick (Egnor), Margaret. 1982. The Changed Mother. Contributions to Asian 158 and ll"ongot "Passion" (liget), 4,
Studies 18:2445. see also Emotion 164-6
1986. Iconicity in P&iyar C.yrng Songs. In A. K. Rqmanujan and S. Black- Affective display and language lev- among Wolof,'l'46
burn, eds., Another itarmony: New Essays on the Folklore of lndin. Berkeley: els, 130 see also Women
University of California Press, pp' 29+-3M- Affective life in EuroPe, 5 Anthropology of emotion, 1-, 25, 27,
1988. Spiriti and Voices in Tamil Songs. American Ethnologist 15n9T215. Affective performance, conventional- 42, 130,162-85
1990. Desire in Kinship. In Nofes on Looe in a Tamil Family. Berkeley and Los ity of, 127 critique of, 25, 28, 40-1, 162-85
Angeles: University of California Press. ,r-" ilso Emotional enfries and psycholory, 1-4,25, 46
In prJss. wandering Lost: A Landles,q Laborer's Sense of Place and self' In Affective registers, 17, 126-59 in, 1-9
strategies
Arjun AppadurailFrank Korom, Margaret Mills, eds., Tale, Text anil Time: and cultuial construction of Per- Culture and PersonalitY;
see als6
lnierpreting South Asian Expressioe Tiditions. Philadelphia: University of son, L57 Psychiatric anthroPologY .- ^-
Pennsylvania Press. see also Registers Appadurai, A., 12, 15-17, 19, 42, 92-
Yalman, I(ur. 1963. on the Purity and sexuality of women in the Castes of AffectivitY 772,154
Ceylon and Malabar. lournal oi the Royal Anthropological Institute 93r2,L58. and llngulotlca,126'7 Asad, T,, 19, 43,78

207

You might also like