Anda di halaman 1dari 337

Publications of the Society for Psychological Anthropology

Naomi Quinn, Department of Cultural Anthropology, Duke University
Editorial board
Anne Allison, Department of Cultural Anthropology, Duke University
Daniel Fessler, Department of Sociology and Anthropology,
Hofstra University
Allen W. Johnson, Department of Anthropology, University of
California, Los Angeles
Takie Sugiyama Lebra, Department of Anthropology,
University of Hawaii
John Lucy, Committee on Human Development and Department of
Psychology, University of Chicago
Claudia Strauss, Department of Cultural Anthropology,
Duke University
Publications of the Society for Psychological Anthropology is a joint
initiative of Cambridge University Press and the Society for Psychologi-
cal Anthropology, a unit of the American Anthropological Association.
The series had been estabished to publish books in psychological anthro-
pology and related fields of cognitive anthropology, ethnopsychology,
and cultural psychology. It includes works of original theory, empirical
research, and edited collections that address current issues. The creation
of this series reflects a renewed interest among culture theorists in ideas
about the self, mind-body interaction, social cognition, mental models,
processes of cultural acquisition, motivation and agency, gender, and
1 Roy G. D'Andrade and Claudia Strauss (eds.): Human motives and
cultural models
2 Nancy Rosenberger (ed.): Japanese sense of self
3 Theodore Schwartz, Geoffrey M White and Catherine A. Lutz (eds.):
New directions in psychological anthropology
4 Barbara Diane Miller (ed.): Sex and gender hierarchies
5 Peter G. Stromberg: Language and self-transformation
6 Eleanor Hollenberg Chasdi (ed.): Culture and human development
7 Robert L. Winzeler: Latah in Southeast Asia: the history and ethnogra-
phy of a culture-bound syndrome
8 John M. Ingham: Psychological anthropology reconsidered
A cognitive theory of
cultural meaning
Claudia Strauss and Naomi Quinn
Duke University
The Pitt Building, Trumpington Street, Cambridge, United Kingdom
The Edinburgh Building, Cambridge CB2 2RU, UK
40 West 20th Street, New York, NY 10011-4211, USA
10 Stamford Road, Oakleigh, VIC 3166, Australia
Ruiz de Alarcon 13,28014 Madrid, Spain
Dock House, The Waterfront, Cape Town 8001, South Africa
Cambridge University Press 1997
This book is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without
the written permission of Cambridge University Press.
First published 1997
Reprinted 1999, 2001
Typeset in Times 10/12 pt [VN]
A catalogue record for this book is available from the British Library
Library of Congress Cataloguing in Publication data
Strauss, Claudia.
A cognitive theory of cultural meaning / Claudia Strauss and Naomi
p. cm. (Publications of the Society for Psychological
Anthropology; 9)
This work grew from a session at the 1989 meeting of the American
Anthropological Association in Washington, D.C.
Includes bibliographical references and index.
ISBN 0 521 59409 X (hardbound)
1. Ethnopsychology - Congresses. 2. Cognition and culture ~
Congresses. 3. Connectionism - Congresses. I. Quinn, Naomi.
11 American Anthropological Association. Meeting (1989:
Washington, D.C.) III. Title. IV. Series.
GN502.S77 1997
155.8'2~dc21 97-14944 CIP
ISBN 0 521 59409 X hardback
ISBN 0 521 59541 X paperback
Transferred to digital printing 2003
To my parents
Lee Strauss
Robert Strauss
To my teachers
Bea Whiting
Roy D'Andrade
List of figures page ix
List of tables x
Acknowledgments xi
Part I Background
1 Introduction 3
Meanings and culture 5
Psychology and cultural anthropology 8
Summary of the book 10
2 Anthropological resistance 12
Interpretivism: meanings are public 13
Poststructuralism and postmodernism: culture and the self are constructed 23
Historical materialism: people can resist cultural meanings 36
Cognition in practice/discourse pragmatics: meanings depend on context 42
Toward a more fruitful resolution: Bourdieu 44
3 Schema theory and connectionism 48
Introduction to connectionism 50
Connectionism for the somewhat more formally minded 60
Final comments: symbols and meanings 82
Part H Implications for a theory of culture
4 Two properties of culture 89
Durability in the individual 89
Motivational force 101
5 Three further properties of culture 111
Historical durability 111
Thematicity 118
Sharedness 122
Vl l
viii Contents
Part m Practice and possibilities
6 Research on shared task solutions 137
Analysis 1. Metaphors for marriage and what they do 140
Analysis 2. A shared schema for reasoning about marriage 160
Conclusion 187
7 Research on the psychodynamics of shared understandings 189
Analysis 3. The psychody namic basis of marital love 189
Conclusion 208
8 Research on cultural discontinuities 210
Analysis 1. How are conflicting discourses internalized? 213
Analysis 2. The disparate motivational effects of different forms of culture
learning 231
Analysis 3. Disjunctures between shared understandings and public culture 245
Conclusion 251
9 Beyond old oppositions 252
Notes 257
References 291
General index 313
Name index 319
3.1 A US American address system - classical model
(Ervin-Tripp 1969:95). page 62
Reprinted with the kind permission of Academic Press.
3.2 A US American address system - connectionist model 63
3.3 Interactive activation model (McClelland, Rumelhart,
and Hinton 1986:22). 64
Reprinted with the kind permission of MIT Press.
3.4 Recurrent network (Rumelhart 1989:154). 65
Reprinted with the kind permission of MIT Press.
3.5 Modified connectionist US American address system 71
5.1 Sally Forth comic strip, originally published April 13,1994. 114
Reprinted with special permission of King Features Syndicate.
3.1 Weights in connectionist model of US American terms of
address - First layer page 68
3.2 Weights in connectionist model of US American terms of
address - Second layer 70
8.1 Lovett's multiple voices 221
The earliest version of what was to become this book was prepared for the
invited session entitled "Assessing Developments in Anthropology" and
organized by Robert Borofsky for the 87th Annual Meeting of the Ameri-
can Anthropological Association in Washington, D.C., November, 1989.
The authors are especially indebted to Roy D'Andrade, Jane Hill, Holly
Mathews, Orin Starn, and Drew Westen for their close readings of all or
substantial portions of one draft or another, and their numerous thought-
ful suggestions. We are also grateful for helpful comments on various
drafts or portions made by Felicia Ackerman, Robert Borofsky, Donald
Brenneis, Donald Donham, Jane Fajans, Byron Good, Peter Hervik,
Jennifer Hirsch, Allen Johnson, Stanley Kurtz, Stephen Levinson, Daniel
Linger, Ernestine McHugh, Sherry Ortner, Gary Palmer, Terry Regier,
Renato Rosaldo, Adam Russell, Nestor Schmajuk, Bradd Shore, Melford
Spiro, Terence Turner, James Van Cleve, Harriet Whitehead, Beatrice
Whiting, and an anonymous reviewer for the American Ethnological
Society Monograph Series (which the manuscript then outgrew; the other
two reviewers, Hill and Shore, identified themselves to us), as well as
fruitful discussions with Dorothy Holland on related topics. In addition,
students in the authors
jointly taught class, "Culture and Cognition" at
the First International Summer Institute in Cognitive Science, SUNY-
BurTalo in July, 1994, and in Claudia Strauss's "Psychological Anthropol-
ogy," "Cognitive Anthropology," "Theories in Anthropology," and "Cul-
ture and the Brain" classes, are to be thanked for their valuable comments
on our ideas at different stages, as are the audiences at a workshop,
"Rethinking the Culture Concept," we jointly organized and chaired at the
Society for Psychological Anthropology in Montreal in October, 1993; at a
joint colloquium we gave in our own department at Duke in 1994; at
Naomi Quinn's presentation to an interdisciplinary workshop in Culture
and Cognition at the University of Michigan, Ann Arbor, in December,
1994; and at the three-day workshop she gave, one of the Bergen Work-
shops on Core Questions in Anthropology, at Bergen, Norway, in October,
xii Acknowledgments
Readers should thank Lee Strauss, as we do fervently, for editing the
entire manuscript and improving its readability and clarity immensely. We
also thank Jessica Kuper at Cambridge University Press for her commit-
ment to this book, and her role in bringing it to publication.
A much condensed version of this book's argument, "A cognitive/cul-
tural anthropology," appeared in a 1994 volume of papers, Assessing
cultural anthropology, edited by Robert Borofsky. The second half of
chapter 6 has been adapted from an article by Naomi Quinn, "Culture and
contradiction: the case of Americans reasoning about marriage," pub-
lished in Ethos in September, 1996. Some of the research reported in the
first part of chapter 8 was adapted from an article by Claudia Strauss,
"Partly fragmented, partly integrated: An anthropological examination of
'postmodern'fragmented subjects," Cultural Anthropology, 1997,12(3).
The lengthy research project reported on by Naomi Quinn in chapters 6
and 7 was made possible by National Institute of Mental Health research
grant #1 ROl MH33037O-O1, National Science Foundation research
grant #BNS-8205739, a stipend from the Institute for Advanced Study,
Princeton, New Jersey, National Science Foundation Visiting Professor-
ship for Women #RII86-20166 (hosted by the University of California,
San Diego), and grants from the Duke University Research Council and
its successor, the Duke University Arts and Sciences Research Council.
People who made the project successful are talented interviewer and
research assistant Rebecca Taylor; undergraduate interviewer Laurie
Moore: interview transcribers Phyllis Taylor, Donna Rubin, and the her-
oic Georgia Hunter; and the twenty-two husbands and wives who par-
ticipated in the long interview process and shared their unique and cre-
ative, albeit culturally informed, ways of understanding their marriages.
The first research project reported on by Claudia Strauss in Chapter 8
was assisted by a grant from the Duke University Research Council. She,
too, is grateful to the men and women who willingly shared their ideas and
life stories with a stranger.
Costs of creating original figures in chapter 3, as well as the permission
fee for reproducing the comic strip appearing in chapter 5, were kindly
borne by the Department of Cultural Anthropology, Duke
Besides being responsible for the chapters addressing their own re-
search, each co-author took primary responsibility, at a certain point in its
drafting, for different parts of this book. Claudia Strauss was the main
author of chapters 1,2, and 3, while Naomi Quinn was the main author of
chapters 4,5, and 9. This said, each of us contributed some prose to every
co-authored chapter, and there is no major idea or point in the book that
we did not discuss and debate in the course of successive revisions.
Once upon a time we anthropologists believed in the concept of culture.
Not only did we believe in it, we proselytized for it, arguing in articles,
books, and speeches that typical patterns of behavior, personality, and
belief within a society are rarely a biological inheritance (due to "race")
but are almost always socially learned ways of acting and thinking (due to
"culture"). The idea of culture was to cultural anthropology what the cell
is to biology or the unconscious is to psychoanalysis: a taken-for-granted
basic concept within the discipline and our primary intellectual contribu-
tion to broader public discourse.
We were successful in exporting the concept of culture, to both aca-
demics in other fields and the public. Within the academy there are
professors of literature doing "cultural studies," historians specializing in
"culture history," and psychologists interested in "cultural psychology."
Outside the academy it is common to hear public discussions of "multicul-
turalism" and "cultural diversity," "organizational culture," or "the cul-
ture of poverty." Psychiatrists, lawyers, doctors, and businesspeople are
learning that they will be more successful if they take "cultural" factors
into account.
Yet, surprisingly, the culture concept now sometimes finds itself more
appreciated abroad than back home. Cutting-edge theorists in other fields
invoke the culture concept, but cutting-edge cultural anthropologists criti-
cize it. While valuable, these criticisms have sometimes gone too far,
eliminating the idea of culture rather than reworking it. This book, by two
cognitive anthropologists, proposes a new analysis of culture. It argues
that in order to rethink culture, we need to understand how human beings
construct meanings.
Culture theory is at an impasse.
From the perspective of the late 1990s, descriptions of "the culture of the
X" seem old-fashioned. In part this is because we have learned how
problematic it is, in a world of shifting and multiple identities, to label any
set of people as "the X." But to a greater extent the problem lies with the
phrase "the culture of the X." In our discipline's past, such descriptions
have too often made it sound as if all of the X thought, felt, and acted the
same way, had shared this way of life for centuries and would have
continued in their traditional ways, unchanged, if colonial education and
modern mass media had not intervened. Past descriptions, too, sometimes
missed the extent to which the story they told about traditional cultural
values and practices was the interested account of one powerful class or
faction or a public, "for show," version that hid alternative accounts,
challenges to the powerful, or even mundane, widely shared practices and
understandings that contradicted informants' conscious beliefs about
what they were doing.
Yet to ignore the force of culture (in something like the old sense) is also
problematic, if culture is not limited to official representations but includes
shared understandings of all sorts as well as the publicly observable objects
and events from which these understandings are learned. Our experiences
in our own and other societies keep reminding us that some understand-
ings are widely shared among members of a social group, surprisingly
resistant to change in the thinking of individuals, broadly applicable across
different contexts of their lives, powerfully motivating sources of their
action, and remarkably stable over succeeding generations. To omit this
older view of culture from current thinking about it is to ignore the fact
that both domination and everyday practices (concerns of many current
anthropologists) rest on shared interpretive schemes, schemes learned in
ways that sometimes render them resistant to change. To leave this aspect
of culture out of consideration is also to ignore the fact that contestation
and change never arise in a cultural vacuum but always originate
4 Background
from existing conceptual systems, which can be widely shared despite
apparent political differences.
Even critics of the traditional culture concept recognize, or seem to
recognize, that we cannot dispense with it altogether. Thus James Clifford
(1986:19) states, "Culture is contested, temporal, and emergent," yet, in
the introduction to The predicament of culture (1988) notes that he is
"straining for a concept that can preserve culture's differentiating func-
tions while conceiving of collective identity as a hybrid, often discontinu-
ous inventive process," adding, "Culture is a deeply compromised idea I
cannot yet do without" (1988d:10). Lila Abu-Lughod argues that "'cul-
ture,' shadowed by coherence, timelessness, and discreteness, is the prime
anthropological tool for making 'other/ and difference . . . tends to be a
relationship of power"; therefore, "anthropologists should consider stra-
tegies for writing against culture" (1991:147; see also Abu-Lughod and
Lutz 1990). Yet Abu-Lughod observes in the same article, "Cross-cultural
work on women also made it clear that masculine and feminine did not
have, as we say, the same meaning in other cultures" (1991:140) and the
(so-called) "essential human has culturally and socially specific character-
istics" (1991:158), making it clear that culture is an idea she cannot do
without, either.
Most anthropologists today would probably agree with both sides of
this debate. Most would probably agree both that "cultures" are not
bounded, coherent, timeless systems of meanings (as we caution our
advanced students), and that human action rests on networks of often
highly stable, pervasive, and motivating assumptions that can be widely
shared within social groups while variable between them (as we teach
undergraduates in Anthropology 101).
The problem facing the discipline
is not which account is right but how to explain the fact that both are right.
How can we explain both cultural reproduction, thematicity, and force -
what Mikhail Bakhtin (1981) called the "centripetal" forces at work in
social life - and cultural variation, inconsistency, and change - Bakhtin's
"centrifugal" forces? More plainly, how do we handle the fact this is not a
homogeneous world without creating separate entities ("the culture of the
x," "the culture of the y," "the culture of the z") to explain the differences?
In this book we argue that to go beyond this impasse we need to take
another look at cultural meanings: what they are, where they come from,
and why sometimes they are motivating, sometimes not; sometimes endur-
ing (in persons and across generations), sometimes not; sometimes shared,
sometimes not; and sometimes thematically unifying, sometimes not. Since
it is the basis for everything that follows, we should address immediately
the first of these questions: what are cultural meanings?
Introduction 5
Meanings and culture
Early anthropological definitions of culture (e.g., Tylor 1958 [1871])
equated culture with socially learned ideas and behaviors. However, as Ulf
Hannerz (among others) notes, "in the recent period, culture has been
taken to be above all a matter of meaning" (1992:3).
But what is meaning?
Philosophers have long debated this, particularly with respect to the
meaning of words and sentences. Some have proposed that the meaning of
a term or sentence is its referent (i.e., the thing or situation in the world it
stands for). Others, most notably John Locke, argued that linguistic
expressions are the external, public mark of ideas in people's heads and
gain their meanings only in relation to those ideas. These two theories have
fallen out of favor in the twentieth century. Taking their place for a while
were behaviorist theories, which defined the meaning of a linguistic expres-
sion as the typical stimulus that gives rise to it and the response it evokes,
and, more recently, theories that look for the meaning of a sign in either its
pragmatic uses or its place in a larger system of signs. Of these last two
possibilities, the first (look at use) was advocated by Ludwig Wittgenstein,
the second (look at the place of a sign in a system of signs) initially by
Ferdinand de Saussure and later by structuralists such as Claude Levi-
Strauss, as well as (coming from different concerns) such philosophers as
W. V. O. Quine and Donald Davidson (Alston 1964, 1967; Tiles 1987).
More recently poststructuralists such as Jacques Derrida (1982) have
started with the structuralist approach to meaning but questioned whether
there are any stable sign systems, which leads them to the claim that
meanings are endlessly "deferred." These last three approaches are cur-
rently dominant in the philosophy of language, and we suspect that if our
fellow anthropologists considered all these possibilities, most would favor
one of the last three also.
We reject all three of the current meaning-is-use
system-of-signs, and meanings-are-endlessly-deferred approaches. The rea-
sons for this will be spelled out more in the following chapters, but can be
summarized as follows. Meaning-as-use pretends that people act without
having anything in mind. Meaning-as-emerging-from-a-system-of-signs
assigns a reality to these abstract systems that they do not have. Finally,
meanings-as-endlessly-deferred delights in the ceaseless play of signs, for-
getting that in the meantime people need some meanings to get them
through the day. Instead, our definition combines aspects of earlier behav-
iorist (meanings are defined by their stimuli and responses) and ideational
(meanings are ideas in people's heads) approaches. While behaviorist and
ideational theories have been seen as opposed in the past, we will present a
cognitive paradigm that brings them much closer together and draws on
6 Background
the strengths of each approach while avoiding some of the criticism
incurred by earlier versions of these theories.
The meaning we will give to "meaning" here is the interpretation evoked
in a person by an object or event at a given time. (Note: this includes, but is
not restricted to, word meanings.) Our notion of interpretation will be
explained in greater detail in chapter 3. For the time being, however, we
should note that a person's interpretation of an object or event includes an
identification of it and expectations regarding it, and, often, a feeling about
it and motivation to respond to it.
This definition makes meanings momentary states, as some current
theorists would argue. Unlike these theorists, however, we also stress that
these momentary states are produced through the interaction of two sorts
of relatively stable structures: intrapersonal, mental structures (which we
will also call "schemas" or "understandings" or "assumptions") and
extrapersonal, world structures. The relative stability of the world and our
schemas has the effect that both in a given person and in a group of people
who share a way of life, more or less the same meanings arise over and
over. Our definition also makes meanings psychological (they are cogni-
tive-emotional responses), but highlights the fact that meanings are the
product of current events in the public world interacting with mental
structures, which are in turn the product of previous such interactions with
the public world. We will have a great deal more to say about schemas
later; for the time being they can be roughly defined as networks of
strongly connected cognitive elements that represent the generic concepts
stored in memory (Rumelhart, Smolensky, McClelland, and Hinton
In other words, we are saying that what something (a word, an object, an
event) means to somebody depends on exactly what they are experiencing
at the moment and the interpretive framework they bring to the moment as
a result of their past experiences. A cultural meaning is the typical (fre-
quently recurring and widely shared aspects of the)
interpretation of some
type of object or event evoked in people as a result of their similar life
experiences (cf. Spiro 1987a: 163). To call it a cultural meaning is to imply
that a different interpretation would be evoked in people with different
characteristic life experiences. Talk about "the meaning" of something
tends to mystify the fact that this phrase is always a shorthand way of
referring to cultural meanings.
Nor is there an entity, "the culture of the X," which contains or causes
the meanings of the X people. As D'Andrade puts it, we have to move
beyond "the idea that culture is a thing" (1995:250). To the extent that
"culture" carries the implication that there exists some entity above and
beyond human products and learned mental structures, we agree with
Introduction 7
recent critics of the concept that it is misleading. We could keep the term
"culture," however, if we stopped thinking of cultures as independent
entities. To the extent people have recurring, common experiences - ex-
periences mediated by humanly created products and learned practices
that lead them to develop a set of similar schemas - it makes sense to say
they share a culture. As Dan Sperber notes, "There exists... no threshold,
no boundary with cultural representations on one side, and individual ones
on the other. Representations are more or less widely and lastingly distrib-
uted, and hence more or less cultural" (1985a:74). Culture, in our formula-
tion as in his, is thus not some free-floating abstract entity; rather, it
consists of regular occurrences in the humanly created world, in the
schemas people share as a result of these, and in the interactions between
these schemas and this world. When we speak of culture, then, we do so
only to summarize such regularities.
This makes "culture," as we use the term, a fuzzy concept, because we
are focusing on people's (more-or-less) shared experiences and the schemas
they acquire on the basis of those experiences. We need to put some
conditions on what sorts of shared experiences are cultural. For example,
we do not think it is useful to use "culture" to refer to shared experiences of
the natural world. But suppose we are referring to a plant, an animal, or a
landscape that has been altered through human intervention? To allow for
that we say that a schema is cultural to the extent that it is the product of
humanly mediated experiences.
Similarly, we do not want to label as cultural those schemas that are the
product of experiences arising from innately programmed behaviors. But
what about the great many experiences (some people would say this is true
of the majority of our experiences) that are the product of general innate
potentials gaining specificity in a particular human environment (Chan-
geux 1985)? A similar sort of fuzzy boundary is necessary: a schema is
cultural to the extent that it is not predetermined genetically.
Finally, an implication of our view is that cultures are not bounded and
separable. You share some experiences with people who listen to the same
music or watch the same television shows you do, other experiences with
people who do the same work you do, and still others with people who
have had formal schooling like yours, even if you live on opposite sides of
the world. This makes each person a junction point for an infinite number
of partially overlapping cultures. Some people would object, rightly, that
the context in which people experience, say, a television show, greatly
affects the particulars of the schemas they acquire from that experience.
We agree and will show later how this happens. Yet we do not want to
return to the assumption that shared cultures belong only to spatially and
temporally contiguous communities. Our proposal turns attention away
8 Background
from boundaries of this sort and focuses attention, instead, on people's
experiences, which can be partiaUy shared even if never identical, across
space and time.
Psychology and cultural anthropology
The following points sum up our initial statements and make explicit some
of the other assumptions guiding our argument:
1 We cannot explain cultural meanings unless we see them as created and
maintained in the interaction between the extrapersonal and intraper-
sonal realms. The force and stability of cultural meanings, as well as
their possibilities for variation and change, are the outcome of this
complex interaction.
2 Intrapersonal thoughts, feelings, and motives, on one side of this inter-
action, are not simply copies of extrapersonal messages and practices,
on the other side, and the dynamics of these realms are different.
3 Therefore, we need to know how the mind works in order to understand
how people appropriate their experience and act on it, sometimes to
recreate and other times to change the public social world.
4 We need to examine socialization in greater detail to learn the concrete
forms of extrapersonal culture in learners' worlds and to examine what
learners internalize at different points in their lives from experiencing
these things.
It follows that we share with many other commentators the belief that
the private/public (i.e., inner/outer or subject/world) divide, as currently
conceived, is problematic.
We differ from most, however, in exactly what
we think the problem is and what would be a better way to think about it.
We think that the inner world or psyche and the world outside of persons
are not isolated realms and that too large a gulf has been posited between
them in current theorizing. It is central to our view, however, that these
realms are different, with distinctive characteristics not found in the other.
In our view the intrapersonal and the extrapersonal realms are distinct but
closely interconnected; they are separated by a boundary, but one that is
A note on terminology: "private" and "public" are the terms usually
used to refer to the intrapersonal and extrapersonal domains, respectively.
The problem with these words is that they have been used to mark a
number of other distinctions we do not want to confuse with the intraper-
sonal/extrapersonal difference. For example, that which is "private" might
be secret or the property of one person only or restricted to the home or not
belonging to the state (as in "private property"). To avoid these connota-
Introduction 9
tions, we will usually use "intrapersonal" and "extrapersonal" or "psycho-
logical" and "social" (or, sometimes, "culture-in-persons" and "public
culture"). Another possibly confusing term is "internalization," because it
carries misleading connotations of taking in whole. We will continue to use
"internalization" but advise readers that the appropriative processes we
describe can be selective and transformative.
If the intrapersonal and extrapersonal realms function differently, as
noted in the third of our listed assumptions, then we need to know how the
intrapersonal realm works; that is, we need to know some psychology.
This point is not new to our colleagues in cognitive anthropology or the
larger field of psychological anthropology, of which it is a part.
As the
psychological anthropologist Melford Spiro has put it, "to attempt to
understand culture by ignoring the human mind is like attempting to
understand Hamlet by ignoring the Prince of Denmark" (1987a: 162).
Still, we know we will encounter strong resistance from colleagues in
cultural anthropology outside of our subfield. One anthropologist in our
department, when told the title of this book, suggested we leave out the
word "cognitive" if we wanted anybody to read it. This parallels the advice
a senior anthropologist at another university gave to an undergraduate
submitting applications to pursue graduate work in anthropology at our
department and elsewhere. Drawing on his experience of finding his pion-
eering work in culture and personality ignored in recent years, he suggested
that she describe her interests as "the self cross-culturally" or some such;
anything other than "psychological anthropology." Nowadays in cultural
anthropology it is perfectly respectable to talk about "the self," "mean-
ing," "identity," "consciousness," "subjectivity," "experience," "reader
response," "the imagined," and "agency";
strangely, however, we are
not supposed to talk about the psychological processes and structures that
help explain
these. If we ignore psychology, however, we are likely to
make false assumptions about the way selves, meanings, identities, con-
sciousness, subjectivity and experience are constructed and about the way
people respond to texts, imagine communities, and resist hegemonic struc-
Consider, for example, "identity" in its various forms (gender, sexual,
ethnic, national, etc.). Given the importance, not just in anthropology but
in the world today, of analyzing and understanding identities, it is very
unfortunate that most academic discourses on identities tend to assume
only two alternatives: Either identities are predetermined and fixed or
identities are completely constructed and fluid. The psychological models
we will present here help us to see a middle path: identity has an implicit
(normally out of awareness)
component, which is neither completely
fixed nor entirely fluid. Without such psychological models, it is all too
10 Background
easy to see either fixed physical attributes or the ever-changing immediate
context as more determining than they are and to underestimate the
out-of-awareness processes that shape conscious choices.
Antipsychologism is not universal among anthropologists outside of
psychological anthropology. Many anthropologists acknowledge the im-
portance of Freudian and post-Freudian (e.g., Lacanian) psychoanaly-
It is more unusual for anthropologists to draw on cognitive psychol-
ogy for inspiration, but Turner (1985) and Bloch (1985,1992a, 1992b) have
done just that. Bloch (1992), in fact, has argued for the importance of the
same cognitive paradigm (connectionism) that we will describe in chapter
3. Bourdieu's Outline of a theory of practice (1977), which we will discuss in
greater detail at the end of chapter 2, relies heavily on the cognitive sciences
of the 1960s and early 1970s. Among psychological anthropologists,
Bradd Shore has recently argued, much as we do here, for a "cognitive
view of culture" (1996:13) and Dan Sperber has been elaborating a cogni-
tive account of cultural meaning and (in his terms) an "epidemiology" of
the spread of mental and public representations for some time (e.g.,
Sperber 1975, 1985a, 1996). His view and ours differ primarily in the
cognitive paradigms upon which we draw. *
We were also much heartened to read the following sensible discussion
at the beginning of Hannerz's recent treatise on the forms of social organ-
ization that regulate the flows of cultural information within and between
complex societies (Cultural complexity, 1992):
As I see it here, culture has two kinds of loci, and the cultural process takes place in
their ongoing interrelations. On the one hand, culture resides in a set of public
meaningful forms, which can most often be seen or heard, or are somewhat less
frequentiy known through touch, smell, or taste, if not through some combination
of senses. On the other hand, these overt forms are only rendered meaningful
because human minds contain the instruments for their interpretation. The cultural
flow thus consists of the externalizations of meaning which individuals produce
through arrangements of overt forms, and the interpretations which individuals
made of such displays - those of others as well as their own. (1992:4)
That is exactly what we are saying. Hannerz goes on to present a sophisti-
cated analysis of some complexities in the distribution of the extrapersonal
forms. In what follows we aim to provide a similarly in-depth look at
intrapersonal culture.
Summary of the book
In the chapters that follow, we expand on these assertions and back them
up with many examples. Part I continues with chapter 2, in which we
describe and reply to four contemporary anthropological arguments
Introduction 11
against studies of internalization. At the end of that chapter we turn to
Pierre Bourdieu's practice theory. His model is similar to ours in many
respects, but we argue that bis theory could be improved if it rested on a
firmer psychological basis. Chapter 3 rounds out part I and describes the
psychological theory we use most extensively, a way of modeling cognition
called "connectionism." As we go on to demonstrate later in the book,
connectionist models alone are not sufficient to understand cultural mean-
ings, but they do provide the core of our thinking about them. Relying on
connectionist and other psychological theories, chapters 4 and 5 (part II)
outline some of the factors that explain both the centripetal and the
centrifugal tendencies of cultural meanings. Part III shifts from theory to
our research. In chapters 6 and 7 Quinn describes the research program she
has undertaken to study cultural meanings; in chapter S Strauss does the
same for her research. Finally, in chapter 9 we consider why psychological
explanations have been excluded from anthropological discussions of
cultural meanings and suggest directions for the future.
Anthropological resistance
Our insistence that anthropologists need to study internalization is not
new: The point has been made by cognitive anthropologists (e.g.,
D'Andrade 1984 and Sperber 1985a) and other psychological anthropolo-
gists (e.g., Obeyesekere 1981,1990; Shore 1991,1996; and Spiro 1987a) for
many years. Others who have made some of the same points include Bloch
(1985), Barth (1975, 1987), and Wikan (1990). Still, our concern with
psychology and our use of the term, "internalization/* is likely to meet
with resistance from many anthropologists. Before we go any further we
need to consider why, and try to disarm these objections.
We expect dissent to come from several quarters, especially from Geer-
tzian and neo-Geertzian interpretivists, Foucauldian poststructuralists
and other postmodernists, some contemporary historical materialists, and
those cognitive and linguistic anthropologists who study cognition in
practice (or discourse pragmatics). Each of these schools is associated with
a characteristic stance on psychology, meanings, and culture. Geertzian
interpretivists have stressed the publicness of meaning, cognition, and
culture. Foucauldian postmodernists have argued for the constructedness
of culture and of the self. Some contemporary historical materialists
highlight the importance of resistance to cultural meanings. Finally, many
of our colleagues in cognitive and linguistic anthropology focus on the way
thought and meaning are situated. We will present and counter these four
possible critiques of our approach, showing that there is no incompatibil-
ity between a reasonable version of those claims and ours.
While there is no logical incompatibility between our views and most of
the others we will discuss, each does lead to a different research program.
Our aim here is not to dismiss any of these approaches; each is valuable for
understanding the extrapersonal realm of culture. (Indeed, too much
anthropological debate these days seems akin to forms of urban renewal in
which whole neighborhoods are razed. The new critics seem to believe that
efforts to create better structures require complete elimination of the old.
Instead, we need a less destructive form of intervention that conserves
what is good in the old while not mindlessly preserving its blind alleys or
Anthropological resistance 13
crumbling edifices.) At the end of each section we will explain how each of
these analyses of culture is weakened by ignoring the interplay between the
extrapersonal and intrapersonal realms. Anthropological understanding
suffers if too many researchers focus on one realm to the exclusion of the
other; we need to look at both.
Interpretivisra: meanings are public
The first line of anthropological objection to meaning as psychological
states comes from interpretivism, the classic and most forceful statements
of which are Clifford Geertz's essays in The interpretation of cultures
(1973e) and Local knowledge (1983c). More than any other theorist, Geertz
(drawing on the work of his teacher, Talcott Parsons)
was responsible for
the shift we described in the first chapter, from the use of "culture" as a
loose covering term for all social learning and practice to current, narrower
equations of culture with systems of meaning (see, e.g., Geertz 1973i:5).
Although psychological anthropologists were concerned with actors' sub-
jectivities long before 1973, it is in good measure to Geertz's eloquence that
we owe a general acceptance of the importance of studying "the actor's
point of view" - not only in anthropology, but in related social sciences.
If Geertzian interpretivists and cognitive anthropologists like us are all
interested in cultural meaning, where do we disagree? The disagreement
comes from a false dichotomy that Geertz and later Geertzians established
between meaning as public (in several senses) and meaning as a cognitive-
emotional state.
Geertz's statements about the publicness of culture, meaning, and mind
(terms he used almost interchangeably) remained fairly consistent over the
years, despite other changes in his model of culture. In his early writings
Geertz retained some of Parsons's functionalism. While disputing necess-
ary compatibility between culture and social structure (1973h), Geertz
assumed that social relations are determined by the interaction of cultural,
social structural, and personality systems (1973g:92, 122, 125 and
1973h: 145-6). In this early functionalist phase Geertz stressed that sym-
bols act as models " of and "for" reality: that is, they not only express
social patterns but also motivate human action (1973g:93). From at least
"Deep play" (1973a) on, however, he rejected functionalist models of
interacting subsystems in favor of the metaphor of culture as a "text" (e.g.,
1973a:448). At the same time he began to downplay the "model for" side
of cultural symbols, their role in shaping action, leaving the "model o f
side, their purely expressive role, to dominate (see, e.g., 1973a:444). In fact,
this may not be as radical a change as it is often taken to be: Perhaps
Geertz was arguing simply for a different way of thinking about culture,
14 Background
while continuing to assume that personality and social structural systems
are as important as he ever thought they were (see, e.g., Geertz 1983b: 14;
certainly, he continued to note that symbols do shape action, e.g.,
1973a:451). Still, this shift from functionalism to interpretivism was an
important one: it paved the way for later views that cultural texts are all
there are
and was explicitly associated for Geertz with a rejection of
explanations in favor of "thick descriptions*' (1973i:14).
Yet, even as Geertz shifted from images of societies as organisms to
those of cultures as texts, he consistently advocated what he called an
"outdoor psychology" (1983d: 153). Compare, for example, the following
discussions of culture and thought, the first originally published in 1962,
the second originally published in 1966, the third and fourth originally
published in 1982:
[Cjultural resources are ingredient, not accessory, to human thought. . . human
thinking is primarily an overt act conducted in terms of the objective materials of
the common culture, and only secondarily a private matter . . . man's mental
processes indeed
take place at the scholar's desk or the football field, in the studio
or lorry^driver's seat, on the platform, the chessboard, or the judge's bench.
The view that thought does not consist of mysterious processes located in what
Gilbert Ryle has called a secret grotto in the head but of a traffic in significant
symbols - objects in experience . . . upon which men have impressed meaning -
makes of the study of culture a positive science like any other. The meanings that
symbols, the material vehicles of thought, embody are . . . as capable of being
discovered through systematic empirical investigation - especially if the people
who perceive them will cooperate a little - as the atomic weight of hydrogen or the
function of the adrenal glands. It is through culture patterns, ordered clusters of
significant symbols, that man makes sense of the events through which he lives.
(1973f: 362-3)
For symbolic action theorists (... to whom, with some reservations, I would give
my own allegiance), thinking is a matter of the intentional manipulation of cultural
forms, and outdoor activities like ploughing or peddling are as good examples of it
as closet experiences like wishing or regretting. (1983d: 151)
It is a matter of conceiving of cognition, emotion, motivation, perception, imagin-
ation, memory . . . whatever, as themselves, and directly, social affairs.
These views are most compactly (and famously) expressed in the following
formula, from "Thick description" (originally published in 1973):
Culture is public because meaning i s. . . The generalized attack on privacy theories
of meaning is, since early Husserl and late Wittgenstein, so much a part of modern
thought that it need not be developed once more here. What is necessary is to see to
it that the news of it reaches anthropology; and in particular that it is made clear
Anthropological resistance 15
that to say that culture consists of socially established structures of meaning in
terms of which people do such things as signal conspiracies and join them or
perceive insults and answer them, is no more to say that it is a psychological
phenomenon, a characteristic of someone's mind, personality, cognitive structure,
or whatever than to say that Tantrism, genetics, the progressive form of the verb,
the classification of wines, the Common Law, or the notion of "a conditional
curse"... is. (1973i:12-13)
What did Geertz mean in stressing the publicness of meaning, culture, and
thought? We find that different commentators stress different things he
might have meant. Some tell us that what he really meant was that public
symbols are observable and psychological states are not and we should
study only what we can observe; others stress that he really meant that
meanings are intersubjectively shared; another contingent takes from
Geertz the insight that meanings are socially established prior to an
individual's learning them; and still another focuses on Geertz's point that
thinking often relies on objects in the world. We agree with all of these points
and think Geertz was correct to give emphasis to them. However, none is
inconsistent with our stress on internalization and our definition of cultural
meaning as the shared cognitive-emotional state that results when the
mental structures of a group of people respond to typical objects and events
in their world. Structures of meaning can be public in all of these senses and
still be psychological states. Let us consider each of these points in turn.
Only public forms are observable and we should study only what we
can observe
Obviously, we cannot observe cognitive-emotional states directly. But
meanings, as Geertz usually used this term, cannot be directly observed
either. (Shortly we will discuss the different meanings of "meaning" for
him.) In his research and ours we start with something that is publicly
accessible (in our case, ordinary talk and mundane actions; in his case,
social dramas and other symbols) and draw inferences from these to
something that is not directly observable (in our case, to people's mental
structures and meanings; in his case, to cultural meanings that are some-
where outside of people).
Meanings are intersubjectively shared (i.e., meanings are not only
shared, but participants know that they are shared)
When Geertz referred to Wittgenstein's and other philosophers' critiques
of "private languages" he probably had in mind statements such as the
following: "if language is to be a means of communication there must be
agreement not only in definitions but al s o. . . in judgments" (Wittgenstein
1958: sec. 242). Although we think Geertz and Wittgenstein may have
16 Background
underestimated the extent of miscommunication that occurs (or the possi-
bility that communication can be pragmatically sufficient without mean-
ings being completely shared, see Wallace 1970), we do agree that there is
substantial sharing of meanings within a social group. To the extent that
people learn to interpret their worlds under similar circumstances or need
to communicate about these, they will acquire similar mental structures
and will arrive at similar meanings. (This topic is addressed thoroughly in
chapters 5 and 6.) It is a fallacy to assume that if meanings are intraper-
sonal (private in the sense of not being observable by others), they are
therefore idiosyncratic (private in the sense of not being held in common).
Meanings are established prior to the individual's learning of them
In Geertz's words, "culture . . . [is] an historically transmitted pattern of
meanings embodied in symbols, a system of inherited conceptions . . . "
(1973g:89), and "From the point of view of any particular individual, such
symbols are largely given" (1973d:45). We agree completely that individ-
uals do not discover or make up cultural meanings on their own. Not only
does most of the publicly observable world pre-exist us, but the mental
structures by which we interpret that world are developed through explicit
teaching and implicit observation of others' talk, actions, and material
products. (Chapters 4 and 5 in this volume address how this learning takes
place, with chapter 5 stressing, in particular, the factors that lead to the
recreation of mental structures in new generations of learners.) Meanings
are both psychological states and social constructions - although, as we
will show, the process of social construction leads to cognitive results that
are not apparent from study of symbols alone.
Thought often relies on material objects
As anthropologists we certainly agree that thought takes place not only in
the psychologist's laboratory or at the philosopher's desk, but also in the
kitchen, the playground, the factory, and the store. In fact, the way people
think in such settings, making use of objects at hand and combining
information contributed by different participants in face-to-face groups, is
currently being addressed by a number of cognitive anthropologists.
does not follow, however, from the fact that most thought takes place in
such settings, using props, and advancing through interactions with other
people, that thinking itself is occurring in some nebulous space outside of
people's bodies. We find the proposal that thought (including interpreta-
tion) occurs in the air between people more mysterious than the assump-
tion that it occurs in the brain and other parts of the nervous system. Our
commonsensical point that thought is both private (in the sense of taking
Anthropological resistance 17
place in our bodies) and public (in the sense of occurring while we are out
in the world, engaged in activity) could be illustrated by an image from the
children's book, Inside, outside, upside down (Berenstain and Berenstain
1968). In that book a little bear crawls into a box, which is then turned
upside down and carried outside. On the last page he tells his mother,
"Mama! Mama! I went to town. Inside, Outside, Upside down!" Similarly,
our understandings are always "going to town" both inside and outside
(and sometimes upside down!).
In sum, meaning can be public in each of the ways Geertz stressed and
still be a psychological state. Next we will show that meaning has to be a
psychological state; there is no other sensible way to conceive of meaning
and still have it play a role in social action.
Three nonpsychological ways to conceive of meaning
Over the course of his career Geertz used "meaning" in three different
ways. None is the psychological definition we propose, and each sense of
meaning is problematic.
A Meanings Are Uses. This was Ludwig Wittgenstein's approach. Wit-
tgenstein argued that "the meaning of a word is its use in the language"
(1958:sec. 43). His arguments on this point correctly and importantly stress
(among other things) the way the meaning of a term (or any sign) will vary
depending on the "language game" in which it is embedded, and the
impossibility of assigning most words clear-cut boundaries of their ap-
plicability (e.g., 1958:secs. 21ff and 66ff). (The fuzziness and context-
variability of meanings will be discussed further in this chapter and in
chapter 3.)
This meaning of "meaning" has the virtue of concreteness: the meaning
of an expression is not an abstract definition or interpretation that resides
nowhere in particular but is simply its use. But this is not the way most
anthropologists would think about it: most of us would say that people use
words or other signifiers in a certain way because of the cultural meanings
of these things. Meanings lead to uses; if that is not the case, then people
are just mechanically acting and meanings are only a way for outsiders to
describe the pattern of others' behaviors.
Indeed, Geertz is typical of most anthropologists in this regard (and
seemingly unaware of Wittgenstein's dismissal of this common view, e.g.,
1958:sec. 146rT and p. 216). On the one hand, Geertz (1973i:17 and
1973f:405n) invokes Wittgenstein and occasionally seems to identify cul-
ture with behaviors (e.g., 1973i:10). But for the most part Geertz is very
careful to distinguish between behaviors and the meanings they impart:
that is the whole point of his repetition of Gilbert Ryle's discussion of the
difference between a wink and a twitch (19731:6-7). In "Thick description"
18 Background
(1973i), Geertz notes that "culture consists of socially established struc-
tures of meaning in terms of which people do such things as signal
conspiracies and join them or perceive insults and answer them"
(1973i: 12-13). Note that he says, "w terms of which people do such things
...," in other words, structures of meanings are prior to the actions and are
part of an explanation of why people act as they do. This implies that
meaning is something added to bare actions; it is not simply use.
In this respect, we agree with Geertz rather than Wittgenstein: we, too,
believe that meanings lead to actions, and are not simply a way of descri-
bing them. But Geertz did not go on to think through, consistently, the
implications of this position. If meanings influence action, they have to be
someplace. Where are they? Geertz gave two answers to this question. In
some passages he says that meanings are "in" public objects and behav-
iors; and in some passages he says or implies that meanings are abstract
structures that are no place in particular.
2. Meanings are stored in symbols. Geertz used that very phrase in the
essay, "Ethos, world view, and the analysis of sacred symbols':
"[M]eanings can only be 'stored'" in symbols: a cross, a crescent, or a
feathered serpent (1973b: 127). In "Religion as a cultural system" he used a
metaphor of symbols as a "vehicle" that carries meaning:
[A symbol is] any object, act, event, quality, or relation which serves as a vehicle for
a conception - the conception is the symbol's 'meaning'... (1973g:91)
The problem with these images of symbols as storing meanings or
carrying them along for the ride, is - as Linger (1994) and Langacker
(1991:508) show very clearly - that they all employ a misleading "conduit
metaphor" (Reddy 1979) to describe the relation of words and other
symbols to meanings. It takes only a moment of thought to see that
meanings cannot literally be found in symbols. Consider, for example,
cultural meanings of money. However hard we look, we will not find these
ideas hidden somewhere in slips of paper, bits of metal, or the physical acts
by which these bits of paper, metal, and their representatives are distrib-
uted. (It follows that meanings cannot be found - as more recent struc-
turalist and poststructuralist theories would have it - in fixed or shifting
systems of symbols either.) Geertz probably realized this, and his unease
with this misleading way of speaking is doubtless what led him to add the
scare quotes around "stored" in the first quote in this section.
But if meanings do not literally rest in symbols and are not identical with
symbolic actions, where are they?
3. Meanings Are Abstract Structures. A third possible answer that could
be given to this question, and one that Geertz gave as well, is that meanings
are nowhere in particular because they are abstract entities:
Anthropological resistance 19
If, leaving our winks and sheep behind for the moment, we take, say, a Beethoven
quartet as an, admittedly rather special but, for these purposes, nicely illustrative,
sample of culture, no one would, I think, identify it with its score, with the skills
and knowledge needed to play it, with the understanding of it possessed by its
performers or auditors, nor, to take care, en passant, of the reductionists and
reifiers, with a particular performance of it . . . that a Beethoven quartet is a
temporally developed tonal structure... is a proposition to which most people are,
upon reflection, likely to assent. (1973i:l 1-12)
And though Geertz notes in the middle of the preceding passage that
neither is culture "some mysterious entity transcending material exist-
ence," he gives as further examples of culture, "Tantrism, genetics, the
progressive form of the verb, the classification of wines, the Common Law,
or the notion of 'a conditional curse'" (1973i:13). These are all abstract
things. Quartets, Tantrism, and so on have concrete expressions, to be
sure, but in these passages he claims that concrete expressions are not
themselves structures of meaning.
But meanings cannot be abstract structures that are nowhere in particu-
lar (or in a cloud hovering over Cincinnati, as a colleague of ours put it
As we have already noted, if they are nowhere in particu-
lar, how can they ever come to motivate action? A reasonable slogan here
would be, "Abstract entities cannot have concrete effects." And Geertz
probably realized this, too, as the following passage indicates. Note,
however, that even in this passage, which describes the "ontological
status" of cultural actions as "things of this world," he also says culture is
"unphysical" - slipping back into the idea that culture is an abstract
Culture, this acted document, thus is public, like a burlesqued wink or a mock
sheep raid. Though ideational, it does not exist in someone's head; though unphysi-
cal, it is not an occult entity... The thing to ask about a burlesqued wink or a mock
sheep raid is not what their ontological status is. It is the same as that of rocks on
the one hand and dreams on the other - they are things of this world. (1973i:10;
emphasis ours)
If culture (a pattern of meaning) is "unphysical," how can it have the same
ontological status as a rock or a mock sheep raid?
Meanings have to be psychological
Geertz's inconsistency
- indeed, the inconsistencies and unease of many
anthropological expositors of meaning
- is the result of his and their
avoidance of the only position that makes sense: Meanings of things have
to be in people's minds (and "hearts"; cf. Geertz 1973i:ll). There is no
20 Background
other place for meanings to be concretely, and they have to be concrete if
they make a difference in the world.
This can be illustrated with our example of cultural meanings of money.
We have already pointed out the absurdity of saying these meanings are
literally embodied in coins, bills, checks, and credit cards (Meanings are
stored in symbols, definition 2). We could speak of these meanings as being
an abstract system or structure (definition 3), but that would be to confuse
the outsider's descriptive abstractions with insiders' internalized under-
standings about money and the concrete practices and objects created by
those understandings (cf. Bourdieu's critique of objectivism, 1977:5). The
most tempting view might be the first: meaning is use. Again, however, this
confuses outsiders' and insiders* perspectives. Certainly an outside ob-
server of the meanings of money for late twentieth century North Ameri-
cans can only ascertain those meanings by observing people's uses of
money, but for the people whose uses are being observed, each monetary
transaction provokes meanings in them, and it is on the basis of these
meanings that they act. For example, someone deciding to buy a lottery
ticket does so because of what a sudden windfall of money would mean to
them. These meanings are a combination of ideas (e.g., about the "good
life"), feelings (e.g., of relief at being free from debt), and motivations (e.g.,
to win admiration through generous charitable donations) in them.
True, these ideas, feelings, and motivations can only be observed from
things people say and do publicly; they are probably held in common with
many other people; they are learned through participation in social institu-
tions and reinforced by being associated with symbols of various sorts
(e.g., calendrical and life-cycle rituals, everyday talk, mass media); and
they are enacted with concrete objects in everyday settings and can be
improvised upon in these settings. But the point remains that these mean-
ings are the actor's meanings: They are the actors' thoughts, feelings, and
motivations, including out-of-awareness psychological states. As others
have insisted before us, meanings can only be evoked in a person.
Implications for current symbolic research
Geertz's denigration of psychological analyses has fettered later theorists
trying to deal with a central dilemma contributing to the current impasse in
culture theory: how do we explain the force of culture (as both symbols
and meanings) while acknowledging that culture (in whatever concrete
forms it takes) does not make anyone do anything? This handicap is
obvious in Ortner's (1989,1990) recent thoughtful attempts to resolve this
dilemma along neo-Geertzian lines.
Ortner rejects both a "hard/internal" position that socialization leads
actors to '"do the cultural thing' under most circumstances" (1990:86) and
Anthropological resistance 21
a "soft/external" position that people generally act on the basis of univer-
sally understandable rational self-interest. Her compromise "internal/ex-
ternar view is that
not all of a culture's repertoire of symbolic frames make sense to all actors at all
times... Yet at moments in the course of events the story seems to make sense of a
person's circumstances and is thus appropriated and internalized .. . Yet the . . .
loose fit, between the structure of the self and the structure of cultural models
means that a cultural frame that has been taken into the self can also be put out
again... (1990:89)
Ortner's analysis makes an important contribution in highlighting the role
of schemas that seem to be highly thematized across cultural domains and
highly persistent across time in a society - in her case, a cultural "schema"
(her word)
for fraternal and similar rivalries among Sherpa men that is
embedded in Sherpa stories and rituals and motivates much Sherpa inter-
pretation and action. The problem is that without a theory of the complex
ways symbolic frames shape people's motives and understandings - of
what it means for something to be "appropriated and internalized" -
Ortner cannot explain why some schemas are more motivating than
others, why some are more thematized culturally than others, and why
some are more historically persistent than others. She attributes historical
persistence to the " 'freezing' of a variety of cultural practices into a
particular narrative shape" (1990:8), an explanation that begs the question
of why narratives are replicated. Instead of assuming that certain narra-
tives will continue to be told, making the schemas they express continuous-
ly available, she might have asked why the values that are expressed in
Sherpa stories about rivalry have remained compelling for Sherpa, leading
them to repeat the narratives generation after generation.
In her book-length treatment of the Sherpa case (1989), while adhering
to the same argument about the relation between culture and the individ-
ual actor (see 1989:126-129 and 198), Ortner does provide an explana-
tion for the sibling-rivalry schema. She (ibid.:33) suggests it emerges from
"the contradictory nature of the brother relationship," which is "at once
hierarchical (in the natural superiority and authority of elder over
younger) and egalitarian (in the rule of equal inheritance)." She (ibid.:34)
adds that "[A] major dimension of the Sherpa ethic of egalitarianism and
achievement is a strong streak of competitiveness in the culture. Given
the theoretical equality of all men, the fact that in reality some do better
than others is a source of resentment and renewed striving." But this
explanation seems to be missing something. For one thing, there are
many societies, including our own, in which the natural authority of
older siblings over younger coexists with equal inheritance (and, as in our
society, with competitiveness). For another, there is something more than
22 Background
merely structural about the Sherpa preoccupation with sibling rivalry in
particular and with rivalry more generally - in the "elaborate vocabulary
pertaining to competition, envy, rivalry and so forth" that Ortner
(ibid.:34) reports, in the preoccupation with competitive feelings and re-
lations that she describes in Sherpa talk (ibid.:35), in the common Sherpa
introspection about "not being able to stand seeing others get ahead"
(ibid.). Something, dare we say, psychological.
Pursuit of this line of thinking might have led Ortner to try to specify the
dynamics behind the contradiction she identifies between hierarchy and
equality. After all, we might reasonably expect a structural contradiction
to translate into significant and sustained felt experience in order to have
psychological consequences like the Sherpa preoccupation with sibling
rivalry she describes. Suggestively, she comments in one place (1989:173)
that, though inheritance is supposed to be equal, "in practice things do not
always work out so evenly/* and goes on to describe how eldest sons are
assured of respect and youngest sons of secure property, including the
parental house, while middle sons* prospects for either status or property
are not so secure. The youngest son, she (ibid.) explains, "is often the
sentimental favorite of his parents.'* In trying to understand the rivalry
brothers feel towards one another, can we ignore the way parents appear
to discriminate among sons?
An altogether different possibility, unmen-
tioned by Ortner - but raised by Obeyesekere (1990:117-118) in the
different South Asian context of Hindu Indian joint households - is that, in
the often coerced arranged Sherpa marriages Ortner describes, in-marry-
ing "stranger'* wives are moved to foster particularly intense symbiotic
relationships with their first sons, thus setting up second and later sons for
feelings of jealousy. Our more general point is that explaining Sherpa
sibling rivalry requires a theory, whatever that theory proves to be, that
addresses how people are socialized, the sorts of understandings and needs
they gain as a result of this socialization, and the implications of this
process for the way they go on to socialize their children.
This, however, would take us deep into the "secret grotto" of psychol-
ogy, entrance into which Geertz has proscribed (1973f:362).
ly, Ortner frames her 1990 paper in the terms provided in Geertz's early
"Religion as a cultural system":
The lines are being drawn in the debate over the role of culture in history. On one
side there is a set of authors denying culture anything other than a minor represen-
tational role. For them, culture operates largely as a set of markers of social
phenomena (particularly as markers of group identity) but rarely as models for
social phenomena, shapers of the social and historical process. On the other side
there is a set of authors insisting that culture, in the form of complex templates/or
thought, feeling, and action, plays a strong role not simply in representing the
Anthropological resistance 23
world but also in shaping its ongoing historical emergence. (1990:57; emphasis in
But her attempts to reclaim the insights of the earlier Geertz are hobbled
by the limits, imposed by the later Geertz, of what is sayable and thinkable.
The following lines are telling:
One may even say that schemas are "in people's heads." Yet ultimately they get
"into people's heads" and become part of people's repertoires for ordering, experi-
encing, and acting on the world, as a result of their enactment in real lived forms -
in stories that are told, rituals that are performed, kinship and political relations
that are practiced. (1990:77)
Why the scare quotes around "in peopled heads"? Probably because
Ortner has not freed herself from the later Geertzian interpretivist dictum
that cultural meanings are public - not only learned from public life, but
strictly extrapersonal, outside of psyches. She starts from the assumption
that culture is outside of persons, then has the difficulty of grappling with
the fact that it is clearly inside persons as well.
In sum, symbolic analysis needs to be supplemented by research that
takes note of which symbols people attend to and which they ignore;
varying ways a given set of symbols is interpreted by different social
subgroups and over time (Roseberry 1989); and the socializing situations
that lead some symbols to be motivating while others make no difference in
what people do (Spiro 1987a).
The simplistic model that publicly access-
ible symbols straightforwardly determine people's understandings could
be called a "fax model" of internalization (Strauss 1992a): According to
the fax model, extrapersonal messages are simply reproduced in people's
Ortner's "loose fit" or "loosely structured relation" (1989:198)
between extrapersonal cultural messages and individual actors' psycho-
logical states does not radically revise the fax model; it only introduces
some noise into the system. The inadequacy of a fax model will be illus-
trated with examples from our own work in chapters 6, 7, and 8. Before
that, however, we have a lot of other ground to cover, including a more
recent version of the view that cultural meanings are public.
Poststructuralism and postmodernism: culture and the self are
A new generation of anthropologists and allies from literary criticism,
history and other disciplines
share with the later Geertz and other
symbolic anthropologists the method of analyzing the meanings of public
"texts" and "performances" (e.g., movies, myths, architecture, and cloth-
ing) but reject what they see as the earlier generation's reification of culture
24 Background
and society. For this newer generation the concept of an internalized
culture is suspect because it seems to imply static, tradition-bound actors
who are out of place in the world today, if indeed they ever existed. This
newer generation stresses that culture and the self are constructs. Theories
that, like ours (or even Geertz's), speak of signs referring to something
"deeper" (in our case, people's schemas; in Geertz's, a cultural ethos) are
suspect because they seem to want to fix meanings and persons too much
(cf. Jameson 1984).
Culture is invented
There are two schools of research that could be summed up by the slogan,
Culture is invented. Some stress the constructedness of anthropologists'
(and other scholars') descriptions of culture. Others show that very often
cultural identities and "traditions" are inventions of the participants for
political or material advantage. We agree with these points but do not see
how they threaten the concept of culture or cultural meanings that we have
proposed here. In fact, our pared-down definition of culture as simply
regularities in the world of public objects and practices as well as more-or-
less shared understandings learned from this public world fits very well
with some postmodern critiques; we share their suspicion of culture as a
superorganic, cohesive, bounded, timeless entity.
Indeed, we have prac-
ticed for many years forms of ethnographic writing that present individual
voices rather than descriptions of abstract systems (e.g., Quinn 1982).
Cultural descriptions are constructions. Following Michel Foucault's
discussions of the forces shaping social scientific descriptions (e.g, 1977; see
also Geertz 1973i:15), a number of analysts have argued persuasively that
older descriptions of cultures overemphasized the stability, harmony, isola-
tion, and uniformity of non-European peoples (Asad 1973; Clifford 1986c;
Diamond 1974; Said 1978). According to these critics, whatever the moti-
vation for earlier cultural research, it tended to support Western European
imperial projects by painting images of people who were stuck in the past
and needed outside assistance to become "unstuck" (Rosaldo 1986).
Yet, as James Clifford argues in a critique of Edward Said's Orientalism,
while we should reject past descriptions of culture as static coherent wholes,
it does not follow that we should ban cultural descriptions. If we are going to
throw out one set of descriptions, they should be replaced by others that still
allow us to talk about "collectively constituted difference" (Clifford
It is true that all representations are partial (we can never tell
the whole truth), but it does not follow that all representations are false (that
we are not telling part of the truth). And except for the lunatic fringe,
2 3
postmodernists would not deny that there are some regularities in people's
Anthropological resistance 25
practices, assumptions, and the institutions creating and created by these - if
nowhere else, then in the regularities that produce hegemonic discourses (as
Foucault's own work is dedicated to showing).
Claimed identities are constructions. Closely related to the critique of past
descriptions of culture are claims that "culture is invented" by players in the
social game themselves, with identities put on or discarded, practices
adopted or rejected, as needed to advance their own interests vis-a-vis other
groups. Some of the authors advancing this argument start with Marxian
rather than Foucauldian premises (and thus focus on social movements
rather than intellectual discourses about movements), but all are post-
modern in their stress on the interconnectedness of peoples and fluidity of
identities. These authors would agree with Geertz, for example, that cultural
meanings are socially established but would add that meanings are not
established for all time but are constantly up for grabs as groups negotiate
differences in their practices. In a well-known case study, Clifford presents
this point of view clearly for the Mashpee Indians:
Metaphors of continuity and "survival" do not account for complex historical
processes of appropriation, compromise, subversion, masking, invention, and
revival. These processes inform the activity of a people not living alone but
"reckoning itself among the nations." The Indians at Mashpee made and remade
themselves through specific alliances, negotiations, and struggles. It is just as
problematic to say that their way of life "survived" as to say that it "died" and was
"reborn." (1988a:338-9)
He makes the same point more generally for everyone's culture: "Culture
is contested, temporal, and emergent" (1986:19). Eric Wolf made similar
claims earlier:
In the rough-and-tumble of social interaction, groups are known to exploit the
ambiguities of inherited forms, to impart new evaluations or valences to them, to
borrow forms more expressive of their interests, or to create wholly new forms to
answer to changed circumstances. (1982:387)
We agree that groups do sometimes negotiate, exploit, reinterpret, borrow,
and create cultural forms; they can also "invent traditions" (Hobsbawm
and Ranger 1983) and "imagine communities" (Anderson 1983). How-
ever, that they do so does not rule out internalized schemas. Even when
intent on reinventing themselves, people do not pluck new cultural forms
out of the air; their imaginings and ^interpretations always rely on under-
standings learned and imbued with motivation. Culturally variable, inter-
nalized schemas shape both the ways people define what is in their self
interest and the means they use to obtain those goals. This does not assume
bounded cultural systems; learning can take place across national or ethnic
borders as well as within them. But new forms are still always incorpor-
26 Background
ated, rejected, and remade in terms of previous schemas (as the movie
Trobriand cricket makes clear; for other good examples see Bohannan 1966
and Hannerz 1992:238.) By focusing exclusively on conscious goal-striving
theorists ignore the complexities of internalized beliefs and underestimate,
especially, the force of out-of-awareness knowledge and feelings. How
many of us vowed as adolescents that we would create new selves, only to
find, as adults, that we are much more like our parents or peers than we
had planned to be?
Indeed, Clifford serves as his own best critic here:
Perhaps this book goes too far in its concern for ethnographic presents-becoming-
futures . . . its Western assumption that assertions of "tradition" are always
responses to the new (that there is no real recurrence in history) may exclude local
narratives of cultural continuity and recovery. I do not tell all possible stories.
Still, this set of critiques is right in pointing out that some accounts of
culture have relied on the implicit assumption that we do not have to
account for why people act in the culturally variable ways they do: People
do "the cultural thing" because "their culture taught them to do it" (Fox
1985; Hannerz 1992). The notion of culture we favor is not subject to this
criticism because in our sense, culture is not a being above and beyond
people's schemas, practices, and the material causes and results of these
things. In our framework it makes no sense to say "culture taught." Rather
(as we will explain in chapters 4 and 5), people teach - both explicitly,
through their words and consciously chosen actions, and implicitly, by
their example - and the socially constructed world teaches as well, in the
sense that by living in it, people acquire certain implicit understandings.
The cultural understandings that are the result of these processes are often
continuous historically but not necessarily so. New understandings can
arise and can then serve as the basis for the production of new cultural
The self is constructed
A much more radical critique of our approach comes from Foucauldian
poststructuralists who argue that any talk of internalization relies on a
particular notion of self- a being with inner states and processes - that is a
flawed, ethnocentric construct and ignores the politics of identity construc-
tion (Gergen 1990). As Judith Butler puts it:
"Pinner'* and "outer" make sense only with reference to a mediating boundary
that strives for stability. And this stability, this coherence is determined in large
part by cultural orders that sanction the subject and compel its differentiation from
Anthropological resistance 27
the abject (sic). Hence, "inner" and "outer" constitute a binary distinction that
stabilizes and consolidates the coherent subject. When that subject is challenged,
the meaning and necessity of the terms are subject to displacement. If the "inner
world" no longer designates a topos, then the internal fixity of the self and, indeed,
the internal locale of gender identity, become similarly suspect. The critical ques-
tion is not how did that identity become internalized? as if internalization were a
process or a mechanism that might be descriptively reconstructed. Rather, the
question is: From what strategic position in public discourse and for what reasons
has the trope of interiority and the disjunctive binary of inner/outer taken hold?
(1990:134; emphasis in the original)
Butler recommends instead that we look at identity as a performance.
Gender, for example, is not a natural part of self but established through
repeated performance of "signifying gestures" constrained by regulating
discourses (in Foucault's sense) (Butler 1990:x; 1993:1).
Butler draws her critique of internalization, and of person-centered
studies, in part from Foucault (e.g., 1977). Foucault does not proscribe
references to empirical individuals: "Of course, it would be ridiculous to
deny the existence of individuals who write, and invent" (1972b:222).
However, he writes about the "construction of subjects" and the creation of
a certain kind of individuality during the Enlightenment in Western Europe:
[Tjhe subject who knows, the objects to be known and the modalities of knowledge
must be regarded as so many effects of these fundamental implications of power-
knowledge and their historical transformations... To analyse the political invest-
ment of the body and the microphysics of power presupposes, therefore... that -
where knowledge is concerned - one abandons . . . the primacy of the subject.
To sum up, it might be said that discipline creates out of the bodies it controls four
types of individuality, or rather an individuality that is endowed with four character-
istics: it is cellular [a separated unit]..., it is organic [endowed with natural functions]
. . . , it is genetic [continuously and progressively developing]..., it is combinatory
[capable of being combined efficiently with other individuals in larger groupings],
(1977:167; bracketed insertions our paraphrases of Foucault's descriptions)
In The history of sexuality, volume I (1978) Foucault critiques, especially,
the idea of the coherent subject: at any given time there are multiple social
discourses "furrowing across individuals themselves, cutting them up and
remolding them, marking off irreducible regions in them, in their bodies
and minds" (1978:96).
Implicit in this is that discourses do not represent
realities: they create them.
What is valuable in Foucault's work, and work (like Butler's) that he
inspired, is the constant attention they draw to the political context of
psychological development (to put their views into our terms - more on
this shortly). Butler, for example, problematizes the prohibition of same-
sex desire that is so important in Freud's theory of the development of
28 Background
gender identity through resolution of the Oedipus Complex for boys and
Electra Complex for girls (Butler 1990:57-72). Foucault forces us to pay
attention to the mutual influence of social scientific descriptions and
military, economic, pedagogical, reformatory, and other "disciplining"
institutions, as well as the implications of all of these techniques and ideas
for everyday modern life in Western Europe and the United States. Both
writers have absorbed, and help demonstrate, the general post-Boasian
cultural anthropological premise that much of what has been considered a
biological inheritance is actually socially learned.
Butler's (1990:128-141) emphasis, furthermore, on the "performative"
aspect of sexual and gender identities, which implies the possibility of
creating intermediate genders and sexualities or moving between them at
different times, has been extremely helpful for people who reject self-
identification as all masculine, all feminine, all straight, or all gay. By
extension, her theory has enormous practical value for other sorts of
identities as well.
Yet, as valuable as this de-essentializing move is for practice, it rests on
very odd premises. If Foucauldians like Butler were saying only that we have
to pay attention to the political context of psyche formation or identity
performances, there would be no conflict between their views and ours.
However, their view is much stronger than that: they would contend that in
finding this area of agreement we have misrepresented their views because
they do not see a person as something separate that is acted upon by social
discourses but rather as a creation, construct, or "effect" of social dis-
courses. In their terms society is not the context within which people
develop. There is no real boundary between people (and their inner work-
ings) and the social world outside them; any attempt to draw such a
boundary itself serves as a means of trying to create normalized, disciplined
selves. Butler claims, furthermore, that such boundaries are not drawn in
societies or historical periods outside the modern West: "mundane oper-
ations of ordinary language - widely documented within anthropol-
ogy. . .regard the subject/object dichotomy as a strange and contingent, if not
violent, philosophical imposition" (1990:144; see also 1990:10). This is
consistent with Foucault's claims that not only were new kinds of subjects,
but also new kinds of ideas about individuals, created in the modern West.
It is helpful for anthropologists to remember, when we suffer crises of
confidence about the value of our studies of other societies, how much this
research is relied upon by thinkers in other fields. However, here we have
to depart from the agreeable, both-sides-are-right stance we have taken so
far in this chapter. This denial of the difference between the inner world of
subjects and the outer world of objects goes too far; furthermore, the
anthropological record does not support the idea that most people in the
Anthropological resistance 29
world see no difference between persons (and their inner thoughts, feelings,
and motives) and the world outside them. In fact, denial of this difference is
the more "strange and contingent'* philosophy. Note that the following
discussion applies primarily to Butler's 1990 book, Gender trouble, rather
than her 1993 book, Bodies that matter, which we will consider as it
becomes relevant later.
Are self/other and inner/outer distinctions not made in other
Although Butler does not cite any of the "widely documented" anthropol-
ogical work to which she refers,
it is true that a number of anthropolo-
gists have stated that the "self is conceived of as more autonomous from
other people and outside influences in the modern West than in other times
and places. The best known of such statements is Geertz's:
The Western conception of the person as a bounded, unique, more or less integ-
rated motivational and cognitive universe, a dynamic center of awareness, emo-
tion, judgment, and action organized into a distinctive whole and set contrastively
both against other such wholes and against its social and natural background, is,
however incorrigible it may seem to us, a rather peculiar idea within the context of
the world's cultures. (1983a:59)
Catherine Lutz (1988) has contributed an influential, Foucauldian dis-
cussion of emotions on the Micronesian island of Ifaluk:
As Clifford has noted of culture itself, emotion is "contested, temporal, and
emergent" (1986:19). Once de-essentialized, emotion can be viewed as a cultural
and interpersonal process of naming, justifying, and persuading by people in
relationship to each other. Emotional meaning is then a social rather than an
individual achievement - an emergent product of social life. (1988:5)
Daily conversations on Ifaluk are pervaded by the assumption that people are
oriented primarily toward each other rather than toward an inner world of individ-
ually constituted goals and thoughts. (1988:81)
Yet, none of these authors shows that, in the societies where they did their
research, the average person sees no boundary between themselves and the
world or between the inner world of thoughts and feelings and the outer
one of objects and others' actions. In fact, Lutz is very careful to say the
Other cultural systems that cannot be characterized as having an ideology of
individualism, such as that of the Ifaluk, also obviously talk about the emotions as
located to some degree "inside" the person. It is clear that emotion words are
everywhere used to talk about the relationship between the self and the world.
30 Background
Her data show this. On the one hand, Lutz found it was a faux pas in Ifaluk
for her to say, "Do you want to come with me to get drinking water?"; she
should have said "We'll go get water now, O.K.?" (1988:88). Her study
does show that emotion statements are used by Ifaluk Islanders to make
claims on each other and in Ifaluk it is assumed that one person's emotions
are generated in response to another's. On the other hand, the Ifaluk use
the term niferash, which she translates as "our insides" and is etymologi-
cally related to feral ("the center vein of a tree, through which sap rises"
1988:235,fn. 6), to refer to thought-feeling states (1988:92). A less com-
monly used term is nifitigosh, a literal translation of which is "inside our
flesh" (1988:235, fn. 6). They also talk about human persons (yaremaf) and
individual will/emotion/desire {tip-) (1988:87 and 93).
While Lutz emphasizes interpersonal emotion discourse but does not
rule out inner emotional experiences, Dorinne Kondo goes further. Draw-
ing on Butler's work, among others, Kondo challenges the "conventional
trope [that] opposes 'the self as a bounded essence, filled with 'real
feelings' and identity, to a 'world' or to a 'society' which is spatially and
ontologically distinct from the sel f (1990:33-4). She takes even Lutz to
task for implying that the Ifaluk have selves who use words in certain
typical ways (Kondo 1990:40-1). Kondo's own Japanese study, however,
goes on to do the same sorts of linguistic analyses and shows that while her
neighbors and coworkers in Tokyo assumed people to be highly interde-
pendent, they still could distinguish self from other and inner thought-
feelings from the outer world. Consider the following summary of the
lectures she and some coworkers heard at a company-sponsored outing to
an "ethics center":
The most relevant term in our discussion of the center's theories of selfhood . . . is
kokoro, the heart, the seat of feeling and thought. To improve and polish the
kokoro was our goal. The kokoro partakes of the energies of ki [breath or vital
force]. These emotions and energies cannot be left on their own, to focus on
themselves, lest the kokoro become intent on the expression of its own selfish
desires, with no thought for others... ideally the desires of the kokoro and those of
society should run parallel. The moral force of the ideal of sunao na kokoro emerges
still more clearly in contrast to its opposite: hinekureta
i.e., twisted, eccentric,
crochety, perverse, and prejudiced. These are the characteristics of someone who is
self-indulgently anti-social, who allows egocentric quirks to disturb smooth social
relations. Were people made aware of their social connectedness, they would also
realize the inappropriateness of such selfish behaviors. (1990:104-5)
Far from assuming the absence of a self/society distinction, these lectures
imply that it takes great effort and discipline to bring self and society into
harmony, effort that depends on training thoughts and feelings stated to be
in "the heart."
What can we conclude from all this? Anna Wierzbicka (1993) and
Anthropological resistance 31
Melford Spiro (1993) helpfully review cross-linguistic and cross-cultural
data regarding the alleged peculiarity of the modern Western concept of
the person. Wierzbicka concludes that while the Western concept of "self"
(with its implications of complete autonomy, stability, and isolation) is
culturally specific, every language has a way of referring to the "person" or
[Tlhe presence in all languages of words meaning "someone" or "who" (with the
same implications of "solidity," "unity," and "endurance" as the corresponding
English words) testifies that in the folk model of ordinary speakers the world is
indeed inhabited by "referentially solid" persons, with an "essential unity" and an
"enduring nucleus," rather than by mere sums and swarms of participations.
Similarly, while the English word "mind" does not have close semantic
equivalents in other languages, there are words for the concepts covered by
the English terms "thinking, "knowing," "feeling," and "wanting" in
every language she knows of (1993:212).
Spiro (1993) found that in many societies people value individualism less
than it is valued in the United States, but none of the reports he analyzed
shows that there are societies where people do not distinguish between
oneself and others or between inner thought-feelings and the outer world.
Unni Wikan (1987), for example, has criticized Geertz's description of Bali
ethnopsychology, noting that Balinese see a "fundamental difference be-
tween the outer forms of conduct and the inner life of experience"
(1987:348-9). Spiro also cites Ernestine McHugh's (1989) claim that
"Geertz's characterization of the Western conception of the person 'would
serve nicely to characterize the Gurung [of Nepal] conception of the person
as expressed in beliefs about the sae [soul]'" (McHugh 1989:83; first
bracketed insertion is ours, the second is McHugh's). Katherine Ewing
(1991) states that the Pakistani women she interviewed, while "firmly em-
bedded in interpersonal dependency relationships," (1991:132), still have
"stable internal representations" (1991:157) differentiated from their rep-
resentations of others. Mark Elvin's analysis of an ancient Chinese poem
shows that, "The speaker has a clear inward vision of herself as a relatively
coherent, enduring, and self-contained entity that makes decisions, carries
responsibilities, is possessed by feelings, and in general has a fate, a for-
tune, and a history" (Elvin 1985:59). (All quoted in Spiro 1993:128-34.)
Are ethnopsychologies right? In light of the preceding review, it is Butler's
fascination with surfaces and rejection of inner depth ("... acts, gestures,
and desire produce the effect of an internal core or substance, but produce
this on the surface of the body" [ibid. 1990:136, emphasis in the original]; "If
the body is not a 'being/ but a variable boundary, a surface . . ." [ibid.
1990:139]) that appears the more peculiar position cross-culturally and
32 Background
historically. Fredric Jameson has suggested that this theoretical discourse is
a product of capitalism at the end of the twentieth century (e.g., the empha-
sis on advertising and other surface imagery in post-industrial societies):
JTJhis is perhaps the moment to say something about contemporary theory, which
has among other things been committed to the mission of criticizing and discredit-
ing this very hermeneutic model of the inside and the outside and of stigmatizing
such models as ideological and metaphysical. But what is today called contempor-
ary theory... is also, I would want to argue, itself very precisely a postmodernist
phenomenon . . . the poststructuralist critique of. . . what I will shortly call the'
depth model, is useful for us as a very significant symptom of the very postmodern-
ist culture which is our subject here. (1984:61/199la: 12)
We should note that Butler's later Bodies that matter (1993) is less fas-
cinated by surfaces and speaks instead of femaleness, maleness and inter-
mediate possibilities as "a sedimented effect of a reiterative or ritual
practice" (1993:10). In this work she does not mean sex as surface ap-
pearances only, because she "concedes" that they are "sexually differenti-
ated parts, activities, capacities, hormonal and chromosomal differences"
(1993:10). At the same time, however, she suggests that the very discourse
conceding such differences contributes to a "further formation" of bodies
(1993:10), an even stronger claim for discursive construction than might
have been inferred from Gender trouble. We will turn to difficulties with
this strong version of construction shortly.
At this point Butler could retreat to the following position. Maybe the
view of people as ontologically separable from their social environment
and experiencing thought-feelings inside them is not so unusual cross-
culturally. Still, that does not make that view right. (For example, maybe
she would argue that universally it exists to enforce power relations.) That,
of course, would be a good point to make. After all, the argument that a
position is right because nearly everyone believes it is no more valid than
the argument that a position is wrong because very few people believe it.
(Notice that the latter argument is used by both the politically correct left:
"Your view is ethnocentric, hence is wrong" and the reactionary right:
"Your view deviates from the norm in our society, hence is wrong.")
However, there are a number of other difficulties with the strong stance
Butler takes against internalization.
Internal states are not necessarily innate or fixed. Butler claims that the
inner (psyche, biological body)/outer (society, performance) distinction
should be discarded, but, implicitly, she relies on this distinction in her
assumption that inner = fixed and outer = fluid. In Gender trouble "inter-
nal," "interior," and "inner" are regularly linked to "fixity" for her: for
example,"... the appearance of its own interior fixity" (1990:70) and "the
internal fixity of the self" (1990:134). She appears to assume that the only
theoretical choices are to state that one is born with one's identity fixed
Anthropological resistance 33
inside oneself or to claim that identity continually shifts on the "surface of
the body." Those are not our only choices. As anthropologists have argued
for most of the twentieth century, to a great extent identities are socially
learned and cross-culturally variable yet are developed by building up
psychological structures that give rise to powerful internal thoughts, feel-
ings, and tendencies to act a certain way. For example, Foucaulfs dis-
cussion of the "inscription" of military, educational and other disciplines
"on the bodies" of soldiers, school children, and others is entirely consist-
ent with anthropologists* observations of learned, cross-culturally variable
aspects of posture and movement and Bourdieu's (1977) and our dis-
cussion of the cognitive schemas produced by and producing such actions
(see end of this chapter and the chapters to follow). It is true that some
earlier anthropologists tended to assume a greater stability of identities in
enculturated social actors than Butler would. In the rest of this book we
will show how learning is consistent with both stability and a degree of
flexibility and potential for change.
Practice cannot signify without internalized schemas. Butler says gender
is performed through repeated "signifying practices" (1990:144-145). In
discussing the meaning of "meaning" for Geertz, however, we showed that
it makes no sense to say that performances carry meanings "inside" them
or are themselves equivalent to meanings or refer to meanings "in the air";
rather, they signify only by evoking internalized schemas.
Discourses about desire (emotions, thoughts, the body, etc.) are not the
same as desire (emotions, thoughts, the body, etc.). Anyone who has not
read too much postmodern philosophy recognizes this from their own
Furthermore, analyzing dominant discourses and avoiding
people's descriptions of their experiences allows Foucauldian discourse
analysts to ignore evidence that might prove them wrong. Studying social
discourses isolated from the meanings people give them can give a spurious
plausibility to theories positing the demise of the meaning-seeking subject.
Social discourses do not directly construct psychological realities. This is
another version of the "fax" theory of culture acquisition that we criticized
in talking about interpretivism.
This is not confused, just wrong, as
much of the rest of this book will undertake to show. For the time being, an
example Spiro (1993) gives from his fieldwork in Burma will help make this
point. Spiro went to Burma to learn how a society functions when people
believe the Theravada Buddhist doctrine of Anatta (no self), which can be
described as follows:
"Buddhism stands unique in the history of human thought in denying the existence
of such a Soul, Self, or Atman. According to the teaching of the Buddha, the idea of
the self is an imaginary, false belief which has no corresponding reality." (Rahula
1959:51; quoted in Spiro 1993:119)
After Spiro had spent a few months in Burma, however, he found that he
34 Background
had to change his fieldwork plans because the Burmese villagers he talked
to had not internalized the doctrine ofAnatta. This was based in part on
their experience (they felt they had a self) and in part on their acceptance of
a different religious doctrine: the idea of reincarnation. As they pointed
out, why bother to avoid temptation in this life if there is no continuous
self that will reap the reward in the next life? Spiro notes there is a Burmese
couplet that expresses (and seemingly mocks) the Buddhist doctrine: "This
body will enjoy, that body will suffer" (Spiro 1993:120). Thus, the villagers
have adapted Buddhist teachings (reincarnation is internalized but anatta
is ignored) to help motivate them to do the right thing (cf. Weber's 1958
description of the dissatisfaction of early Calvinists with the doctrine of
Readers familiar with Butler's Bodies that matter are sure to point out
that she leaves room there for effects of discourse that run counter to
hegemonic norms (e.g., 1993:10,140) and thus she might not find Spiro's
observations in contradiction with her own. Yet, lacking a theory of
internalization, Butler cannot explain how dominant discourses and "re-
iterative or ritual practice" (1993:10) interact, leaving "gaps and fissures"
for difference and change.
Implications for current research on the constructedness of
culture and self
It is odd the way Foucauldians talk about the constructedness of "sub-
jects" while ignoring the work of psychological anthropologists who have
shown some of the subtle ways in which thoughts, feelings, and motives are
shaped by power that runs all through society, as Foucault (1978) puts it,
and is present not just in the state. In fact, substituting "culture" for
"power" in Foucault's writings produces statements that look a great deal
like Ruth Benedict's classic culture-and-personality theory in Patterns of
culture (1959).
Anthropologists interested in construction of cultural identities and
products (rather than of selves) are similarly disadvantaged if they never
look beyond such aspects of extrapersonal culture as linguistic structures,
rituals, or the mass media. For different reasons in each case - slow evolution
in the case of language, the aim of vivifying dominant values in the case of
most rituals, the need to appeal to a wide audience in the case of mass media -
texts of this sort tend to be conservative. While we can find some examples of
inventive performances by studying these things, many more examples and
much more insight into the forms such invention takes would be accumu-
lated by talking to people and observing their everyday actions to see how
they selectively internalize or read new meanings into texts that would seem
Anthropological resistance 35
on the surface to simply reproduce dominant social structures (e.g., Radway
1984; Willis 1990). Moreover, without a model of how people learn and act
on what they have learned, it is very hard to explain why some sorts of
cultural inventions or creative juxtapositions occur and others do not or why
some "normalizing** practices (as Foucault puts it) are resisted and others
are not.
At its worst (typically in the hands of non-anthropologists practicing
cultural studies), ignoring subjective meanings leads to ethnographic
works that merely display exotic practices or surprising juxtapositions of
the exotic with the analyst's own (nonexotic?) practices. Thus, even though
many practitioners of this approach denounce "exoticizing" or "Other-
ing," their practice tends to have the very effect they denounce. An
egregious example is the film Reassemblage (Trinh 1982), which consists of
lengthy footage of scenes from a Senegalese village. Guidance on this
village tour - which focuses on flies, dead animals, naked breasts, and
other exotica - is provided by a series of quirky voice-over remarks by
Trinh Min-ha that are unrelated to the scenes shown, the villagers them-
selves remaining voiceless and unrepresented. While Trinh*s aim may be to
question the authority of the ethnographic filmmaker (Rice 1993), she ends
up producing a typically postmodern spectacle that objectifies its subjects
to a much greater extent than does the ordinary ethnographic film.
Some descriptions go too far in the other direction, projecting onto the
people they study the moral exhaustion and rootlessness that is probably
more descriptive of postmodern theorists than of the subjects of this
Again, such attribution is a danger if one analyzes the
productions of cultural elites, such as advertising or architecture, without
finding out the meanings of those things for their consumers (Strauss,
Some anthropologists in this camp recognize this. George Marcus and
Michael Fischer, for example, praise psychoanalytic ethnographies such as
Robert Levy*s Tahitians: mind and experience in the Society Islands (1973)
on the same grounds we would:
If Geertz seems to suggest a direct relation between public forms and emotional
dynamics, Levy suggest a division between public surfaces and private behavior.
Geertz is in the tradition of Durkheim (ritual or public form helps generate
sentiment) and of philosophers George Herbert Mead and Gilbert Ryle (there is no
private language; all consciousness is intersubjective, mediated by public com-
municative forms). Levy is in in the tradition of Freud (concentrating on layered
notions of personhood and the self), yet he is able to establish the shared intersub-
jective nature of the most private behavior. Locating cultural organization at the
level of personal emotional expression and self-definition is thus one of Levy's
main achievements. (Marcus and Fischer 1986:50)
36 Background
Oddly, three pages earlier, Marcus and Fischer laud Geertz's "Person,
time and conduct in Bali" for ignoring the intrapersonal realm: "The most
appealing and effective aspect of Geertz's essay is that he does not resort to
discussions of psychology, even though he is certainly talking about 'the
Balinese mind"' (ibid.:47; emphasis added). Yet it is precisely Levy's
"resort to . . . psychology" that enables him to produce the innovative
ethnography Marcus and Fischer admire!
Historical materialism: people can resist cultural meanings
Concern with the changeability of ideas and practices typifies not only
writers in the field of postmodern cultural studies, but also many contem-
porary historical materialists. Criticisms of psychological approaches
come even from anthropological historical materialists who, unlike an
earlier generation of Marxists, acknowledge the force of cultural
Hegemony and cultural meanings
For a long time most historical materialists - in anthropology as well as
other disciplines - hewed to Marx's emphasis on the means and social
relations of production and eschewed the "superstructural" realm of idea
systems. Structural Marxists such as Althusser who were influenced by
Levi-Strauss abandoned that stance beginning in the 1960s (as the Frank-
furt School influenced by Freud had before them), and the shift toward
acknowledging the importance of cultural meanings accelerated with
Raymond Williams's (1977) promotion of (his version of) Antonio
Gramsci's (1971) discussions of "hegemony."
In Williams's terms,
hegemony is
a whole body of practices and expectations, over the whole of living: our sense and
assignments of energy, our shaping perceptions of ourselves and our world. It is a
lived system of meanings and values . . . It thus constitutes a sense of reality for
most people in the society... It is, that is to say, in the strongest sense a "culture,"
but a culture which has also to be seen as the lived dominance and subordination of
particular classes. (1977:110)
In other words, some (all?) cultural meanings are part of a system of
exploitation. In recent years, it has become increasingly common for
anthropological historical materialists to grant cultural meanings a signifi-
cant role in social action.
Even in this intellectual environment, some anthropological Marxists
are uncomfortable with theories of socialization that would explain how
"shaping perceptions," "meanings and values," or a "sense of reality" are
Anthropological resistance 37
learned. For example, Richard Fox (1985) approvingly quotes Williams's
definition of hegemony and acknowledges, in a later work (Fox 1989), that
while his main concern is cultural innovation, "We also know that cultural
understandings are compelling and condition human intention and ac-
tion" (1989:266). Yet he adds disapprovingly, "Even in the recent guise of
cultural scenarios, culture is taken to set the terms of social action in
advance of individual cultural practice. Similarly, the individual is often
seen as an assemblage of personality traits set in childhood socialization
and subsequently compelling the adult's life-course" (1989:274). Rejecting
both the images of individuals as "butterflies, flying carefree above their
culture" and as "ants, grounded by the weight of cultural tradition"
(1989:266), Fox argues that the best image is that of the grasshopper:
"There are individual leaps, not flights, of cultural innovation, but they are
of short duration, and they are propelled by bouncing off the resistance
offered by the dominant cultural beliefs" (1989:272). In another example,
Eleanor Leacock (1985) makes a laudable effort to develop a Marxist
theory of individuals, arguing (exactly as we do in our introduction) that
"although the societal is the more inclusive level, the psycho-social is not a
mere response to social conditions but involves a specific dialectic of its
own" (1985:79). However, in seeking a framework within which to investi-
gate that dialectic she turns to work in symbolic anthropology, rejecting
the possibilities offered by psychological anthropology, which she sees as
currently too fragmented and atheoretical and formerly too dominated by
a culture-and-personality approach she excoriates: "Personality and cul-
ture anthropology was functionalist to an extreme; all was primarily
governed by internalized values; change, extrinsic to society, was reduced
to happenstance; and resistance to exploitative systems no more than
personal deviance" (1985:71).
In response, we agree that work by culture-and-personality theorists in
the middle of this century as well as some of the work on "cultural
scenarios" now (Fox does not cite anyone, so we are not sure which work
he has in mind) does paint an overly static picture of individual personali-
ties and cultures. One of the goals of this book is to theorize socialization in
a way that explains both change in individuals' understandings during
their life and change in dominant cultural understandings historically.
Theories of socialization are not necessarily functionalist, even if mid
twentieth-century versions tended to be - along with most anthropological
theories of that time.
We think it important, however, to preserve what is valuable in func-
tionalism. As we discussed earlier regarding arguments that "culture is
invented," neither individuals nor societies constantly or completely re-
make themselves. Fox's whimsical image of the grasshopper "bouncing
38 Background
off the resistance offered by the dominant cultural beliefs" can be mislead-
ing if we forget that while human grasshoppers can make deliberate
choices to bounce off certain cultural beliefs, the motivation to hop and
the path of their travel will depend on internalized schemas of themselves
and the world they occupy. Sometimes these schemas lead to a reproduc-
tion of current power structures. Not to acknowledge this centripetal
tendency or understand it is particularly problematic politically, because
it gives rise to the unrealistic expectation that it is easy to bring about
fundamental social change (as Fox 1989 acknowledges and illustrates
with the case study of Gandhi's only partially successful attempts at
cultural innovation).
Equally, cultural schemas can motivate action that leads to change.
Thus we question the assumption made by Williams, Fox, and possibly
Leacock, that all culture is a "weight" that drags people down and rein-
forces the power of the dominant class. We see in this metaphor an echo of
the Enlightenment's rejection of religion and the other ruling ideologies of
the time (cf. Marx's 1978:595 reference to "the traditions of all the dead
generations"). Once we appreciate that "culture" is merely a name for all
of the learned schemas that are shared by some people, as well as all of the
diverse things from which these schemas are learned, we can see that even
in their most creative and progressive actions, people are culturally moti-
vated. The question is not, when is action culturally compelled and when
does it escape culture and by how much? Rather, we should ask questions
such as: what are the diverse experiences from which people gain their
interpretive frameworks? Do some experiences create schemas that chal-
lenge the schemas gained from other experiences? On which schemas do
people act in a given situation and why?
Implications for current studies of resistant consciousness
One major aim of some anthropologists who are historical materialists is
to explain and describe the "consciousness" - i.e., understanding of social
inequalities and of their place in the social structures reproducing those
inequalities - of some group of people. This is a topic, obviously, to which
psychological anthropologists can contribute a great deal and any attempt
that omits their contribution is likely to miss the mark. This can be seen
from one of the best such recent attempts, Jean and John Comaroffs
discussion of hegemony, ideology, consciousness, and culture in the Intro-
duction to Of revelation and revolution, volume I (1991),
The Comaroffs note that it is typical of Western social scientists to speak
of consciousness as if it is either present or absent. They suggest instead
that there is
Anthropological resistance 39
a chain of consciousness - . . . a continuum whose two extremes are the unseen and
the seen, the submerged and the apprehended, the unrecognized and the cognized
. . . Between the conscious and the unconscious lies the most critical domain of all
for historical anthropology and especially for the analysis of colonialism and
resistance. It is the realm of partial recognition, of inchoate awareness, of ambigu-
ous perception, and sometimes, of creative tension: that liminal space of human
experience in which people discern acts and facts but cannot or do not order them
into narrative descriptions or even into articulate conceptions of the world; in
which signs and events are observed, but in a hazy, translucent light; in which
individuals or groups know that something is happening to them but find it difficult
to put their fingers on quite what it is. It is from this realm, we suggest, that silent
signifiers and unmarked practices may rise to the level of explicit consciousness, of
ideological assertion, and become the subject of overt political and social contesta-
tion - or from which they may recede into the hegemonic, to languish there
unremarked for the time being. As we shall see, it is also the realm from which
emanate the poetics of history, the innovative impulses of the bricoleur and the
organic intellectual, the novel imagery called upon to bear the content of symbolic
struggles. (1991:29)
The first pole of this continuum - that of extreme lack of awareness - the
Comaroffs correlate with hegemonic representations, those representations
of the world that "are so habituated, so deeply inscribed in everyday
routine, that they may no longer be seen as forms of control - or seen at
all" (1991:25). At the other pole of this continuum-complete awareness -
are ideological representations, the explicitly articulated values and beliefs
of a particular social group (1991:24).
This model is very suggestive.
We should, indeed, expect ideas that are
explicitly laid out in public representations (e.g., a speech about free
enterprise) to be easier for individuals to bring to awareness and articulate
than those that are only implicitly represented in symbols and institutions
(such as the markets for stocks and bonds). Yet, too much is glossed over
in this scheme. Much like Bourdieu's opposition between what is said and
unsaid in society (dogma vs. doxa), which we will criticize in the final
section of this chapter, several different cognitive states are lumped to-
gether at the hegemonic end where power is naturalized and in the "liminal
space" in between hegemony (uncontested) and ideology (contested ideas).
These different cognitive states have very different potentials for fuller
consciousness and eventual (individual and social) change.
Thus, at the hegemonic end, there is a clear difference between what is
unsaid because it is unknown (e.g., ideas and practices of a very different
society); what is unsaid because it is very well known, but as a motor habit
or image rather than as a set of propositions; and what is unsaid because it
would require new connections among scattered bits of knowledge people
have. Let us take as an example of the first, most non-Native Americans'
40 Background
ignorance of the role of the berdache in many Native American societies.
An example of the second might be the typically unstated, although
well-known, expectation in late twentieth-century North America that
professional women should wear subtle makeup and have short, well-
groomed hair. An example of the third would be the way power and
authority are compatible with male sexuality but not with female sexuality
for a large majority of middle-class North Americans today. In the first
case there is no representation of these ideas in some people's schemas:
unless they have taken cultural anthropology courses and read, for
example, Walter Williams's The spirit and the flesh (1986), most non-
Native Americans (and probably many contemporary Native Americans
too) simply have no beliefs whatsoever about berdaches. Regarding the
second example - that of the way professional women are supposed to look
from the neck up - we would guess that a majority of North American
adults have this schema, but it would come to awareness as an image,
rather than a statement. This schema would be one of many that underlie
the third example of an unsaid: that female sexuality is considered incom-
patible with power and authority by many contemporary middle-class
North Americans. Other schemas that specify parts of this idea include
those representing professional women's dress codes; those giving proto-
typical images of men in different positions of authority; those suggesting
that popular girls in high school should not sound too sure of themselves in
class; and those representing the appearance of the type of women who is
supposed to be sexually attractive (schemas that are somewhat different
for each person, but still similar in many respects for large numbers of men
and women at any given time). Like the second sort of unsaid knowledge,
this third sort is cognitively represented - but in this case not as one set of
strongly interconnected cognitive elements but scattered among a large
number of schemas that may be only loosely connected to each other.
In the first case people trying to create social change can bring the unsaid
to awareness by introducing new ideas into popular discourse, but these
ideas are unlikely to be accepted if people see no way to connect them to
the knowledge they have. The second sort of unsaid knowledge is the
easiest to bring to consciousness, but the result is likely to be an isolated bit
of awareness that affects only an isolated bit of behavior. For example, in
the movie Working girl, the central character is a secretary (played by
Metanie Griffith) who wants her male bosses to stop seeing her as a sex
object and start taking her ideas seriously. After she observes a new boss, a
woman, Griffith's character cuts her long hair and scrubs off her heavy
makeup. This was easy for her to do, but it does not change much. It is in
the realm of the third sort of unsaid knowledge that there is the greatest
potential for fundamental cognitive and behavioral change. Pointing out
Anthropological resistance 41
connections between previously isolated bits of people's assumptions can
create both greater awareness of those bits and new cognitive links among
them, the result of which will be a much more frequently accessed schema
that becomes increasingly salient. This is what the consciousness-raising
technique of "naming" does: by giving a wide-ranging set of practices a
single label (e.g., "sexual harassment") one is not so much making it
possible to say what could not be said but rather creating a set of links
among previously disconnected memories and assumptions. This will tend
to bring the whole set to awareness when one part is experienced. When
what was unsaid becomes said in the first case this is likely to produce a
distanced, "Huh, that's interesting" reaction; in the second a smile or nod
of recognition; but in the third it may evoke that excited "aha!" that comes
from feeling one has an important new insight.
Similarly, there are different sorts of ambiguous, contradictory mental
representations that lie in the space between the ComaronV poles of
hegemony and ideology. In chapter 8 Claudia Strauss discusses several
possibilities, including horizontally compartmentalized schemas (schemas
deriving from conflicting social representations-that can come to a person's
awareness with equal ease, but are activated in different contexts); vertically
compartmentalized schemas (schemas deriving from conflicting social rep-
resentations, one of which is more readily articulated by the person than the
other and which are activated in different contexts); and integrated schemas
(ones deriving from conflicting social representations, but which are inter-
linked in the person's cognitive network, hence activated together). Alexei
Yurchak (1997) offers another possibility: a divided awareness in which, like
Strauss's horizontally compartmentalized schemas, each can be voiced, but
unlike Strauss's examples, there is subliminal awareness of the conflict
between the parts that produces psychic discomfort. This divided aware-
ness, he claims, was widespread in the 1970s and early 1980s in the former
Soviet Union, with a large majority of the population expressing (and to a
certain extent feeling) both allegiance to the established order in official
settings and cynicism about it when they were with their friends. Critiquing
the Comaroffs' model, he argues that this sort of divided awareness does not
lead to resistance (in this case, because people do not imagine that any
resistance could possibly succeed), but it does produce apathy and with-
drawal that are debilitating to the established order. Obviously, there are
many other cognitive possibilities as well, including psychic conflict of the
sort described by Quinn (1992). This suggests that work combining psycho-
logical anthropology and critical theories of consciousness would be very
fruitful (as it was, for example, in Mintz's 1960 book Worker in the cane,
which combined materialist and psychological explanations of one Puerto
Rican worker's life choices).
42 Background
Cognition in practice/discourse pragmatics: meanings depend on
A final set of objections to theories of internalization has been raised by some
of our fellow cognitive and linguistic anthropologists, including some of our
closest colleagues and interlocutors. The concern raised by these theorists is
that a focus on internalization ignores the way actions are shaped through
ongoing interactions of people with each other and their material environ-
ment in the course of action.
Within cognitive anthropology this approach
("cognition in practice*' or "situated cognition" or "activity theory")
sometimes contrasted with our "cultural models" approach.
Meanings and context
Most of these authors have no objection in principle to studies of culture in
persons. Edwin Hutchins, for example, states very clearly that his studies
of the way cognition is distributed among people and cultural artifacts is
not meant to "dissolve the individual and the psychology of the individ-
ual" but only to "connect what is in the person to what is around the
person" (1996:64, 65).
Alessandro Duranti, while giving "situated dis-
course" central priority in the ethnography of speaking, also states that the
ethnography of speaking "studies . . . how speech is related to and is
constructed by . . . speakers' assumptions, values, and beliefs about the
world" (1988:210). We hope it is obvious that an adequate ethnography of
speaking would have to do this, because the context alone is not sufficient
to determine people's action. For our part, we agree wholeheartedly that it
is important to look at situated cognition: Interactions between people and
their material environment or human peers can indeed lead to results that
are surprising (e.g., Holland 1992a, 1992b). We have also found very
helpful, and will apply in chapters 5 and 6, activity theory's attention to
mediating structures that ease routine tasks.
Some followers of Wittgenstein might be concerned, however, that even
to speak of "internalized cultural schemas" as we do is problematic
because it suggests that the learner is programmed with a dictionary or
encyclopedia of clearly defined signs accompanied by context-independent
rules of application, leaving no room for meaning to be affected by
context. Wittgenstein provides insightful discussions of why such a model
is utterly inadequate to account for the way people use, and learn from,
most signs. We agree with him, and in chapter 3 we present an alternative
model of internalization that can accommodate the complexities of social
learning and the dependence of meanings on the context. We suspect that
many of the objections that have been raised to meaning as a psychological
state determined in part by mental structures are responses to older
Anthropological resistance 43
cognitive models of knowledge representation. With connectionist models
it is much easier to see how meaning in our sense is completely compatible
with context effects. In fact, one of the virtues of connectionist models is
that they explain how people deal with unforeseen events. In the third
chapter we will also show that meanings are not necessarily conscious; to
say that "meanings shape action" is not to assume that people consciously
deliberate about what to do most of the time.
Implications for current research on situated meanings
Older models of both culture and cognition may be responsible for more
extreme statements, such as the following characteristic remark of Jean
[Tjt has become increasingly clear that "culture" and "cognition" are not the
analytic units whose relations need clarification in order to proceed [with the study
of everyday cognitive activity]. Certainly neither one is an element in "the activity
of persons-acting in setting," the unit of analysis adopted here. (1988:177)
But is neither an element? Lave's groundbreaking studies are useful in
emphasizing, for one example, the degree to which shoppers rely on the
packaging of food products and their organization on supermarket shelves
to simplify their shopping. In one typical micro-analysis she describes a
shopper who, suspicious that a wrapped chunk of cheese has been mis-
priced, avoids doing mental mathematics by simply comparing the price of
that chunk with that of another, similarly-sized, one out of the same barrel.
On the other hand, there is much else about shopping behavior that Lave
leaves unexplained, such as - to continue her own example - why some
shoppers habitually choose organic goat's cheese, some American cheese,
and others brie. Shopping behaviors such as these depend on cultural
understandings (about, say, gourmet cooking or health foods) that shop-
pers bring with them into the store and which, with sufficiently subtle
methods, could be elicited in discussions held outside of the store. Are not
such assumptions, too, "an element in 'the activity of persons-acting in
setting,'" one that is just as significant a component of shopper's behavior
as the store setting? By ruling out culture and cognition as inappropriate
units of analysis, Lave's model leaves out this half of the "constitutive
order" (1988:177ff.) and forecloses the possibility of analyzing what people
bring with them from context to context and interaction to interaction, as
it must be analyzed, in its own distinctive terms. And as useful as
Hutchins's studies of cognition distributed over multiple actors and ma-
terial artifacts have been, we must not lose sight of the fact that distributed
cognition within a person works differently than distributed cognition
44 Background
among several people or between them and artifacts. Clearly, both learned
dispositions and ongoing interactions are important; our emphasis on the
former here is a necessary step toward a more encompassing model that
incorporates both.
Toward a more fruitful resolution: Bourdieu
It is not enough to criticize: Our goal is to present a way of modeling
internalization that will yield a better understanding of cultural meanings
and social action. We end this chapter with a discussion of Pierre Bour-
dieu's Outline of a theory of practice (1977), because his model is similar to
ours in some crucial respects. While representatives of each of the positions
we described have drawn on Bourdieu's insights (e.g., Duranti 1988; Fiske
1992; Fox 1985; Lave 1988; Ortner 1984, 1990; and Rosaldo 1989), all of
them have missed his central insight into the importance of internalization.
A fresh look at his arguments will set us on a more productive path toward
a resolution of the dilemmas in current culture theory.
Bourdieu begins Outline (1977) by declaring the need to provide an
alternative to two extremes of social theory: theories in which action is
simply the mechanical enactment of learned rules or unconscious structures
versus theories in which action can be utterly unconstrained by social
conditioning (in the shorthand of recent social theory, the structure/agency
dispute). Bourdieu cites British structural-functionalism and Levi-Strauss's
structuralism as examples of the first approach and Sartre's existentialism
as an example of the second.
Bourdieu's alternative is not to say (as does Ortner 1990) that sometimes
humans enact learned structures and sometimes we are "free." Instead, he
argues, we are always constrained by the dispositions learned from our
experiences, but our habitual responses rest on knowledge that is not
learned from or cognitively represented as hard-and-fast rules. Our inter-
nalized (in his words, "incorporated" or "embodied") knowledge is less
specific than rules because it is learned through everyday practice. Every-
day practice is somewhat variable from one day to the next but still tends
to remain within the boundaries of what is culturally acceptable. The
knowledge acquired from practices of this sort is thus not highly precise,
but rather consists of more general categorical relations that can be
realized in different ways, depending on the context. This form of internal-
ization enables people to react flexibly to new contexts instead of enacting
the same practices over and over again. This imprecise knowledge Bour-
dieu calls habitus (to distinguish it from invariant "habits"): "systems of
durable, transposable dispositions . . . principles of the generation and
structuring of practices and representations which can be objectively 'regu-
Anthropological resistance 45
lated' and 'regular' without in any way being the product of obedience to
rules . . . " (1977:72; also 82-3). That "habitus" refers to the internalized
results of cultural experience is particularly clear from lines such as "The
habitus is . . . laid down in each agent by his earliest upbringing . . . "
(1977:81). Thus, Bourdieu (unlike some of his followers) does not contrast
practice with learned understandings; instead, he contends that we cannot
understand practice without rethinking the nature of learned understand-
One of the most important parts of Outline is Bourdieu's discussion of
the way a person's habitus is structured by his or her experiences. Al-
though he also discusses learning from explicit precepts and "structural
exercises" (such as games) he places the greatest stress on "apprenticeship
through simple familiarization" (1977:88) with regular patterns of associ-
ations enacted in everyday life. Apprenticeship by familiarization is nicely
illustrated by his example of the way Kabyle children can learn from the
arrangement of objects and space in the typical Kabyle house (1977:90-1).
The child is not taught a rule: "Always put the loom on the wall facing the
east" or "Light and heat are male" for example, but assimilates a general
pattern: looms and other culturally valued objects are usually found in the
part of the house that faces east and the objects men typically use are
almost always found in the brightest, warmest parts of the house.
To put it in the terms we set out at the beginning, the habitus (intraper-
sonal knowledge) is structured by objects and practices in culture learners'
public environments (the extrapersonal realm). This is just one side of the
interaction between the intrapersonal and extrapersonal realms, for the
habitus in turn structures public culture as people act, creating new (or
recreating old) objects and practices. In Bourdieu's theory, the tendency is
toward social reproduction. However, the overall tendency toward repro-
duction is accomplished through individual actions that are never exactly
the same from one context to the next because they are enactments of
dispositions themselves acquired from varying practices.
Our account is similar to Bourdieu's in several respects. We agree that a
close examination of internalization is the key to explaining both the
centrifugal and centripetal forces of culture, as well as advancing beyond
old dichotomies of structure and agency.
We, too, will argue that inter-
nalized cultural knowledge typically consists of flexibly adaptive under-
standings, rather than unvarying rules, (Understandings as we model
them, however, are not as reductive as the dualistic classification schemes
that Bourdieu's own habitus, steeped as it is in Levi-Straussian structural-
ism, leads him to emphasize.) Finally, we share with Bourdieu the view
that social life is a process of interaction between these imprecise private
understandings and the public objects and events which are both their
46 Background
source and product. There are several other respects, however, in which
our account diverges from his.
First, it is never clear in Outline whether Bourdieu is setting forth a
universal theory of the role of habitus or a theory that applies mainly to
nonliterate societies without formal schooling (1977:89, 186). One of the
greatest flaws of Outline is that too often he comes down on the latter side,
ignoring the fact that in all societies by far the greatest proportion of what
we know is not cognized in the form of rules.
Second, Bourdieu lays down rather hard-and-fast lines between the
"universe of discourse'* and the "universe of the undiscussed" (1977:168)
assuming, at several points in Outline, that the unsaid is unsayable - for
both cognitive and social reasons. He believes that the knowledge em-
bodied in the habitus is unsayable because "schemes are able to pass from
practice to practice without going through discourse or consciousness"
(1977:87). This does not follow, however. Learning by modeling, which
may occur largely out of awareness, is not forever after barred from
awareness. While it is true and significant that such knowledge tends to
remain backgrounded in consciousness, it is entirely possible to fore-
ground it and describe it: Novelists and the authors of humorous "how to"
texts (how to dress like a preppy, act like a real man, be a Southerner, etc.),
among others, recognize and verbalize this typically unspoken knowledge
all the time. Nor is this an ability limited to "advanced" societies.
Throughout Outline, Bourdieu constantly cites or refers to traditional
Kabyle proverbs and informant commentary that show a very clear recog-
nition of what, a paragraph or so earlier, he had labeled unconscious
knowledge (1977:8, 94, 98). Furthermore, the inability to recognize what
one knows is consistently confused in his account with the social impropri-
ety of discussing what one knows, particularly the impropriety in Kabyle
and other precapitalist societies of discussing the potential material bene-
fits of status maneuvers.
Although in some places he claims that these
benefits are not seen as such (1977:191,192,196), elsewhere he indicates
that such benefits may well be recognized by Kabyle men and women and
even discussed privately, although not alluded to in public discourse
(1977:12,43,50,198, n.7).
Third, while Bourdieu provides an insightful discussion of one aspect of
centrifugality - flexible responses to new contexts that are similar to but
not exactly the same as previously encountered situations - he underesti-
mates other centrifugal effects such as change over time. One of the key
sections of Bourdieu's Outline of a theory of practice (1977) is "The
dialectic of objectification and embodiment" in chapter 2 (1977:87-95).
This section makes the same point we are trying to make in this book:
extrapersonal culture is the product of intrapersonal culture, that is, of
Anthropological resistance 47
minds structured a certain way, just as intrapersonal culture is structured
by extrapersonal culture. Yet, in this section Bourdieu overstates the
extent to which these will reflect each other: "the mental structures which
construct the world of objects are constructed in the practice of a world of
objects constructed according to the same structures . . . The mind is a
metaphor of the world of objects which is itself but an endless circle of
mutually reflecting metaphors" (1977:91). Bourdieu makes the same point
for rituals, customary law, and other social practices and institutions. This
metaphor of reflection greatly oversimplifies processes of cultural produc-
tion - even in the preliterate nonurbanized societies where Bourdieu
thought the dialectic of objectification and embodiment particularly
As we will show in the next section, the same theory of cognition
that accounts for the centripetal effects that Bourdieu emphasized will also
allow us to account for centrifugal effects that he largely ignored in his
overemphasis on the unchanging mutual reflection of mental structures
and extrapersonal structures (e.g., ibid.: 91).
Fourth, while Bourdieu has been an influential proponent of "em-
bodied" knowledge, his use of "embodiment" ignores the role of emotion
and motivation in action. His actors have gut reactions, but in his theory of
practice it does not seem to matter how much they care about what they
are observing and doing. Thus, he ignores the role of motivation and
emotion as forces for either reproduction and uniformity or diversity and
change, as well as the role of deliberate instruction, which (as we will show
in chapter 4) sometimes has the purpose of arousing strong feelings to
make lessons memorable.
This last weakness is related to a further problem with Bourdieu's theory.
He seems to assume that familiarity with social practices is sufficient for
acquisition of their regularly associated features. We will argue, on the basis
of psychological and neurological findings, that the process is more compli-
cated than that: Not all regularities in practice are remembered equally well
because the learner's motivational state makes a difference in what learners
pay attention to and how well their experience "sticks."
Finally, for Bourdieu, actors' intentions are mostly or entirely epiph-
enomenal (1977:76-77; 1984:474). We disagree. Intentions can give rise to
deliberate problem-solving thought that builds upon but may transcend
the workings of the habitus.
This kind of problem solving sometimes
leads to radical cultural innovation, other times to new ways of reproduc-
ing existing social structures.
In other words, we believe Bourdieu's theory of the habitus is promising,
but not completely satisfactory.
In the following chapters we will present
an alternative theory that is similar in many ways to his, but avoids the
problems we have just described.
Schema theory and connectionism
We have said that meanings are based on cultural schemas,
schemas that
have come to be shared among people who have had similar socially
mediated experiences.
What are schemas?
To illustrate what schemas are and how they work, imagine two televi-
sion commercials for different brands of beer. In one, the scene is a party
filled with attractive and well-mannered men and women in their late
twenties or early thirties. The other ad features potbellied lumberjacks
joking loudly. You, the viewer, immediately make some assumptions
about these people. You might assume that the first group of people are
middle-to-upper-middle class, probably college educated, while the second
are probably not college educated. Based on these assumptions, you have
some guesses about the kinds of beer they are touting: a somewhat more
expensive brand in the first commercial than in the second. The very words
we used to describe these commercials probably led you to some further
assumptions. You no doubt assumed that the lumberjacks were men. You
may even have a picture of what clothing they were wearing (flannel shirts
and pants held up with suspenders). Even if you did not arrive at these
particular assumptions, surely you had some interpretations - some mean-
ings - that were not contained in the information we gave you.
On what basis did you arrive at your interpretations? The description we
gave elicited your schemas for lumberjacks, for class differences, for beers
or for consumer items, generally, and for television commercials. If you
know what a lumberjack is, your schemas contained information about the
general appearance and schooling of a typical one as well as the education
and income of typical "attractive and well-mannered" people in their late
twenties and early thirties. (Including, possibly, the sense that even if the
current income of the latter group is not high, they probably aspire to an
upper-middle-class lifestyle.) You may have a very detailed beer schema
that distinguishes different brands by taste and price, but if not, at the very
least (if you live in a society with a free-market economy), you have a
general schema about consumer items that includes the information that
goods usually come in different brands intended for different "market
Schema theory and connectionism 49
niches," and that one of the functions of advertisements like television
commercials is to arouse the interest of members of the relevant group of
potential consumers. Somehow all of this information was readily avail-
able, without any effort on your part, and led you to interpretations that
took you beyond the information given. This is what schemas do.
The use of "schema" in this sense comes from a tradition of work in the
cognitive sciences that can be traced back through Jean Piaget and
Frederic Bartlett to Immanuel Kant.
The essence of schema theory in the
cognitive sciences is that in large measure information processing is me-
diated by learned or innate mental structures that organize related pieces
of our knowledge. As we will explain shortly, schemas, as we think of
them, are not distinct things but rather collections of elements that work
together to process information at a given time. Cognitive scientists have
traditionally used the term "schema" to refer to generic knowledge of any
sort, from parts to wholes, simple to complex, concrete to abstract. Thus,
there can be a flannel-shirt schema as part of a lumberjack schema, a beer
schema and an alcoholic beverages schema, a television commercial
schema and a postindustrial economy schema. A great many schemas are
cultural schemas - you share them with people who have had some
experiences like yours, but not with everybody. Another name for cultural
schemas (especially of the more complex sort) is cultural models (D'An-
drade 1995; D'Andrade and Strauss 1992; Holland and Quinn 1987).
Some published examples of cultural models are Trobriand Islanders'
models of land tenure rights (Hutchins 1980); Mexican men's and women's
expectations about the gendered life course (Mathews 1992); and US
Americans* models of romance (Holland and Skinner 1987), marriage
(Quinn 1982, 1987, 1991, 1996, chapters 6 and 7 in this book), and the
environmental movement (Kempton, Boster, and Hartley 1995).
Schemas sometimes reconstruct our memories of past events,
mine the meanings we impart to ongoing experience, and give us expecta-
tions for the future. Schemas also fill in missing or ambiguous information:
just think of everything that can be left unsaid in any conversation because
speakers assume their interlocutors share their schemas. Without these
learned expectations regarding the way things usually go, it would be
impossible to get anything done, plan for the future, or even interpret what
is happening; and without schemas that were at least partly shared, social
interaction would be impossible as well.
While schemas are a necessary part of most action, they are not always
good. Negative stereotypes are one sort of schema. Like any other kind of
schema, negative stereotypes sometimes shape our interpretation of the
present, memories of the past, and anticipations of the future.
example, if you assumed that attractive and well-mannered people in their
50 Background
late twenties or early thirties are likely to be college educated, do you also
hold a negative stereotype of people without college educations as being
less attractive or less well-mannered? (And if we misjudged your assump-
tions about this, what misleading schemas might we hold about the typical
reader of this book?) To understand how such stereotypes are formed and
the prospects for changing them, we need a more detailed explanation of
schemas. The rest of this chapter provides one current way of thinking
about schemas and the meanings that arise from them.
Introduction to connectionism
While schema theory is not new, there has been a radical change in the way
some cognitive scientists model schemas. This new approach (actually, a
revival of some older approaches) is called "connectionism'*
(or "parallel
distributed processing'*
or "neural network modeling"). Why should
anthropologists be interested in it?
As we stated in chapter 1, theories of cultural meaning (and meaning in
general) are caught between older views of meanings as fixed, shared and
consistent in a society and newer ideas of meanings as variable, differen-
tially distributed, and sometimes inconsistent. Many anthropologists (and
theorists in other fields)
assume that if they analyze meanings as psycho-
logical, they are committed to the older view of meanings and only a
nonpsychological theory of meaning can explain the way meaning varies
with context in a particular community, the changeability of meanings
over time, contradictions in the meanings on which people act, or intracul-
tural variation in meanings. (For example, recall, from the last chapter,
our discussion of Judith Butler*s equation of interiority with fixity.) Such
anthropologists may even be uncomfortable with terms like "cultural
models,'* which perhaps bring to mind images of rigid structures.
Current connectionist models of cognition J^VQ US a much less rigid way
of understanding schemas and the meanings to which they give rise (that is,
less rigid than alternative ways of thinking about schemas that we will
describe shortly). Meanings generated by schemas, in connectionist
models, are mental states but are shaped by the learner's specific life
experiences and are sensitive to activity in a particular context. While often
similar from person to person, context to context, and one period of time
to another, they can vary and change. For our purpose of reanalyzing
cultural meanings, then, connectionist models are useful as a heuristic:
They unstick our assumptions, helping us to imagine a way that meanings
can be subjectively imposed yet responsive to the objective world. Connec-
tionist models help provide new answers to other disputes in cognitive
science as well. For example, are cognitive mechanisms domain specific or
Schema theory and connectionism 51
general purpose? Do nonhuman animals have mental representations? As
we shall explain later, connectionists usually favor general-purpose mech-
anisms but always assume some domain-specific constraints, and connec-
tionist models would apply as well to the mental representations of other
species. Connectionism is not sufficient for a complete theory of cultural
meaning, but it gives us a productive way to begin the process of rethinking
what meanings are and how they arise.
Metaphors for knowledge
When we think about the form of internalized knowledge, what do we
imagine? One way to think about the knowledge in our heads is to use a
language metaphor. This metaphor suggests that when we learn some-
thing, there is a sentence in our head representing that knowledge. So if we
know, for example, that guests should be offered something to drink, the
language metaphor pictures that knowledge spelled out in a sentence that
sits in our brain somewhere (OFFER GUESTS DRINKS - only not in
English, but in whatever symbols and syntax the brain uses to store
information). According to this model, learning is the inscription of these
sentences into our brains. Applying this knowledge is a process of drawing
logical inferences or satisfying if-then rules (THESE PEOPLE ARE GUESTS,
THEREFORE OFFER THEM DRINKS); and revising this knowledge means
deleting or amending old propositions (for example, learning the new rule,
sometimes called a language of thought model (Fodor 1975).
intelligence modelers who use a language of thought model are often called
"symbolic processing" modelers.
Connectionist artificial intelligence modelers use, instead, a neural meta-
phor to picture knowledge. Suppose, connectionists suggest, we think of
knowledge not as sets of sentences but as implicit in the network of links
among many simple processing units that work like neurons. Neurons are
arranged in layers. Sensory neurons are activated by particular features of
the world (e.g., the sight of bright red or smell of smoke); motor neurons
are connected to muscle and gland cells, sending the signals that lead to
behavior and feelings; and interneurons lie between, combining signals
from, in some cases, over a hundred thousand other neurons and learning
to react to complex combinations of features. In each layer beyond the
first, all a neuron does is receive excitatory and inhibitory signals from
other neurons, combine them, and, if it is activated past its threshold, send
excitatory or inhibitory signals on to other neurons (or muscle or gland
cells). No single neuron knows much, but thousands of them working in
parallel produce intelligent action. The chemical synapses that predomi-
52 Background
nate in the brain are modified with experience; with learning, neurons
undergo structural changes so that in the future the firing of one is
thereafter more (or less) likely to contribute to the firing of another. In
other words, learning leads to neural changes that determine the pathways
through which activation spreads and the eventual interpretation and
response that is evoked in someone by a given event or thing. Finally, cells
seem to be clustered in functional assemblies or groups with every human
action drawing on a great many of these cell assemblies widely distributed
over the nervous system (Kandel, Schwartz, and Jessell 1991, 1995; Mer-
zenich and Sameshima 1993).
Connectionist computer models draw on these features of neurons to
simulate human knowledge, learning, and action. In these models, knowl-
edge is not represented by symbols strung together in sentences, but by
simple processing units arranged in layers (input, output, and one or more
layers in between). Aside from units in the input layer, which are activated
by (computer-simulated or actual recorded) experiences, each unit simply
sums the positive and negative signals it receives from other units and
passes on a weighted positive or negative total to the other units with
which it is connected. (Or so it is imagined. All that is really happening is
numbers changing, but these numbers are supposed to represent units
exciting and inhibiting each other.) The weights on connections between
units are modified through repeated exposure to examples of associations
that need to be learned. Typically, many units will be working in parallel
until some units in the output layer are excited past their threshold and a
stable answer is reached. No single unit knows much, but the combined
action of many of them, linked by weights modified by repeated experi-
ence, leads to intelligent outcomes. Finally each particular connectionist
model consists of a group of units that participate jointly in responding to
a related set of inputs.
In such models the belief that guests should be offered something to
drink is not a sentence in a "language of thought" but rather a tendency of
units representing aspects of guests and situations to activate other units
(and those, still other units) that in the end, with sufficient training, initiate
the behavior of offering an appropriate beverage. In this framework a
schema (e.g., for company manners) is not a set of sentences but rather a
pattern of interaction among strongly interconnected units. It follows that
for connectionists, schemas vary in their schematicity, depending on the
strength and density of the interconnections among the units of which they
are composed. Schemas such as the one for company manners are typically
learned when repeated participation (either as guest or host) in these sorts
of social interactions creates a gradual strengthening of the weights of
association among these units.
Schema theory and connectionism 53
If talk and other language are part of a model's input, they can also be
represented, but unlabeled, undiscussed sensations (e.g., the distinctive
cast of a person's face, Cottrell and Metcalfe 1991, summarized in Church-
land 1995) can be represented as well. Language is not privileged in this
formulation. "Embodied*' ideas can be represented as well as (and perhaps
even more readily than) highly abstract ones.
All kinds of experiences,
including those provided by texts and other symbolic material, can be
represented in a connectionist model.
In these models new knowledge does not add or subtract sentences but
rather consists of changing connection weights that shift the likelihoods of
what units will activate which. (Hence the name "connectionism": knowl-
edge is in these connection weights.) This process is much as Bourdieu
describes it for learning the dispositions of the habitus (see chapter 2). Like
Bourdieu, connectionists assume that knowledge (even knowledge that
comes to consciousness as sentences) is not represented sententially in our
heads and that much learning proceeds without need for explicitly stated
rules (compare Bourdieu 1977:8 with Smolensky 1988b:4ff.).
Schemas understood in this way can be highly context sensitive because
they consist not of the simple pairs of links postulated in older association-
ist models, but of whole interlinked networks. For example, from repeated
participation in encounters with guests we learn not a one-to-one associ-
ation between guests and offers of something to drink but a whole network
of links between features of the guest (Someone we hope will leave soon or
someone we want to stay? Child, adolescent, or adult? Male or female?);
features of the situation (Did they drop in or were they invited? Is it a
morning, afternoon, or evening visit?); possible offers of hospitality (No-
thing? Hot drink or cold? Alcoholic or non? If coffee, regular or decaffein-
ated?), and more. The meaning of any one situation is dependent on a
whole network of learned associations of this sort.
These networks process information holistically. An event activates all
the units that respond to the features of that event; these units, in turn, then
activate all the others to which they are strongly linked by associations
learned from past experience, exciting some units and inhibiting others.
This process continues until the network reaches a response that satisfies as
many of the constraints as possible in the situation. As we will illustrate
later, the combined influence of the units activated by the particular
features of any given event can lead to different outcomes from one
situation to the next.
In other words, schemas as construed in connectionist models are well-
learned but flexibly adaptive rather than rigidly repetitive. They can adapt
to new or ambiguous situations with "regulated improvisation," to use
Bourdieu's term (1977:11). The reactions that are the output of connec-
54 Background
tionist networks are improvisational because they are created on the spot,
but regulated because they are guided by previously learned patterns of
associations; they are not improvised out of thin air. This would always be
the case, for instance, with the kind of "situated" cognition with which
practice-oriented analysts (e.g., Lave 1988) are concerned,
and the same
argument would apply to the "invented traditions" pointed out by many
anthropologists and historians (e.g., Hobsbawm and Ranger 1983). Both
in people's everyday creative practices and in the more unusual world-
historical instances of creation of salient cultural symbols, new practices
draw upon associations learned from the past. (This does not mean that all
creativity can be explained in connectionist models; see chapter 5.)
This distinction between relatively stable cognitive networks and the
ever-changing reactions that are the response of these networks to particu-
lar events is an important one. In the rest of this book we will use separate
terms for each: "cultural models," "schemas," "networks," "understand-
ings," "knowledge," and similar terms for the relatively stable cognitive
structures; "meanings," or "interpretations," for the thoughts, feelings,
and less conscious associations evoked when people's schemas meet the
world at a given moment.
Some of the dilemmas of culture theory
(indeed linguistic theory and philosophy of language as well) are easier to
resolve when we keep this distinction in mind. Using these terms, then, as
we said in chapter 2, we readily agree with Wittgenstein that the meaning
of a sign depends in part on the language game in which it is embedded. In
fact, the associations evoked by a sign depend not only on language games
as a generic type of practice, but on all the particulars of that practice at a
moment. To return to our coffee-hospitality example, the meaning of an
offer of a cup of coffee following a date is different from the meaning of an
offer of a cup of coffee during an afternoon kaffeeklatsch. Furthermore,
the offer of a cup of coffee following a first date, when it is undecided what
will happen next, could be different from the meaning of an offer of a cup
of coffee once it has become part of that couple's routine. Yet, just because
meanings depend in part on the context in which they are evoked, and this
context can shift, it does not follow that the meanings themselves are
completely unstable. If we think of meanings as the associations elicited in
actors at the moment, we can see that these associations are guided by the
actors' well-learned understandings.
Such interpretations of experience
occur automatically and rapidly, often triggering actions before we are
aware of them, which is why philosophers like Wittgenstein, trying to rind
them through introspection, are unable to pin them down. It would be
wrong to conclude from this failure of introspection that relatively stable
understandings do not exist.
Schema theory and connectionism 55
Caveats and clarifications
It may already be apparent that while connectionism is guilty of being an
associationist cognitive model, it is far more sophisticated than earlier
associationist theories like Skinnerian behaviorism. Earlier associationist
theorists ignored, indeed denigrated, the important role of cognitive
schemas that are highly structured and enable our interpretations to go
beyond the information given by the immediate stimulus. Connectionist
models, by contrast, take seriously and explain the mediating effect of
schemas, while providing an account of how schema-like properties of
mind arise. At the same time, connectionism retains what was always the
greatest contribution of associationist psychology: its focus on learning.
Even if some associationist psychologies go too far in assuming the mind is
a tabula rasa (an issue we will address at the end of this chapter), the extent
of cross-cultural variation in human knowledge attests to the degree to
which meanings are learned and makes it imperative to study the process
of culture acquisition.
A more complicated and easily misunderstood issue is what is really at
stake in the difference between using a language metaphor (as "symbolic
processing" modelers do) and a neural metaphor (as "connectionist"
modelers do) to understand cognition. It is easy to get confused about this
and reach incorrect conclusions (e.g., that connectionists cannot model
processing of symbols or that symbolic processors cannot model pro-
cedural knowledge). To clear up any possible confusion, let us differentiate
six things one could have in mind in championing a connectionist versus a
symbolic processing approach:
1 how best to model cognition on computers,
2 the content of what is learned,
3 what the learner is aware of knowing,
4 how information is learned,
5 how information is processed, and
6 the neural basis of cognition.
After we sort out these issues, it will be easier to state the strengths and
limitations of connectionism for our purpose of understanding cultural
L How best to model cognition on computers. Much of the confusion
about the claims being made by connectionists comes from the fact that it
is not always clear whether they and their critics are talking about psycho-
logical processes, brains, or computer programs. Connectionist models
are, usually, computer programs. The superiority of one computer pro-
gram over another can be judged by various criteria. Which sort of
56 Background
programming is easier to do? Which sort of model lets programmers come
closer to getting machines to perform like human beings? Finally, which
acquires its knowledge and reaches its responses in ways most similar to
the ways humans do?
Some of the excitement generated by connectionist models is due to the
fact that, for some problems, they require less labor to create and respond
more like humans than symbolic processing models do, (For other prob-
lems, on the other hand, they require more labor and may respond less like
humans.) We are not computer programmers and do not care which style
of programming is easier. Nor do we even care which style of program-
ming lets machines mimic human abilities, if they reach those ends in ways
very differently from human beings. What excites us about connectionist
models is that they seem better than symbolic processing models at simula-
ting how people build up cultural schemas and use them to extract mean-
ing from events, a point to which we will return after considering the rest of
this list.
2. The content of what is learned. Connectionist models are sometimes
seen as specializing in procedural knowledge ("know how," e.g., knowing
how to whistle), and symbolic processing models are said to specialize in
declarative knowledge ("know that," e.g., knowing that the Indus River
cities of Mohenjo-daro and Harappa existed around 2000 B.C.E.). (See
Giddens 1979 on "practical" versus "discursive" consciousness.) This is
wrong: symbolic processing models often simulate procedural knowledge
(think of all of the industrial robots that run on standard symbolic process-
ing artificial intelligence programs) and connectionist models can simulate
declarative knowledge (see, for example, Rumelhart, Smolensky, McClel-
land, and Hinton 1986, in which they present a model that could be said to
have the declarative knowledge that a room with a refrigerator is very
likely to have a stove, a room with a bathtub is quite likely to have a toilet,
and so on). This focus on content misses the central point that symbolic
processors represent procedural knowledge as if it were declarative, while
connectionists represent declarative knowledge as if it were procedural. At
stake is not so much what is known, but how it is known.
3. What the learner is aware of knowing. Following from the last section,
it would seem that treating all knowledge as being like declarative knowl-
edge should lead symbolic processing modelers to consider all knowledge
as fully accessible to consciousness, while treating all knowledge as being
like procedural knowledge should lead connectionists to consider all
knowledge as largely out of awareness. It would be too hasty to reach this
conclusion, however: Awareness is an ancillary concern for both connec-
tionists and symbolic modelers, and, as far as we know, the issue is not
often addressed by them.
Some connectionists have proposed, sugges-
Schema theory and connectionism 57
tively, that "the contents of consciousness are dominated by the relatively
stable states of the system" (Rumelhart, Smolensky, McClelland, and
Hinton 1986:39). We read that to mean that schemas tend to remain out of
awareness, with only the responses a network settles into arising to con-
sciousness. If true, that would help explain the fact that we tend to find it
difficult to articulate fully the contents of our schemas. (See chapter 6 for
further discussion.)
But there is no reason why symbolic processing
modelers could not make a similar claim.
Nor is the form of knowledge in consciousness - sentences versus
something less wordy - a reliable way of distinguishing connectionist from
symbolic processing models. As we will discuss under (5), connectionist
models can model verbal formulae such as an oft-repeated rule or poem or
4. How information is learned. This is a significant difference between
the two approaches to modeling. Although symbolic processing models
can learn (Carbonell 1990; Shavlik and Dietterich 1990), often they start
with all the knowledge they need. This does not mean that they simulate
innate knowledge, necessarily, only that symbolic processing models do
not always deal with the issue of how noninnate knowledge is acquired.
Connectionist models do not start out as blank slates either. However,
instead of beginning with a complete understanding of some domain, they
always start with some initial constraints only, and gradually acquire the
rest of their knowledge through exposure to a variety of specific examples
and repeated correction of inferences about those instances. As we will
discuss further at the end of this chapter, this approach does not account
for the full range of human learning. On the other hand, it does give us a
way of starting to think about the issue of learning, one that could be built
upon and modified as necessary to explain various forms of culture acqui-
sition. In some interesting applications, connectionists have provided
possible explanations of critical periods (Munro 1986) and stage-effects
(McClelland 1994) in learning.
5. How information is processed This is another significant difference
between the two approaches. Connectionist models are particularly good
at simulating automatic information processing; they are usually thought
to be much less adept than symbolic processing models at simulating what
Norman (1986) calls "deliberate conscious control."
An example of
deliberate conscious control is the effortful thought that most of us go
through in solving complicated arithmetic problems (What is the square
root of 62?); an example of automatic information processing is our rapid
response to single-digit multiplication problems (What is 5 times 5?). While
it is difficult to pin down all the criteria separating these two forms of
thought, we think that the distinction is important.
58 Background
This difference granted, we should not make the mistake we have
already exposed, in our discussion of point (2), of confusing process with
content. As the multiplication problem indicates, we can automatically,
effortlessly process many sorts of verbal formulae, such as the times table,
jingles, poems, cliches, stories, and oft-heard rules. In fact, knowledge of
these "verbal molecules'* (Strauss 1992b) is better explained by asso-
ciationist models of involuntary automatic information processing than by
models of consciously controlled information processing, because earlier
parts of the text are needed to trigger our memory of later parts. It is hard
to dip into the middle of a verbal molecule; instead we have to start from
the beginning and repeat the whole thing until we get to the relevant part.
Quick: how many days are there in November? Arriving at the answer to
this is not effortful in the way figuring out the square root of 62 is effortful;
it simply requires time to let each part of the formula, "Thirty days hath
September . , ." activate memory of the next. (See chapter 5 for further
discussion of this particular cultural mediating device.) This is one reason
why the term "symbolic processing" is misleading; connectionist models
have no difficulty processing symbols. The difference between them and
so-called symbolic processing models is not whether they can process
symbols, but that the latter models treat symbol processing as being like
deliberate conscious thought, while connectionist models treat this task as
being like an automatic pattern-completion task.
6, The neural basis of cognition. Earlier we said that connectionist
modelers use a neural metaphor for cognition. Are they not, instead, just
describing how cognition really works, because knowledge really is stored
in neurons? But connectionist computer modelers do not aim to model
brains in all of their details. A "unit" in a connectionist model might stand
for the activity of thousands of neurons. And connectionists ignore some
features of the brain altogether (for example, the disparate effects of the
dozens of different neural transmitters). This is why connectionists call
their models "neurally inspired" only (Smolensky 1988b).
processing modelers also agree that knowledge is stored in brains, and they
agree with connectionists that the brain consists of neurons arranged in
layers, processing in parallel, and so on. However, symbolic processing
artificial intelligence modelers think the action of neurons passing on
excitatory and inhibitory messages implements a mental language that
looks like any standard computer programming or natural human lan-
guage (i.e., symbols arranged in sentences).
Connectionist and symbolic processing modelers also differ in whether
they care about the ultimate translatability of their models to brain states
and processes. Symbolic processing modelers do not even attempt to
explain how representations and processes of the sort they describe could
take place in real brains; connectionists do make this attempt and in
Schema theory and connectionism 59
general outline, at least, it corresponds to the way brains work (Church-
land 1995).
Where we stand
Sometimes humans do engage in conscious, deliberate problem solving.
This is usually thought to be difficult to model in the connectionist para-
digm, we have noted, and may be a serious drawback of such models. On
the other hand, this sort of effortful thought is not central to our focus in
this book on cultural meanings. (Also - as we will explain in chapters 4 and
5 - even deliberate conscious thought still draws on understandings
learned in the way connectionists describe.) Interpretations, such as the
ones you had in response to the two beer commercials described at the
beginning of this chapter, typically arise automatically; they are not pon-
Also, the general view that knowledge is internalized as sentences in the
head leads us to imagine cultural knowledge and meanings as inflexible
and may be in good part to blame for the current generation of anthropol-
ogists' rejection of the relevance of internalization. Connectionist models
help us sort out what is likely to be stable, what changeable, in our
knowledge and meanings. They are thus a very promising foundation for a
solution to the problems of cultural meaning we are dealing with here.
Yet, while we find connectionist models stimulating because they pro-
vide a comprehensive framework for understanding important properties
of everyday human cognition, in the rest of the book we will not hesitate to
depart from such models when they favor computational ease and effi-
ciency over neurobiological, behavioral or social veridicality.
puters do not operate the same way as embodied and culturally embedded
people, and even connectionist modelers grossly oversimplify culture
learning and the complexities of human motivation. Some of these criti-
cisms are included at the end of this chapter; other amendments are
proposed in the next section of the book.
This brief introduction to connectionism may indicate why we find it
useful for a theory of cultural meaning that accommodates both older and
newer views of culture and of meaning. Readers who do not want to learn
about connectionism in any greater depth than this are encouraged to skip
ahead to the Final Comments section at the end of this chapter and then go
to part II, where we begin to draw out the implications for thinking about
culture. In the course of that discussion of culture, we will have occasion to
provide examples that should make our model clearer. Some readers,
however, may have found the preceding discussion too vague. For such
readers, a fuller, somewhat more technical explanation follows.
60 Background
Connectionism for the somewhat more formally minded
The first thing we have to make clear is that there are many types of
connectionists - and several intermediate points between formal modelers
who clearly are connectionists and those who clearly are not, (The follow-
ing discussion applies only to people doing artificial intelligence modeling;
people like us who use it heuristically without doing any formal modeling
could be called "connectionist inspired.")
The prototypical connectionist constructs models of cognition in which:
the building blocks are "units," each of which is activated by
the environment or other units and either passes activation on
to other units or is potentially part of the output representation,
or both;
units are connected by "weights," numbers which give the
association (positive or negative, great or small) between
and which express most of the current knowledge of the
system (i.e., its disposition to react one way rather than another
in a particular context);
concepts to be learned are "distributed," i.e., represented as
patterns of activity over a set of units rather than by symbols,
and many different concepts can be represented by patterns of
activation over the same set of units;
propositions are likewise represented by patterns of activa-
tion over many units, patterns that do not have a syntactic
information processing occurs both serially (one step at a time)
and "in parallel" (simultaneous multiple actions);
decisions are reached through a process of finding the best fit to
a number of soft constraints rather than following invariant
9 the system builds up its knowledge by learning associations
(positive or negative correlations) among the features of a
number of specific cases, rather than by being taught any ex-
plicit rules; and,
the system begins with weights of zero or randomly chosen
small connection weights and the same general-purpose learn-
ing algorithm for all computations.
The now classic presentation of work of this sort, especially of the more
effective versions of this approach sometimes called the "new connection-
ism" (Quinlan 1991), are the Parallel distributedprocessing volumes edited
by David Rumelhart and James McClelland (1986).
Schema theory and connectionism 61
The opposite pole (that inspired by the language metaphor), is typically
called "symbolic processing" or "classical" or GOFAI (for Good Old
Fashioned Artificial Intelligence) or sometimes, misleadingly, "computa-
tional" theories of mind.
In these models:
concepts are represented by symbols, which are the basic cogni-
tive building blocks;
propositions are represented by strings of symbols that have a
syntactic structure and information processing is highly sensi-
tive to this syntactic structure;
information processing occurs serially;
the system learns by being programmed with explicit rules and
these rules are then represented as such in the system and
comprise its knowledge (i.e., its disposition to respond one way
rather than another in a specific case).
Allen Newell and Herbert Simon (e.g., Newell and Simon 1972) and Jerry
Fodor (e.g., Fodor and Pylyshyn 1988,1995) are among many well-known
proponents of this approach.
These distinctions are not hard and fast: between these extremes, for
example, are models that most people would call connectionist but that
have much more than the usual amount of built-in knowledge (e.g., Regier
1996) or models that are basically symbolic but process information
through spreading activation (e.g., Collins and Loftus 1975) or allow for
"fuzzy" outcomes or parallel processing, for example.
As is the case with
any concept, the concept of "connectionist model" is somewhat messy. It
does not have sharp boundaries and cannot be defined by a list of necess-
ary and sufficient features. Instead, there are family resemblances among
models usually considered to be connectionist, and whether any particular
model is included in the family depends upon how many and which
connectionist features it has. An example should help clarify the general
difference between classical and connectionist approaches in their architec-
tures, ways of processing information, and learning procedures.
Architectures and representation
Connectionist models have a very different architecture from classical
models. Contrast, for example, the following two ways of depicting the
schema US Americans rely on in deciding which term (e.g., "Ms. So-and-
so," "Aunt Rachel," "Rachel," and "Miss") to use in addressing another
person. Figure 3.1 presents Susan Ervin-Tripp's (1969) classical model;
figure 3,2 presents our connectionist model.
Ervin-Tripp's model consists of boxes, diamonds, and arrows arranged
Figure 3.1 A US American address system - classical model
in the flow-chart fashion of traditional artificial intelligence programs. In
figure 3.1 the diamonds indicate features to be considered in deciding on
the correct term of address (e.g., Is the addressee an adult? Is this a status-
marked setting?); the arrows explain the subsequent features to be con-
sidered (if yes [+], then consider . . .; if no [-], then consider . . .); and the
boxes on the right-hand side are possible decisions or outputs. (FN = first
name; LN = last name; the slashed zero = avoid using any name.)
Figure 3.2 depicts a prototypical "feedforward" connectionist network.
In this model, like most connectionist models, there is a row of circles on
the bottom (it could also be the left) representing the input units, those
activated by features of the event or object. There is another row of circles
at the top (or it could be the right) of the picture, representing the output
units, that is, the possible decisions of the model. In between is a row of
circles labeled H
. These are hidden units, which receive activations
from the input units and pass them on to the output units. Before the
program is trained, the hidden units do not represent anything. After
training, the hidden units still do not represent any one namable thing, but
each is activated more strongly by some recurring patterns of inputs than
by others. Thus, unlike symbolic-processing models, there is no one-to-one
correspondence between symbols in the program and concepts to be
learned by the program. Instead, there are only learned, varying probabili-
Schema theory and connectionism
title &
last name
no name
or ask
context ego
enhances wants
closeness to stress
& closeness
similarity &
Figure 3.2 A US American address system - connectionist model
ties that one group of units, when activated, will excite another group of
units beyond their thresholds and activate them. As Rumelhart and Nor-
man (1981:3) put it, on such a model "information is better thought of as
'evoked' than 'found'" (quoted in van Gelder 1991:46).
The lines between these circles indicate paths of activation from the
input units to the hidden units and from the hidden units to the output
units. For each line, there is a numerical weight that stands for the strength
of the positive or negative connection between the two units. For example,
a large positive weight between the first kin unit and Hj would indicate that
if the first kin unit were activated, it would strongly excite H
making it
more likely to be activated past its threshold. A large negative weight
between the first kin unit and Hj would indicate that if the first kin unit
were activated, it would strongly inhibit H
making it less likely to be
activated past its threshold.
Models with the feedforward design are generally used for problems in
which the task is to figure out the appropriate response to some input.
There could be and often is more than one layer of hidden units or there
need not be any layer of hidden units. However, models with just two
64 Background
Figure 3.3 Interactive activation model
layers, one for input and the other for output, do not have as much
computational power as multilayered models. (In fact the limited capabili-
ties of two-layer models helped kill the whole approach for a while in the
artificial intelligence community, Rumelhart 1989.)
Many other designs are possible. Very often there are arrows between
units in the same layer. There can also be connections from higher layers
back to lower layers. Figure 3.3 is an example of one such, more complex
model: McClelland and Rumelhart's interactive activation model of letter
and word recognition (1981; see also McClelland, Rumelhart, and Hinton
1986). The triangular arrow heads indicate excitatory (positive) connec-
tions, the ball heads indicate inhibitory (negative) connections.
Processing in a model of this design will take place over several stages, as
some units gradually become more strongly excited and others inhibited
when the activations received from higher and same-level units start kick-
ing in. In this example, seeing the letter parts represented by the units in the
input layer at the bottom activates the hidden units representing letters in
Schema theory and connectionism 65
Figure 3.4 Recurrent network
the middle. Activation of those units, in turn, can be thought of as exciting
the output units for words that contain those letters and inhibiting both
the output units for words that do not contain those letters and the units
for other letters. And so on. This is more typical of connectionist programs
than simple feedforward models and probably more typical of actual brain
processes as well.
Some models use the outputs at one time as inputs for another phase of
processing and look something like figure 3.4. This structure is useful for
continuously monitoring the context and using that information to deter*
mine the final output (Jordan 1986).
However, as connectionist modelers sometimes complain, drawing
models with circles and arrows is not enough to make you a connectionist.
The way such models process information and learn is what makes them
distinctively connectionist.
Information processing in a good-old-fashioned artificial intelligence pro-
gram of the sort modeled by Ervin-Tripp takes place serially (one step at a
time) as we can see from Ervin-Tripp's example. As figure 3.1 indicates, the
program considers first if the addressee is an adult, then (if yes) is this a
status-marked setting, where any hierarchical differences between speaker
and addressee are highlighted or (if no) is the addressee's name known, and
so on, one feature at a time. Under most circumstances this model works
perfectly well. However, it does not perform very well if some of the
information we gave it was wrong (it would take the wrong turn and
probably end up at a different decision) or missing (it would stall at the
66 Background
node where no information was available). This, of course, is precisely
what most computer programs do because they are constructed along
conventional artificial intelligence lines, and it is one of the things that
makes them so frustrating to use. Finally, although this is not so much a
problem with Ervin-Tripp's model, some classical models do not handle
novel situations very well.
Connectionist models, on the other hand, are much more naturalistic in
the way they deal with missing, incorrect, ambiguous, and novel informa-
tion. In these cases, the model still proceeds to make a guess on the basis of
the information it does have and, if it has been well trained, many times the
guess will be right (a property called "graceful degradation"). This is due,
in part, to the fact that in connectionist models all of the relevant features
of the situation are processed at the same time (parallel processing) and in
part to the way each feature affects the decision the model reaches. In our
example, the relevant features are those described at the bottom of figure
3.2: whether there is a kinship relation between ego (the speaker) and alter
(the addressee); ego's age and local social status; alter's age and local social
status; alter's general approachability (Is alter typically friendly and infor-
mal or aloof and formal?); how close ego feels to alter; and two features of
the particular context: whether the situation at hand is one in which status
differences or lack of cordiality are enhanced or diminished and whether in
this situation ego wants to emphasize or downplay any differences in status
or closeness of the relationship (Brown and Levinson 1987).
Also im-
portant here is whether ego is assertive or shy, but we will work that into
the model in a different way, to be described a bit later.
Each of these attributes is represented by its own set of input units, with
different patterns of activation in each set corresponding to the possibili-
ties to be represented. For example, with the two units representing alter's
approachability, full activation of the first could indicate someone who is
extremely friendly, full activation of the first and partial activation of the
second would indicate someone who is fairly friendly, and so on. In the
output layer are single units whose activations correspond to possible ways
one could address someone else: kin term; pet name (e.g., "Sweetie" or
"Darling"); first name (or everyday nickname) alone; title and last name;
or generic name used for strangers (e.g., "Sir," "Miss," or "Buddy"). If no
other output unit is sufficiently activated, the model reaches the decision to
either avoid using any term of address or to ask alter for their preference.
Of course, once any of these choices is made, there is a further decision
about which kin term, pet name, title, or generic term to use, but we can
imagine that there are one or more other models that simulate how that
further decision is reached.
What happens when you present this model with an actual example of
Schema theory and connectionism 67
someone? Let us say alter is a relatively friendly female professor who
appears to be in her mid thirties, ego is a twenty-year old male college
student taking a large lecture course with the professor for the first time,
and they run into each other while at the college but outside of class. In this
common situation should students address their professors with a title and
last name (Professor So-and-so, Doctor So-and-so, or Mister or Ms.
So-and-so) or by the professor's first name? There are probably some
colleges where professors prefer first names and some colleges where all are
addressed using last name and title, but at least at this point in the late
twentieth century and at the US universities with which we are familiar, it
varies from professor to professor (or sometimes department to depart-
The computations the model performs are pretty simple. First, the
appropriate patterns of input units are activated. Just focusing on the
status, closeness, and approachability units, these might be patterns such
as full activation of the first ego's status unit (ESj) and no activation of the
second (ES^) to represent the student's fairly low status in the college
pecking order; no activation of the first alter's status unit (ASj) and full
activation of the second (AS^ to represent the professor's fairly high
status in the local hierarchy; partial activation of the first and full activa-
tion of the second closeness units (C
and C,
) to indicate that they know
each other but are not especially close; and full activation of the first and
partial activation of the second approachability units (AP
and AP
) to
indicate the student's perception that the professor is a reasonably friend-
ly, informal person. At the same time other input units are being activated:
those representing the context (one in which their normal relationship is
emphasized and the student has no need to act more formal or less than
usual); the student's young-adult age; the professor's early-middle age; and
so on. Let us give each fully activated input unit a value of one, each
partially activated input unit a value of 0.5, and each nonactivated unit a
value of 0.
For every activated unit, its activation value is then multiplied by the
connection weight between that unit and each of the others with which it is
connected. We have said that the connection of a unit with each unit it
activates is represented by a numerical weight that stands for (roughly) the
strength and positive or negative association between the two units. Al-
though no two people's knowledge will be modeled with exactly the same
weights (and a given person's weights will change while they are learning
the system, as we will explain in the next section), we can make some
plausible guesses about the weights a United States undergraduate at the
time we are writing might have for the portion of the model representing
ego and alter's status, their closeness, and alter's approachability. (We will
68 Background
Table 3.1 Weights in connectionist model of US terms of address -first
-0. 3
-0. 9
-0. 9
-0. 8
-0. 2
-0. 2
-0. 2
-0. 2
-0. 2
-0. 4
-0. 6
-0. 2
-0. 2
-0. 4
-0. 6
-0. 8
ignore connections to the possible choices of a kin term or pet name, which
are highly unlikely in the present example.)
Since it would be hard to fit all of these numerical weights into the
appropriate places in figure 3.2, we have presented them in a more conveni-
ent, easy-to-use matrix (table 3.1). The -0.1 in the top, left-hand corner
stands for the connection weight between the first Ego's Status unit and the
first hidden unit; the 0.1 to its right stands for the weight between the first
Ego's Status unit and the second hidden unit; and so on. Activation levels
for the particular example we gave of the young college student trying to
decide how to greet his thirty-five-year-old professor are given to the left of
the input unit's abbreviation.
Now we can begin to see how the model will simulate what happens in
this case of a student greeting a fairly friendly lecture-class professor. First
we have to calculate the combined effects of the student's status, the
professor's status, their closeness, and the professor's approachability on
each of the hidden units,. This requires, for each input unit, multiplying its
activation values (0, 0.3, or 1) by the weight connecting that unit to the
relevant hidden unit and then summing all those products.
So, for
example, the effect of the first input unit (ESi) on the first hidden unit is 1
(the activation level of ESj) times -0.1 (the weight of the connection
between these units), which comes to -0.1. To put it another way, activa-
ting ESi slightly inhibits H
. ES
and ASi are not activated in the present
example, so they do not affect the activation of H
, but AS
is and
contributes further inhibition (lx-0.3 = -0.3), which is offset by the excita-
tory contribution of C
(0.5 x 0.7 = 0.35). And so on.
Readers who are numerophobic should keep in mind what is theoreti-
cally interesting about this math. The weights in the body of the matrix
represent knowledge the model has accrued over time - these are its
general understandings or schemas at the moment. The activation levels of
Schema theory and connectionism 69
the input units represent the particularities of this encounter. By multiply-
ing weights by activation levels the model is simulating the regulated
improvisation of people who bring their learned schemas to bear on a new
situation. In this case, the simulation goes as follows (multiplying activa-
tion levels by connection weights, in turn, down the columns)
Input to Hidden Unit l =(i x-0. 1)+(0x-0. 1) + (0x-0.1) +
(lx-0. 3) + (0.5x0.7) + (lx~0.9) + (lx0.1) + (0.5x-0.9) = -1.3
Input to Hidden Unit 2 = (1 x0.1) + (Ox-0.1) + (0x0.1) +
(1 x -0.1) +(0.5 x 0.8) + (1 x -0.9) + (1 x 0.7) + (0.5 x -0,8) = -0.2
Input to Hidden Unit 3 = (1 x-0.1)+ (0x0.2)+ (0x0.2) +
(1 x -0.2) +(0,5 x 0.4) + (1 x -0.2) + (1 x 0.8) + (0.5 x -0,2) = 0.4
Input to Hidden Unit 4 = (1 x 0.2)+ (0x0.2) +(Ox 0.1) +
(1 x-0.2)+(0.5 x 0.1) + (1 x-0.2)+ (1x0.7) +(0.5 x-0.1) = 0.5
Input to Hidden Unit 5 = (1 x -0.4) + (0 x 0.4) + (0 x 0.6) +
(lx-0.6)+(0.5x-0.2) + (lx0.5) + (l x-0. 2) + (0.5x-0.1) =
Input to Hidden Unit 6 = (1 x 0.6) + (0 x-0.4) + (0 x-0.7) +
(1x0.9) +(0.5x-0.6) + (1 x0.6)+ (l x-0. 8)+ (0.5x0.7) = 1.35
Now we repeat this process to figure out the combined effects of the hidden
units on the output units, which are the units that simulate the model's
decision about what to do in this case. For this next step, the knowledge of
our simulated typical undergraduate - that is, the weights connecting
hidden units to output units - is captured in table 3.2. If we say that
activation values can range only from 0 to 1 (other choices are possible but
this will be the way we do it here), then negative total inputs mean an
activation of 0 and positive inputs higher than one are capped at 1.
Activation levels between 0 and 1 are the ones computed in the last step.
Table 3.2. Weights in connectionist model of US terms of address - second
No name
(or ask)
70 Background
Based on these weights and activation levels, the computation would be
as follows:
Input to First Name = (0 x 0.8) + (0 x 0.9) + (0.4 x 0.8) +
(0.5 x 0.4) + (0 x 0.2)+ (l x-0. 4) = 0.12
Input to Last Name = (0 x -0,9) + (0 x -0.9) + (0.4 x -0.6) +
(0.5 x -0.4) + (0 x -0.2) + (1 x 0.9) = 0.46
Input to Generic Name = (0 x -0.9) + (0 x -0.9) + (0.4 x -0.9) +
(0.5 x -0.6) + (0 x 0.2) + (1 x 0.2) = -0.46
Input to No Name = (0 x -0.9) + (0 x -0.9) + (0.4 x -0.4) +
(0.5 x 0.2) + (0 x 0.2) + (1 x 0.4) = 0.34
If the output units for terms of address had a threshold of 0, that is, they
were turned on if they attained any positive value and turned off if they had
any negative value, then the only decision the model would reach on this
basis of this input would be to avoid a generic term of address, like
"Ma'am." So we need two modifications of this simple model. First, we
will stipulate some threshold values that activation levels have to reach for
that interpretation or behavior to be chosen. This is where the student's
general level of assertiveness can be modeled easily. Suppose, because he is
a bit shy, his general preference in all these cases is to stick with the safer,
more respectful title-plus-last-name choice rather than risk the rebuff that
might result from addressing people by their first name. That general
preference can be simulated by a higher threshold on First Name than on
Last Name, perhaps 0.3 for Last Name and 0.5 for First Name. In other
words, the combined inputs to the Last Name unit only have to be 0.3 or
higher for that unit to be activated but they need to be 0.5 or higher for the
First Name unit to be activated. If we add that the threshold for the
Generic Name unit is 0.2 and the threshold for the No Name unit is 0
(which is realistic, because this is what you do if you do not know what else
to do), now we have, in the case of our somewhat reticent student greeting
the fairly friendly lecture-class professor, two units that are activated past
their thresholds: the Last Name unit and the No Name unit. Since the Last
Name unit is activated more strongly than the No Name unit, that one
should prevail. To model that we need to add some inhibitory links among
the output units, modifying figure 3.2 (see figure 3.5).
The effect of these links is that if the First Name or Last Name units are
activated past their thresholds, they send large negative inputs to the other
output units, effectively turning them off. The same goes if the Kin Term
and Pet Name units are activated past their thresholds (remember - we left
them out of this example only because they were extremely unlikely terms
of address for a student greeting a teacher). There are weak negative
~ Inhibitory link
Figure 3.5 Modified connectionist US American address system
72 Background
weights among the first four output units as well, so that in the end, the
most strongly activated output unit will win out. These horizontal inhibi-
tory links require more than a single processing cycle - just imagine that
the computations performed in the last step are repeated, adding excita-
tions or inhibitions from the other output units as many times as is
necessary until one of the output units is pushed past its threshold. In our
example, if we were to add an inhibitory link from Last Name to No
Name, the effect would be to leave the Last Name unit the only one
activated in the end, so that is the choice the student would make, at least
on the basis of the his status, the professor's status, their (low) degree of
closeness, and the professor's perceived approachability. Now the student
only has to decide whether to use "Doctor," "Professor," "Ms.," "Miss,"
or "Mrs.," which could be simulated by another model! (A friend of
Claudia Strauss's from India, conducting her graduate work in the United
States, complained about the complexities of the American terms-of-ad-
dress system, which required knowing such things as her female ad-
dressees* gender politics, hence whether they preferred Ms. to Mrs. or
Recall that to keep the math less cumbersome we did not take into
account all the relevant factors. What if the student were an adult return-
ing to college and about the same age as the professor? If the student and
professor were doing volunteer work together in town? Or if the student
wanted to ask the professor for a favor? Multiplying the activation of these
inputs by their weights would change the activation of the hidden units,
hence the answer the model would reach in the end. Or suppose that a year
later the same student takes another course with that professor, say, a
small upper-level class. By then he will probably feel much more at ease
with the professor, which could make him much more inclined to greet the
professor with her first name, and, in general, interpret their relationship as
closer than it had seemed before. If you want, you can try playing with the
model, changing activation levels or making up new weights, or changing
threshold levels to see the effects on the outcome. We tried it in the case of
the same student now taking a second, much smaller class with the same
professor (everything was the same except the activation level of C
was 0.5
instead of 1) and found in this case our shy student was unsure and avoided
using any term of address. Of course, if the professor simply announces at
the beginning of the semester, "Please call me Jane," or if there is some
well-established assumption shared by the students about whether to use
first or last names, then this model is not needed to simulate the process.
(Unless, as often happens, students take these factors into account as
influential but not decisive constraints on their behavior.) Kin terms, for
example, usually have well-defined conditions of use and so are not chosen
Schema theory and connectionism 73
through a process of this sort. (Fictive kin terms, on the other hand, might
be more a matter of choice, especially in some groups.)
From this example we hope it is clear that each of the factors detected by
the input units acts as a "soft constraint," influencing but not decisively
determining the outcome. These models can also reach a decision even if
some of the information is missing or incorrect: if most of the information
is close enough, the decision will be the right one for the occasion. Sup-
pose, for example, the professor appears to be thirty-five years old but is
really forty-five years old, or the student has no sense of the professor's
approachability. The model might still reach an appropriate decision
about which form of address to use. The model can handle novel situ-
ations: These situations simply activate the relevant input units - in this
case, a combination not previously activated - and it proceeds in the usual
way to reach the best answer it can. (On the other hand, just like human
beings, these models sometimes falsely generalize as well, assuming that
two things that are alike in respects that are known to it are also alike in
respects it does not know, a process that helps explain what psychoanalysts
call "transference," Andersen and Baum 1994.) We should also remember
that automatic procedures like these do not always come up with an
answer about what to do, and sometimes another part of our mind comes
in at that point, deliberately ponders the matter (maybe creatively combin-
ing elements of most strongly activated output possibilities),
and makes
a decision. We have already discussed this limitation of connectionist
models and will discuss it further in chapter 4.
We have said that one limitation of a classical model of cultural knowl-
edge, such as Ervin-Tripp's, is that it does not handle missing or incorrect
information very well. An even more important limitation, if we want to
understand where cultural models and meanings come from, is that it is
difficult to imagine how a carefully constructed decision-tree of the sort
pictured in figure 3.3 would be learned in real life. In the communities with
which we are familiar, nobody tells you things like, "If someone is fifteen
years or more older than you, don't call them by their first name." You
might be told, "Address your elders by their last name" or "That's
Grandma Hilda." You then have to figure out from that instruction, along
with what you see other people doing, how old someone has to be to count
as an "elder" and when you really do use a last name versus when you use a
kin term plus first name (as in "Grandma Hilda"). How do children go
from not knowing what to call anyone to knowledge that can be simulated
by a decision tree like Ervin-Tripp's but surely is not learned in such a
74 Background
logical form? Connectionist models are especially good at answering this
In the last section we saw that the way connectionist models respond to
input depends on the connection weights between the units. Almost every-
thing the model knows is in those weights, which determine how it reacts in
any given context. From our anthropological perspective one of the most
attractive aspects of connectionist models is that these weights are not
predetermined from the start by an omniscient programmer. Much as
children gradually acquire some of their dispositions to react in a culturally
appropriate way simply by observing and participating in daily routines,
so connectionist models gradually acquire their knowledge through expo-
sure to numerous examples.
In our example (from the last section) of a twenty-year-old male US
college student, some of his exposure came from people he learned to
address as he was growing up: his friends; other children; relatives, some
younger, some about his age, some adult; his friends' parents; his teachers;
and assorted other adults. At the same time, he was learning terms of
address by observing how it was done by other people he knew, as well as
how it was done by the characters he saw on television, in movies, and so
on. In these ways he had been exposed to hundreds of examples of people
addressing each other.
To simulate this learning, connectionist models are trained with a set of
inputs that, ideally, corresponds to the sorts of inputs a human learner
would encounter. In this example, we should divide these inputs by life
stages, thinking about different sorts of examples encountered by this US
college student when he was a preschooler; as he entered school and
greater demands were placed on him to respond appropriately to people in
authority and his media exposure increased; and as he entered his late
teenage years and had participated in or been exposed, at least vicariously,
to an even wider set of interpersonal encounters. Learning does not stop at
that point, of course; but in the present example, we do not need to
simulate the subsequent changes that the schema undergoes in adulthood.
To begin with, the model starts with weights of zero on all connections,
or with very small, randomly chosen positive and negative weights.
Imagine table 3.1 filled in with zeros or numbers like 0.02 and -0.01. It is
then presented with its first example: say, mother. The appropriate input
units are activated, responding to the fact that she is kin, ego is very young,
she is an adult, and so on. These activation values are then multiplied by
the initial weights and the products are summed, to compute the activation
of the hidden units and that process is repeated to arrive at the activation
of the output units. If we start with connection weights of zero, the end
product will be zeros for all of the output units, which is the wrong answer:
Schema theory and connectionism 75
kin term should be activated past its threshold and more strongly than all
the other possibilities. Even if we start with random, small weights, the
model is very unlikely to reach the right answer.
What happens next varies from model to model, but generally speaking,
a small amount is then added or subtracted from each weight that was
Weights are not fixed all at once to be the best ones for the
example just presented, only adjusted a little to nudge them in that direc-
tion. Typically, this learning is "supervised." This means that the pro-
grammer stipulates the correct output pattern for each input pattern and
adjusts weights to reduce the difference between a unit's actual output and
its desired output. Then the process is repeated for the next example in the
training set (e.g., father). Again, an output is computed on the basis of the
current weights in the model - the ones that were the result of the last
round of changes - and small changes, either additions or subtractions, are
made to the weights that contributed to the output. This process is repeat-
ed for the whole training set and the training set itself is typically repeated
for many "epochs" - sometimes hundreds of epochs. Although you might
think that the model would get confused having its weights adjusted first
for one situation, then for another, in fact for a surprising number of tasks
that connectionists have considered, at the end of this process it has
reached a set of weights that lets it produce not only the desired outputs for
the examples it was trained on, but also for a new set of examples similar
to, but not included in the training set.
If the learning process is divided
into stages, as would be realistic in our terms-of-address example, a new
(much larger and more diverse) training set would then be introduced that
contains examples of the sort a US school-aged child might encounter and
the process would be repeated. The process might be repeated a third time
with a further enlarged and still more varied training set appropriate to the
experiences of this boy as he reaches his late teens and early adult years.
We have said "a US school-aged child" but that can be made as culturally
specific as we want: for example, "a Mexican-American boy growing up in
Los Angeles at the end of the twentieth century." Note that there is no
assumption of cultural stability built in to the model. The boy could move
to a very different community or ideas about appropriateness could change
as he is growing up: the learning procedure needs only to change the
training sets to adjust accordingly.
The end of the process is that, without the model ever having been given
a rule such as, "Address people who are fifteen years or more older than
you by a title and their last name," it has a set of weights that will enable it
to act as if it knew such a rule (or whatever is the appropriate term of
address for elders in its experience). It should also have learned not to use
the same form of address all the time for a given person, but to vary that
76 Background
form as appropriate for the context. As Bourdieu puts it, it has acquired
"schemes enabling [it] to generate an infinity of practices adapted to
endlessly changing situations, without those schemes ever being con-
stituted as explicit principles" (1977:16). In his words again, it has acquired
a "sense" of the situation.
Anthropological critique
That does not mean that we are completely satisfied with connectionist
models of learning. While connectionists are starting to pay attention to
studies documenting the distribution of inputs to which a learner might be
exposed, they oversimplify the variety of ways in which cultural knowledge
is transmitted. To our knowledge the only attention that is paid to this
issue currently is in the form of discussion by connectionist modelers about
the relative merits of "supervised" and "unsupervised" learning.
In connectionist models that rely on supervised learning, as the model is
being trained, the output the model reaches is compared to the desired
output (i.e., what the programmer has decided in advance is the right
answer) after each presentation of a case. Then all the relevant weights are
adjusted up or down as needed to bring the model a little closer to the
desired output the next time it is presented with that input. As many
connectionists have noted, however, in many real-life situations there is no
teacher to correct every wrong guess we make about the world and reward
every right one and often it is difficult to specify in advance what the
correct output should be.
One alternative is an unsupervised model of learning, such as "competi-
tive learning." In the version of this described by Rumelhart and David
Zipser (1986) the basic idea is that at any layer from the first hidden layer
on up, units are arranged in clusters. Within each cluster the unit that is
activated most strongly by the input inhibits the activation of all the rest
until it is the only unit on. Then all the weights connecting that one unit to
units in the layer below are made a little greater and all the weights
connecting to other units in the cluster are decreased a bit, making it more
likely that the next time that stimulus is presented, that unit will be the one
most strongly activated. The result is that after repeated exposure to
different cases the units in a cluster get to be specialized, responding to
different combinations of inputs, without needing a teacher to say what
was the desired output (because changes are made solely on the basis of
which unit in a cluster was most strongly activated). There are other
unsupervised approaches as well that rely on different formulae for adjust-
ing weights (Hinton 1989).
In real life, however, just as most culture (and language) learning is not
Schema theory and connectionism 77
completely supervised, so most such learning is not completely unsuper-
vised. As Roy D'Andrade and others have pointed out, learning is typi-
cally partially supervised. Very often "you try some of it by yourself, and
other people help by giving occasional procedural advice and crucial
instruction in classification when you get stuck" (D'Andrade 1981:186).
D'Andrade calls this "learning by guided discovery" and claims it is the
most typical form of learning: "Looking at cross-cultural studies of sociali-
zation, one is struck with both the small amount of explicit step by step
instruction and the large amount of occasional correction that character-
izes cultural learning all over the world" (D'Andrade 1981:185). Thus,
adults and knowledgeable older children intervene occasionally, correcting
errors as necessary to keep neophytes out of social, intellectual, moral, and
physical danger; by their enthusiasms, classificatory terms, and topics of
commentary they draw the learner's attention to some features of the
world and away from others. The result is that for a great many tasks,
completely supervised connectionist learning procedures unrealistically
build in too much feedback while unsupervised connectionist learning
procedures unrealistically have too little.
Nor is the issue here one solely of quantity of feedback. D'Andrade's
discussion of guided discovery contrasts not only constant feedback with
occasional feedback but also explicit instruction with correction in specific
cases. Here connectionists might point out that they are more anthropol-
ogically correct than symbolic processors because connectionist models
learn not by being programmed with explicit rules but rather through a
gradual process of trial-and-error in response to specific cases. Yet, some-
times culture learning does proceed through explicit statements, as is
typical with rules of etiquette ("No elbows on the table"); maxims and
proverbs ("Feed a cold and starve a fever," "The squeaky wheel gets the
grease"); heuristics taught to a novice in some field ("Shift to second gear
when the car is going 15 miles an hour"); and certainly, narratives, abstract
theories, and sacred texts (D'Andrade 1995:144; Quinn and Holland
1987:22). The process is not always a simple one, however, that could
easily be modeled in the old-fashioned way by spelling out the knowledge
sententially. In part, the problem is one that connectionists have already
noted: The shift from a novice's explicit rules of thumb to an expert's
intuitions, for example, is not so much a process of learning more rules but
rather of building up an entirely different sort of knowledge base, one that
may not have any rules in it (Smolensky 1988b:5; see also Bourdieu
1977:19). But there is a further problem: explicit statements are sometimes
helpful and sometimes misleading or ambiguous, so that learning is neither
a process of blindly following explicit rules nor one of ignoring them in
favor of implicit learning from multiple trials. For example, some teaching
78 Background
merely states explicitly what learners could observe for themselves but
might not have noticed (or ignored, as a two-year old might with the rule,
"No gargling at the table"), while others state what turn out to be "com-
pany manner" rules that apply in statistically rare but socially important
occasions but not the rest of the time (for example, "No elbows on the
table," in many US households). Often rules are ambiguous as stated (for
example, "Address elders by their last name." Does that mean all elders?
How much older do they have to be?) and need to be supplemented by
Another form of cultural learning- an extremely important
one - proceeds by naming an abstract entity, e.g., honor or love, that
learners then can learn to identify and acquire a fuller knowledge of
experientially (D'Andrade 1984:91; Quinn and Holland 1987:3; Strauss
and Quinn 1992).
By the same token, even implicit learning through exposure to multiple
examples is more complicated than depicted in connectionist models at
present. For one thing, learners have to keep straight the difference be-
tween what is presented as normal in books and other media (e.g., house-
holds with one adult male, one adult female, and one or more children, the
adults being parents of the children and married to each other) and what is
statistically typical in the child's experience (e.g., blended or single-parent
families or households that include other relatives, friends, or lovers,
Strauss and Quinn 1992).
For another, examples do not all have equal
salience. Social learning theorists for example (Maccoby and Martin
while stressing observational learning, note the importance of
selective attention in this process. Sometimes attention is directed toward
people who control desired resources (e.g., mothers, for young children),
sometimes to the behavior of people like the learner (e.g., adults of the
same sex as the learner), and sometimes to the behavior of people whom
the learner aspires to be like or expects to be like someday (e.g., adults of a
higher class or of a society to which one has emigrated) or the behavior of
someone who is funny or otherwise attention-getting (as Claudia Strauss
has observed with her four-year-old son's partial adoption of the speech of
Teenage Mutant Ninja Turtles), This process could be simulated by craft-
ing sets of inputs that include only the most salient examples for a particu-
lar sort of learner, but it would be more interesting if connectionists were
to try to model how learners learn to attend to some things and not others,
paying particular attention to the role of emotional and motivational
Sometimes examples contrary to what one expected are not just
one more example to learn from but are so surprising that they are
especially memorable. As we will explain later, connectionist models are
often insufficiently biological in not giving more weight to affect-laden
learning, as learning something that is shocking or surprising may be. And
Schema theory and connectionism 79
it would help if connectionists would model social knowledge much more
than they do, which would also force them to stop positing characterless
generic learners.
What all of this points to is the need for connectionists to work with
field workers who will study the amount and kinds of explicit teaching that
learners in different communities receive for a given task, as well as the
particular examples learners are both exposed to and notice, especially for
culture learning.
Beyond the tabula rasa
A more controversial issue at present is whether connectionists posit too
little innate knowledge. For most cultural anthropologists this is not a
problem. As cognitive anthropologist Lawrence Hirschfeld (1988) notes
(citing Quinn and Holland 1987, among many others), a central dogma in
cultural anthropology is that human behavior is predominantly learned.
In response to this central dogma, Hirschfeld (1988; Hirschfeld and Gel-
man 1994) and some other cultural anthropologists (e.g., Atran 1990,
Brown 1991) have joined a large and perhaps growing number of practi-
tioners from other fields in suggesting that much more human knowledge
is innate than anthropologists had previously supposed. From their per-
spective, the fact that connectionist models start with little innate knowl-
edge is a serious drawback, rather than the virtue we have made it out to
This criticism of connectionist models misunderstands the extent to
which they do contain what we could call innate knowledge. While the
prototypical connectionist model does not include any explicitly pro-
grammed rules, it is not a blank slate before it is trained. As some
commentators (e.g., Lachter and Bever 1988) have pointed out, all connec-
tionist models start with a set of output units that represent whatever
answers correspond to the task the model is supposed to learn, a set of
input units representing those features of the stimulus that, the program-
mer decides, are most relevant for doing the task, and nonrandom connec-
tions among these units.
These connections and the choice of indicators
to be represented by the input units are extremely important constraints
that are built into the program. Thus, in our terms-of-address example
earlier, we decided in advance that the relevant inputs were ego and alter's
age, ego and alter's local status, the closeness of their relationship, and so
on. We ruled out such inputs as the weather at the moment or ego and
alter's hair color (Hirschfeld and Gelman 1994:11). Less trivially, we also
did not include features that might be relevant in another society. Among
the Wolof of Senegal, for example, greeting routines are affected not only
80 Background
by the relative status of the two people interacting but also by whether
their kin groups stand in a "joking relationship" (Irvine 1989). If a connec-
tionist model has input units for the features relevant in no more than one
society, it starts out more highly constrained than people do.
At this point, if not a lot sooner, many anthropologists might say, "This
is exactly why we do not find artificial intelligence very useful. Why bother
studying what computer programs can and cannot do, when they are so
different from human beings?" But artificial intelligence modeling issues of
this sort can actually lead in a fruitful direction for anthropologists be-
cause they could force us to consider whether learning cross-culturally
variable social knowledge is facilitated by any universal propensities to pay
attention to the correlations in one's own society among a limited, cross-
culturally recurring, set of possible features of the situation.
Social status and interpersonal closeness might be two examples of such
cross-culturally relevant features. Roger Brown and Albert Gilman argued
many years ago, in "The pronouns of power and solidarity," (1960) that in
Romance languages at least, choice between informal (e.g., tu) and formal
(e.g., vous) second-person pronouns depends on the relative status and
closeness of the speaker and addressee. More recently, Penelope Brown
and Stephen Levinson (1987) argued that many features of speech in a
wide variety of languages are affected by power/status differences and
closeness (as well as a feature of the context they called "weightiness of the
imposition" - a feature that is particularly relevant to situations in which
the speaker has a favor to ask the addressee). The same features of
closeness and relative status can also be found, according to Geoffrey
White (1980), as underlying the similarity judgments of personality-trait
descriptors made by speakers of three disparate languages (A'ara, spoken
in the Solomon Islands of Melanesia; Oriya, spoken in Orissa, India; and
US English), further evidence that we may be prewired to recognize these
features of social relationships. It also strengthens the argument for innate-
that many species of animals can recognize their place in the local
"pecking order" and tell the difference between friends and strangers (e.g.,
Cheney and Seyfarth 1990).
A similar recognition of probable universals helped Terry Regier with
the problem of writing a connectionist program that could learn how to
use spatial relation terms like the English words "above," "on," and
"through," for any natural language.
In English, relative position and
path of a trajector (movable item of interest at the moment) in relation to a
landmark (stable point of reference in the scene) are the only relevant
features. We say, "He is on top of the hill" and "The kite was on the roof of
the house," using "on" in both cases because both of the trajectors (he and
the kite) are above, but in contact with and supported by, the landmarks
Schema theory and connectionism 81
(the hill and the roof). In other languages relative location is not all the
learner needs to notice. For example, in Mixtec (spoken in Oaxaca,
Mexico) spatial relations are described with bodily metaphors and the
shape of the landmark is important. So He is on top of the hill is expressed
in terms of being at the hilTs head, while / was on the roof of the house is
expressed in terms of being at the house's back (Lakoff 1987:313-4,
drawing on Brugman 1983, 1984). In other languages (e.g., German) still
other features (e.g., orientation of the landmark's surface) are important
(Regier 1995). So a program that simulates the ability of a young human to
learn correctly how to use the spatial relation terms of their native lan-
guage cannot be restricted to input units for features that are relevant in
one language only.
Regier's solution to the problem of learning these spatial relation terms,
given all this potential variability, was to give the model a head start by
building in some plausible universal feature detectors. His model has units
sensitive to different orientations of the landmark and of the trajector in
relation to landmark. That is reasonable: there is very solid neurobiologi-
cal evidence for specialized cells that respond to a given orientation of a
visual stimulus. Regier's model also has units that discriminate topo-
graphical relations (e.g., trajector is inside landmark versus outside land-
mark). Here, too, he can draw on evidence for brain cells that respond to
center-surround and other topographic relations (cited in Regier 1995).
And his model has special-purpose procedures for remembering these
orientational and topographic relations at the beginning and end of a path
if the trajector is in motion, as well as calculating average, largest, and
smallest values of relevant states during the path (for using terms like
"through"). With these specialized systems in place, Regier's model was
able to learn spatial terms from English, German, Japanese, Russian, and
Mixtec (not as exotic a collection as we anthropologists might like, but still
The paradoxical moral of the story may be that we need to postulate
some human cognitive universals in order to explain human learners'
ability to acquire widely variable cultural knowledge. This is not such a
novel suggestion: Noam Chomsky has argued the same point for many
years with respect to language learning.
But connectionist modeling like
Regier's may show how it is possible to start with only a small core of
evolutionarily plausible mechanisms
and learn the rest. We probably do
not need very detailed innate structures; just enough to give us a start in the
right direction. Indeed, the remarkable ability of brain cells to respond to
different features of the world, depending on the learner's history, suggests
we should not go too far in assuming hardwiring (Merzenich and
Sameshima 1993). It may be that we are born with propensities to attend to
82 Background
and represent certain features of the world, but these initial propensities
are only neural first guesses that can be modified with experience. If that is
the case, one problem with connectionist models as they stand now would
be not that they are underconstrained but that they are overconstrained
because their input units have fixed representations.
Final comments: symbols and meanings
We hope it is clear now what it means to say that connectionist models are
an alternative to "symbolic processing" models. Connectionist models
certainly can learn to interpret culturally meaningful symbols. In these
models, however, there is no one-to-one correspondence between the
concepts to be learned and mental symbols that represent them;
or ideas are not mentally represented as strings of such symbols (Strauss
1988); and our involuntary forms of thought are not simulated on the
analogy of deliberately talking our way through a problem, one sentence at
a time.
Furthermore, modeling cognition connectionistically gives us a new way
of thinking about meaning: one that is responsive to both postmodern
critiques of static cultural meanings and old-guard warnings that cultural
meanings do not change continuously or whimsically. We said in chapter 1
that meaning is "the interpretation evoked in a person by an object or
event at a given time" and a "cultural meaning" is the typical interpreta-
tion of objects or events evoked in people who share life experiences. We
can put those claims more precisely now, as follows: the meaning of an
object or event for a person at a given time is the output of something like a
connectionist network.
The "cultural meaning" of an object or event is
the typical output of the networks of people who have similar histories.
There is no other meaning for that thing, no essential meaning floating in
the ether somewhere. Nor does meaning rest in an abstract symbolic
structure, although these structuralist claims about the importance of the
whole do make sense if reanalyzed in connectionist terms, for the interpre-
tation of any one thing depends on the weights in an entire cognitive
network (see Hinton et al. 1986). Finally, while poststructuralists are right
that meanings are never decided for all time, we are constantly producing
meanings based on whatever schemas - however incomplete or partial - we
have at the moment.
Thinking of personal and cultural meanings that way has several advan-
tages. A few can be outlined now. Thus, the contextually variable, change-
able nature of meanings can be explained. To take a famous Geertzian
(and before him, Rylian) example: what is the meaning of an eye blink? If
this gesture were the input to a connectionist network, its output (the
Schema theory and connectionism 83
interpretation of one eye closing) would depend very much on all the
features of the situation. If the gesture were accompanied by a sly smile, it
might be interpreted as meaning "We're in this conspiracy together"; if it
were followed by eye rubbing, it might instead evoke the interpretation
"There is something in that person's eye." Interpretations also depend on
the learner's history of experiences and can change over time. To return to
our example of learning forms of address, a student could initially take a
number of courses in a department where older professors like to be
addressed formally and younger ones informally, but then take a number
of courses from professors who expect to be addressed informally in
seminars but formally by students in large lecture classes. These changing
experiences will change the connection weights in the student's model for
using terms of address, changing the interpretation, i.e., the effects, of an
indicator such as age. At the same time the professor in this example is
building up a network that interprets the terms being used to address her.
Over time, this professor's interpretive model can change too. Perhaps
once only advanced graduate students first-named her but now she is
routinely first-named by all students. Whereas before an undergraduate's
calling her by her first-name would evoke an unpleasant feeling of not
receiving sufficient respect, she now interprets that informal mode of
address differently. To the extent that the changes in her experience are
shared by other professors, we can say that the cultural meaning of
first-naming (in this situation) has changed.
It is also easy to understand social variation in meanings when they are
modeled this way. To the extent that different subcultural groups have
different typical experiences, their cognitive networks will develop differ-
ently, and the interpretations evoked in them by a given object or event will
diverge. There are also likely to be subcultural differences - some large,
some subtle aspects of the context - in the objects and events being
interpreted. To put it in terms of the model: There are likely to be
subcultural differences, indeed, differences within any group, no matter
how small, both in connection weights developed over time and input
features at any one point in time.
On the other hand, this model for meaning gives us new ways of
picturing the possibility of somewhat stable and shared meanings. Thus,
situations can change, activating different combinations of input units,
without the underlying knowledge structure changing. That is why we
have stressed the difference between meaning and schemes, where the
former is one's interpretation of a particular situation and the latter are the
learned patterns of connections among units, different parts of which will
be activated in any given situation. Similarly, it is possible that networks of
connections might overlap substantially across subcultural groups and
84 Background
that some of what we call intracultural variation could be due to differen-
ces not so much in schemas as in the combinations of inputs that are most
typically activated. This model allows us to think about meanings that are
rooted in experience without ever hardening beyond all possibility of
change and are somewhat shared by different groups of people, but only to
the extent that their experiences in the relevant domains are shared.
The implications of our connectionist-inspired model for a theory of
cultural meanings will be elaborated, with many more examples, in the
next part of the book. Not that we will be restricted to a connectionist
paradigm. As we have stated more than once, connectionism is useful as a
way of showing that meanings can be mental while being learned from and
sensitive to the public world. But it will take other psychological and social
theories to explain fully how meanings become motivating (or not), per-
sistent in people and communities (or not), spread thematically (or not),
and widely shared (or not). In the chapters that follow we will show, for
example, that cognition cannot be considered apart from emotion and
motivation; that both implicit learning and explicit instruction are import-
ant for culture acquisition; and, finally, that just as an emphasis on public
culture alone is inadequate, so is a consideration of cognition without
discussion of its interactions with the world outside the mind. As we
elaborate our arguments we will also indicate the limitations and indeter-
minacies of connectionist models. We will amend these models as necess-
ary, suggesting, for example, how emotional arousal can strengthen neural
connections and how control processes can override responses that other-
wise might be activated by connectionist networks. It is finally time to start
analyzing the centrifugal and centripetal properties of culture - the prob-
lem with which we started.
Part II
Implications for a theory of culture
Cultural understandings have five centripetal tendencies. First, they can be
relatively durable in individuals. Secondly, cultural understandings can
have emotional and motivational force, prompting those who hold them to
act upon them. Thirdly, they can be relatively durable historically, being
reproduced from generation to generation. Fourthly, they can be relatively
thematic, in the sense that certain understandings may be repeatedly
applied in a wide variety of contexts. Finally, they can be more or less
widely shared; in fact, we do not call an understanding "cultural" unless it
is shared, to some extent, in a social group. Various writers and traditions
in cultural anthropology have been concerned with one or another of these
properties or a subset of them; our account addresses all of them within a
unified framework.
Furthermore, these centripetal tendencies are reconciled within this
framework with the equally evident centrifugal tendencies of culture,
which have earned so much recent anthropological attention. That is,
cultural understandings can be changeable in persons and across gener-
ations; they can be unmotivating; they can be contextually limited; and
they can be shared by relatively few in a society. We should be clear that we
are doing something more than acknowledging both tendencies. We want
to begin accounting for a complex world in which some kinds of cultural
beliefs, institutions and practices are enduring and others fleeting; a world
in which some beliefs, institutions and practices are imbued with great
psychological force and other kinds hold no particular emotional or
motivational force for individuals; in which some recur over many con-
texts while others are quite context specific; and in which some are widely
shared and others are shared by only a few people or even idiosyncratic.
In other words, we are initiating an explanation of why the centripetal
effects we have named occur when they do, and why the opposite, centrifu-
gal effects occur when they do. As was noted in the opening pages of this
book, if we think just about the centripetal effects of culture it is easy to
imagine that these can be read from the public world itself, and that we do
86 Implications for a theory of culture
not need an anthropology of intrapersonal meaning at all. Public culture
is, after all, the most tangible evidence of what endures in us, what
motivates us, what lasts over time, what themes repeat themselves across
the contexts of our lives, and what we share. But trying to account for why
cultural practices are only sometimes durable and other times transitory in
our lives and historically, only sometimes motivating for us and other
times not, and so forth, forces us to examine the intrapersonal processes
that explain these variable effects.
Throughout this part of the book, our central point will be that cultural
meanings are intrapersonal - that they arise in people. Connectionism
makes two key contributions to an explanation of how this happens. First,
a connectionist account preserves a theoretical balance between the flexi-
bility of cultural meaning and its structure - what we are calling the
centrifugal and the centripetal properties of culture. By this we mean, not
just that connectionism gives centrifugal and centripetal tendencies each
their due, but that it gives an account of how the two interact to produce
highly complex results. The chapters in part II will give illustrations of this
complexity and show how it can occur.
Secondly, connectionism helps to explain how the meanings that arise in
people do so in response to the outer world they encounter. The next two
chapters address this aspect of cultural meanings as well. They will de-
scribe how people get their schemas from public objects (including other
people) and public practices. They will also suggest - an issue to be taken
up further in part III - that this is not a straightforward, simple process, so
that we cannot form any expectations about what public objects and
practices will become private meanings, or how enduring, widely shared,
and so forth, these meanings will be, without studying the processes by
which they are learned. The chapters in part II will also address what
happens when schemas learned under earlier conditions then meet new
conditions in the world.
We want to make it very clear, however, that connectionism does not
alone do the job of explaining how cultural meanings arise in individuals; it
must be joined with other theoretical paradigms. Humans are multifaceted
organisms living in a complex environment - from which it follows that
many different processes contribute to any phenomenon of human life.
Unlike connectionism, most of the other theoretical paradigms on which
we will draw are not particularly new. Indeed, another of its virtues, we
think, is that our account revives and reuses psychological theory of a sort
that contributed to the anthropology of an earlier day, but that has been
eclipsed by the more recent emphasis on the publicness of meaning. In
stressing the importance of teaching, modeling, social evaluation, or
psychodynamics we claim no originality, then; indeed, sometimes it may
Implications for a theory of culture 87
seem that we belabor the obvious. We do so for a reason, however: we aim
to refocus anthropological attention on these various psychological pro-
cesses underlying the kinds of social practice that shape cultural meaning
for people.
In what follows we will discuss each of the centripetal tendencies we
have identified, noting also the centrifugal tendencies that parallel and
counteract each. If we grant less space to these centrifugal properties of
culture it is only because its centripetal properties are what require explain-
ing. A centrifugal world in which everything "falls apart," making no
lasting impression on individual lives or on future generations, failing to
propel people to any concrete action, and never spreading beyond narrow
contexts or few people, is what we could expect to happen without some
impetus to the contrary. It may also be that in the contemporary
sociopolitical mood a centrifugal world is more commonsensical - as
evidenced by the readiness of contemporary anthropologists and other
cultural theorists to imagine the world in this way. In any event, the
centrifugal properties of culture follow unproblematically in the same
terms we have developed for explaining the centripetal ones, we will show.
Nevertheless, these centrifugal tendencies have an equally central place in
the theory of culture we are proposing. We will take care to illustrate them
To illustrate centripetal and centrifugal tendencies and further points we
will be making in chapters 4 and 5, we introduce a fictional character,
Paula. Although entirely a product of our imaginations, she is a composite
constructed in large part from the lives of women friends, acquaintances,
and ourselves. Thus, while "Paula" is not the pseudonym for a real person,
she is a realistic US American of a certain type (for example, the type
depicted in the comic strip Sally Forth). Lest some readers conclude that
we are trying to make this fiction do the work of actual evidence, we stress
that we mean Paula's experiences to be purely illustrative. They are in-
tended solely to remind readers of the properties of cultural understand-
ings that we then attempt to explain. These properties themselves should
not be controversial or require further proof; rather, what is needed is to
account for them all. That explanation is the purpose of this part of the
book. In part III we will describe research on actual people's understand-
ings, conducted by each of us, that incorporates and exemplifies the
approach we are delineating.
Another purpose served by illustrating our points with the lives of a
fictional character and her family, friends, coworkers and acquaintances is
to emphasize that cultural understandings are not Platonic abstractions
but always belong to people who are shaped by specific life circumstances
in given places and times. Paula, for example, was born shortly after the
88 Implications for a theory of culture
end of World War II in the baby-boomer generation and grew up in a
white middle-class family living in a suburb of a medium-size US city. Her
father was an engineer and her mother stayed home and raised Paula and
her younger brother, Daniel. Paula went to college, married rather later
than her mother did, and unlike her mother has held a paid job continu-
ously since she graduated from college. Now, in her early forties, with two
young children, she holds a responsible managerial position in a large
telecommunications company. Our focus on Paula's particularities does
not mean our theory is limited to US Americans like her; we challenge the
reader to think of any cultural setting in which the properties we describe
do not hold. However, if we are to break through the current impasse in
culture theory, we need to stop talking about "culture'* abstractly as a
property of reified social groups, and look instead at how cultural under-
standings are shared and vary among particular people - in our society as
well as others.
Two properties of culture
In this chapter we consider the two properties of culture that draw us most
fully into a consideration of the individual who bears and acquires culture,
properties that hence require the most complex psychological arguments
about the way in which culture is internalized.
Durability in the individual
Some beliefs, values and other cultural understandings that people have
stay with them a long time, sometimes their whole lives. Culture theories
that focus on public forms of culture at the expense of the understandings
people acquire from those public forms naturally have a hard time explain-
ing the durability that schemas can have in individuals. If the world of
messages surrounding us is rapidly changing and we are constructed by
these discourses, why are our understandings not rapidly changing as well?
On the other hand, how can a model that explains durability likewise
account for people's obvious ability to adapt to change? Answering these
questions will require the longest discussion in part II.
This discussion entails a consideration of how the world is organized to
ensure that the same associations will be made repeatedly. Explaining
durability also leads to recognition of the role of emotional arousal in
making some schemas durable, including some learned very early in life.
And it invites us to discuss in some detail how teaching achieves its end of
making learning durable. We go on to address the roles of other people's
expectations and of institutional constraints on the durability of some
cultural practices and the understandings attached to these. Turning next
to centrifugal forces against durability, we explore how new situations
sometimes present themselves in ways that ensure attention to new infor-
mation, or demand new interpretations. Finally, we consider circumstan-
ces under which tendencies toward durability are overcome by deliberate
effort to change cultural practices and understandings.
90 Implications for a theory of culture
Centripetal tendencies toward durability in the individual
Part of the reason why cultural understandings are durable in the individ-
ual is that they rest on neuronal connections that are not easily undone.
Until now we have talked in terms of connectionist computer models; but
these models are plausible, we have suggested, because they are inspired by
certain properties of the brain. There is a neural basis for associative
learning. When neurons are consistently activated by co-occurring features
of experience, physical changes in the neurons strengthen the connections
between and among them (Kandel, Schwartz and Jessell 1995:479,667ff.;
Merzenich and Sameshima 1993:190).* Thereafter, if one of those neurons
is activated, it will be more likely to activate another in that group.
Remember that the environment determines which features of experience
co-occur, and that a large part of that environment is culturally construc-
ted. Growing up in an environment of a given cultured shape brings with it
a distinctive pattern of experiences and corresponding neural changes.
These neural changes determine the pathways through which activation
spreads until a particular response is evoked. The synaptic changes that
make this happen cannot be erased like sentences from a text (or even
altered as quickly as the weights in a connectionist computer model).
Change in the world can lead to new patterns of strong neural connections,
but it does not completely destroy earlier learning.
For example, in
getting to know Paula, the fictional character whose life was summarized
in the introduction to part II, one of the first things you would realize is
that she is a good feminist. She and her husband, Michael, are serious
about sharing housework equally. Even so, Michael's ingrained disposi-
tion not to notice when something needs to be done around the house and
Paula's unbreakable habit of being bothered by undone housework some-
times causes them to drift back into an old gendered division of household
labor that they then realize seems unequal.
Another part of the explanation for the stability of schemas is that they
tend to be self-reinforcing. For the most part the schemas of infants are
highly malleable. In the course of development, however, some schemas
become increasingly well established, with the result that subsequent ex-
periences are much more likely to be understood in their terms than the
schemas are to be altered by these experiences.
This process can be
understood connectionistically as follows: once a network of strongly
interconnected units has been created, it fills in ambiguous and missing
information by activating all the units in an interconnected network, even
those not directly stimulated by current experience. Subjectively, we may
experience all the features of the typical event when only some of its
features are present, reinforcing our original expectations (an effect psy-
chologists, e.g., Loftus 1979, have demonstrated experimentally).
Two properties of culture 91
This tendency of schemas to fill in for us can block disconfirming
evidence. For example, thinking well of ourselves as most of us do, we may
go through life blithely unaware of bad habits and other socially objection-
able behaviors of ours that annoy, frustrate, or anger other people, be-
cause these negative traits do not accord with positive views that we hold
of ourselves and get filtered out. Equally, our dislike of someone can lead
us to read everything they do as bad. Paula has a colleague at work with
whom she is on bad terms. In the past he has demeaned her in subtle and
not-so-subtle ways, and Paula has built up the expectation that he will
continue to do so. As a result, many of his comments, even those innocent-
ly intended, trigger her "there-he-goes-demeaning-me-again" schema,
with the result that she reads malevolent intentions into them. As far as she
is concerned, her experience contains recurrent evidence that this colleague
looks down on her, and each new interaction only strengthens the negative
associations she has already constructed.
Furthermore, negative social schemas can lead people to avoid interac-
tions that might provide disconfirming evidence that would change their
schemas. Brent Staples (1994:201-204), in his autobiography Parallel
Time: Growing Up in Black and White, gives a striking description of such
patterns of interaction. As a black graduate student at the University of
Chicago in the 1970s, he discovered that he could not walk the streets at
night without having white people run from him, stare straight ahead to
avoid his eyes, or freeze and still their conversations until he passed;
women would clutch their purses, couples would lock arms or reach for
each others' hands for security, and people in cars would hammer down
their door locks. At first he tried to make himself more innocuous, turning
out of his way to avoid people, whistling popular tunes as he passed them,
and waiting for them to exit building lobbies before he entered so they
would not feel trapped. Then, angered, he began deliberately to bear down
on people or aim himself between them, laughing at their fright as he
passed them.
Paula is no different from the whites Staples encountered on Chicago
streets. Given her middle-class suburban childhood and the fact that she is
white, she has had little first-hand experience with poor inner-city blacks.
Her stereotype is that poor African-American men are likely to be violent
criminals. Recently, alone in an unfamiliar large city on a business trip, she
was approached by a shabbily dressed black man. Something about the
way he looked triggered Paula's fear-laden associations, and she turned
and ran. It didn't occur to her that she looked lost and he was approaching
to offer directions. For Paula, the experience reinforced the connection in
her mind between poor African-American men and violence, and she has
since told others about the incident as evidence that it is not safe to walk
the city streets anymore.
92 Implications for a theory of culture
Stereotypes are good examples, as well, of other ways in which schemas
can be self-reinforcing. Consider a possible problem with using an example
like the last one. A reader who held the same stereotype as Paula, and who
read this example quickly, might well miss our cautionary point about the
way in which false stereotypes can persist and even strengthen in the
absence of empirical support. The effect then would be to reinforce this
stereotype, exactly the opposite of what we intend. A different and even
more insidious way ingrained stereotypes can be self-reinforcing is that
they become part of the experience of the stereotyped themselves. This is
what Claude Steele and Joshua Aronson (1995) call "stereotype vulner-
ability." These psychologists show, for example, that blacks underperform
in relation to whites on a difficult verbal test if they are led to believe the
test is diagnostic of ability, although performing the same as whites
(adjusted for SATs) on the same test if led to believe that it does not
diagnose ability. In the first condition the black participants are presum-
ably under pressure of the stereotype of racially inferior intellectual ability,
a stereotype made plausible by the difficulty of the test. This threat, in turn,
can interfere with performance in a variety of ways, such as causing an
arousal that reduces the range of cues participants are able to use, or
diverting attention to task-irrelevant worries. Steele and Aronson
(1995:798) point out that the stereotype need not even be internalized as a
belief by the stereotyped person for that person to be vulnerable to these
effects. The person may only need to know that his or her performance is in
danger of confirming the stereotype, and being judged by it. Unfortunate-
ly, of course, it often is - thereby reinforcing the stereotype in the minds of
whites in a position to judge and interpret such underperformance.
Reinforcement of stereotypes in this way is a special instance of a more
general case: by behaving in a way that confirms others' schemas about
them, people often prompt responses from these others that reflect the
others' interpretations and that re-elicit, and thus reinforce, the same
behavior that confirmed the others' schemas in the first place. Sensing as he
does that Paula does not like him, her colleague at work tries to avoid her,
behavior that she interprets as a lack of interest in her colleagueship, and
that feeds her more general interpretation that he does not respect her.
Psychologist Drew Westen (personal communication) provides the further
example, from attachment research, of avoidant infants shutting off their
needs for attachment, causing others to respond to them in ways that
reinforce the infants' attachment avoidance. When a group of people has a
stereotypical view of another group and that view is confirmed in this way,
the resulting circle of reinforcement can be equally difficult to interrupt,
and more consequential for society as a whole.
A further reason why some cultural schemas are durable is that emo-
Two properties of culture 93
tional arousal during or immediately after an experience (providing it is
not excessive) strengthens the neural connections that result from that
experience (Freeman 1991:81, 82; McGaugh 1989; Squire 1987:39-55).
Thus, Paula's anger in interactions with her coworker, and her fear in the
inner-city incident, altered the neurochemical environment in which rela-
tions among features of those experiences were internalized to render her
schemas stronger than they would have been had she not cared so much
about what she was observing. Each time she perceives her coworker to be
demeaning her, for example, her anger kicks in to heighten the effect of the
incident and the strength of the associations she brings away from it. The
resulting schema becomes still stronger and even more likely to be ac-
tivated in the future, at the expense of alternative understandings.
Notably, the role of emotional arousal in making schemas durable can
explain the indelibility of schemas learned in infancy, if we assume that
early experiences tend to be linked to exceptionally strong feelings related
to survival and security. To the degree that such early experience is widely
shared, the durable schemas shaped by this infantile experience will be
cultural schemas. We will have occasion to call upon this assumption
about the indelibility of early schemas later in this chapter when we discuss
the durability of feelings about being a good or a bad person.
Teaching is almost always designed and conducted to the end of making
learning durable. As obvious as it is, this truism about how teaching can
work to establish enduring cultural schemas is altogether missing from
Bourdieu's (1977) account and others (e.g., Ortner 1990) that draw on his.
Common teaching techniques make perfect sense given connectionist the-
ory (and associationist theories of learning more generally). At the heart of
most teaching is the engineering of the learner's repeated exposure to and
practice of what is being taught, so that networks of connections are
gradually established and strengthened. Most of us, like Paula, can still
recite multiplication tables that we learned this way in school decades ago.
Since repeated exposure can have no strengthening effect on connections
unless what is presented is actually perceived,
effective teaching must also
use devices designed to focus learners' attention on the relevant informa-
tion. Also important are devices to create the emotional arousal that, as we
described above, helps make learning "stick." Other teaching techniques
make learning durable in ways that further complicate and extend connec-
tionist frameworks as currently conceived. To summarize these compli-
cated processes, parents and other socializers achieve durable results
through meting out rewards and punishments for specific behaviors, con-
veying more general evaluations of the learner as a good or bad person,
and giving or withholding love.
Some rewards, such as (in American society) giving children ice cream
94 Implications for a theory of culture
for being well behaved or money in exchange for good grades, are effective
only as long as they keep being provided by socializers. Other rewards and
punishments, such as praising and frightening, can make learning durable
because the feeling of being praised or frightened not only motivates
learners when it is originally experienced, but is reproduced in them to
motivate these learners when occasions for similar behavior arise in the
future. Thus the praise one has learned to expect for doing as one has been
taught may be self-administered, and the memories of fright, aroused by
contemplation of doing something one is not supposed to do, can them-
selves be frightening enough to inhibit the forbidden behavior. Let us
consider a somewhat fuller example from Paula's childhood, one to which
we will have occasion to return later in this chapter and in the next. Paula
no longer needs to be reminded to be self-reliant, as she was prompted to
be by her parents when she was little; her pride in her own competence, or
(at other times) her embarrassment and disappointment in herself at
having to ask for help, are feelings automatically engendered in her by
situations that call upon her to do something for herself. Moreover, just as
with any skill in which she becomes expert, the more competent she
becomes at a task that she associates with being self-reliant, the more she
finds the task intrinsically rewarding. Her pleasurable feeling of compet-
ence becomes stronger and more consistent, and she may also come to
enjoy a feeling of superiority to others less expert than she. These intrinsic
rewards make the skills she has learned in order to be self-reliant, and her
self-reliance more generally, even more durable. Notice that this descrip-
tion of learning already takes us well beyond any simple behaviorist
mechanisms that might be implied by our references to rewards and
Human learning depends on more than appropriate rewards and pun-
ishments. One class of cultural understandings can often play a general
role in making learning durable. These are social evaluations. Ideas about
goodness and badness inform social life everywhere, whatever the distinc-
tive cultural shape of these understandings. Sometimes they are cast in
terms of what is moral - e.g., being a good or bad Christian or a good or
bad child; and sometimes in terms of what is natural - e.g., being psycho-
logically normal or abnormal, or being a real man or not, or a feminine
woman or not. These ideas are typically learned not as abstract intellectual
possibilities, but as that which is approved and disapproved about other
people, their behavior, and the groups they belong to, and/or about
oneself, one's own behavior, and one's own group. When such approval
and disapproval engender emotions such as admiration or love for others,
hatred or fear of others, pride or satisfaction in oneself, shame or guilt over
one's own behavior, then judgments of social approval or disapproval are
Two properties of culture 95
learned in association with these emotions. Once learned in association
with given ideas of good and bad behavior, the emotions themselves - even
in the absence of real or fantasied social approval or disapproval or other
rewards or punishments attendant on these - motivate the good behavior
and inhibit the bad, just as other positive and negative emotions can
motivate behavior with which they have become associated. Thus, one
feels guilty for not having been Christian or ashamed of being a wimp.
Social evaluations need not always have an emotional side to them, of
course. But when they do - when learned through approval and disap-
proval and associated with the strong emotions that accompany those -
social evaluations of oneself and others can be very durable indeed.
Ultimately, evaluative discourses directed at children by parents (and
other primary socializers) draw force in all societies from the child's desire
to insure the parent's love (Spiro 1987b). The early experiences of being
loved for being good and having love withdrawn for being bad are durable
lessons because, linked in the preverbal child's understanding with security
and survival itself, these experiences are so highly emotionally arousing.
While Melford Spiro (1987b:136-138) emphasizes the "moral anxiety"
that love withdrawal arouses, we would expect that early experiences of
being loved for being good would also be memorable. The early, global
concern with being good and not being bad prefigures more culturally
defined ideas about being this or that kind of good or bad person. It colors
other important early experience, such as that of separation and individua-
tion, to be discussed in the next section of this chapter in the context of
learning self-reliance. The same emotions that, we will see, make such early
learning highly motivating also make its lessons indelible. Still another
aspect of early experience that similarly produces enduring as well as
powerful effects is the infant-caretaker relationship as this shapes the adult
experience of love. This will be considered in the context of Naomi Quinn's
research described in chapter 7.
In and out of schools, teachers capitalize on the durability of social
evaluations of the self. Knowing that learners can almost always be
counted on, over time and context, to want to feel like a good person rather
than a bad one, teachers attach the behaviors they are teaching to evalu-
ative discourses so that learners' ongoing desires to be good will activate
the associated behavior in appropriate future contexts. "Don't be bad,
now," we say to children habitually, or, "What a good girl!" By the time
Paula had to learn the multiplication tables she had already learned to
pride herself on being a top student, and she wanted to learn her tables so
that she could demonstrate how good she was at reciting them. Desires
such as those to feel like a good student or a good mother or a self-reliant
person thus lend their considerable durability as well as (we will see in the
96 Implications for a theory of culture
next section) their force to the behaviors that have been linked to these
ideals. Furthermore, to the extent that evaluations become aspects of self
identity, they can act as very stable goals for which people strive through-
out their lives in order to remain true to their self images. These goals are
stable and powerful enough to prevent people from drifting into other
patterns of behavior that would be easier accommodations to practical
obstacles. Emphatically, such consistently rewarded, well-learned, highly
stable self identities are not mere creations of shifting social discourses that
individuals strategically adopt and perform, which is the constructionist
view of them that we critiqued in chapter 2. This point holds for the
gendered, ethnic, and national identities with which constructionists are
concerned as well. We have already given an example of how durable
understandings of oneself as a good mother, a feminine woman, or a real
man can be. National and ethnic identities can be as enduring, we know,
surviving emigration and the incorporation of immigrants into new nation
states, or erupting again in ethnic wars after long periods of quiescence or
state suppression.
As Bourdieu (1991:52) puts the same point,
[t]he power of suggestion which is exerted through things and persons and which,
instead of telling the child what he must do, tells him what he is, and thus leads him
to become durably what he has to be, is the condition for the effectiveness of all
kinds of symbolic power that will subsequently be able to operate on a habitus
predisposed to respond to them.
Bourdieu also reminds us that durable identities, once learned, exert an
ongoing effect on behavior that need not always be self-conscious, but can
(like any other learned behavior) be quite habitus-like. Paula, for example,
sees herself as a good mother, and works at being one continually and
consciously, making sure she attends PTA meetings regularly, spending
sufficient quality time with her children, producing handmade Halloween
costumes for them every year, and generally making accommodations in
her busy schedule to do other things that she identifies with her ideal of
motherhood. At the same time, she quite unconsciously emulates a style of
interacting with children that she has learned to associate with being a
good mother, almost always showing exceptional patience and adopting a
cheery tone of voice with them, and becoming very nurturant when they
are hurt or upset and very protective when they seem to be in trouble.
Resistance to change rests on more than the properties of one's own
schemas and the way they were learned. Sometimes others' well-learned
dispositions require so much effort to overcome that change is very diffi-
cult. For example, we have observed that Paula and her husband Michael
are committed to a more egalitarian division of household labor than they
Two properties of culture 97
observed as they were growing up. They have been particularly successful
in sharing responsibility for cooking and dishwashing. Even so, when they
are busy entertaining, they find that their guests* expectations force them
back into gendered patterns they and the guests observed as they were
growing up, Paula and the other women clearing the dishes at the end of
the meal while Michael and the other men sit engrossed in the dinner-table
conversation. Ordinarily Michael and Paula clear the table and wash the
dishes together, but in front of company, when female guests are volun-
teering to help and male guests are not, the two feel awkward about
insisting on their usual role violation. It is not simply changes in their
behavior that are impeded by a resistant social world: her clearing and
washing the dishes with the women and his sitting idly at the table with the
men reinforces, in both of them, a whole array of unarticulated assump-
tions and feelings about men's and women's places. This is perhaps more
so in the case of Michael, in whom the experience evokes no anger,
frustration, or other conflicted feelings of the sort evoked in Paula, that
might contradict his sense of ease with gender roles as they had always
It is not only at the interpersonal level of others' habitus, but also at the
level of institutional practices and policies, that the world can be un-
cooperative about changes such as those Paula and Michael have tried to
make in the gendered division of labor. When they had a baby, for
example, they discovered that Paula could take maternity leave from her
job but Michael did not have the option of taking paternity leave, so they
found themselves reproducing their parents' child-rearing roles during the
baby's first months. Paula's staying at home with the baby and Michael's
continuing to work thus reinforced the associated ideas, motivations and
feelings each of them had about mothers staying home and men pursuing
their careers. Institutional arrangements such as this workplace policy can
be especially diflicult to change, resting as they do not only on the beliefs
and attitudes of those in charge of making them, but also on an intricate
web of other institutional arrangements that would also have to be altered.
When he inquired about it Michael found, for example, that administra-
tors in the human resources office of his company considered paternity
leave an unrealistic benefit option because, unlike maternity leave, it could
not be folded into the existing short-term medical leave benefit offered in
the company's medical insurance plan. This plan's subsumption of child-
birth and hence maternity leave under short-term medical leave was stan-
dard in the medical insurance industry, Michael learned. Were paternity
leave to be added to comprise a parenting-leave policy, leave could no
longer be justified in terms of childbirth defined as a medical condition and
funded in this way. The number of maternity-leave requests being substan-
98 Implications for a theory of culture
tial, the cost to Michael's company would be considerable. Of course, to
the degree that local practices and policies like this one are governed by
state-wide or even national organizations or laws, individual companies
like Michael's may not even be free to devise their own solutions to the
needs of their employees.
Centrifugal tendencies against durability in the individual
Now we turn to centrifugal tendencies. In our account the tendency of
schemas to durability does not require that they will always be so. First, it
is important to stress that schemas do not act as perceptual filters, keeping
incongruent information out. It is typical of anthropologists to hold an
extreme epistemological relativism according to which culture (or our
"subject position") determines the very look, sound, feel, etc. of the world
for us. This brand of relativism is infused with thinking that accepts
unexamined a "strong" Whorfian hypothesis, according to which the
labels for things, the syntax of a language, and other of its features screen
the perceptions and hence completely determine the conceptual distinc-
tions of its native speakers (see, e.g., Kondo 1990:26-33; Sahlins 1985). If
this were true, it would be difficult for the knowledge structures of adults to
change, because their well-learned schemas would prevent them from
becoming aware of anything that contradicted these understandings.
While it is true that our schemas may cause us to notice some things and
overlook others and that schemas reproduce learned expectations when
faced with missing or ambiguous information, there is no evidence that
schemas, including cultural schemas, act as gatekeepers that bar percep-
tion of unexpected events (Alba and Hasher 1983). For example, if Paula
had stayed long enough to listen to the man who approached her in the
city, her stereotype would not have prevented her from hearing his offer of
help. As Roy D'Andrade (1991) has pointed out, connectionist models can
explain why schemas do not screen out new knowledge and hence why
culture is not a perceptual "veil." During any new experience, missing
information is filled in by the weights of association learned earlier, but
present information, if attended to, will ordinarily override the activation
of those old associations.
The result can be that new schemas are learned
and older ones changed. Indeed, there is evidence that under some circum-
stances schema-inconsistent behavior will be especially noticeable and
memorable, and especially likely to change a stereotype. One such circum-
stance is when schema-inconsistent behavior is exhibited by an otherwise
good exemplar; another is when the behavior exhibited clearly contradicts
the schema (Schneider 1991:534-537). It is possible, in addition, that novel
experiences may - just because they upset one's commonsensical view of
Two properties of culture 99
the world - arouse strong emotions, leading, for the reasons we explained
earlier, to a large change in cognitive structures faster than normally
occurs. Had she stayed to hear the oifer of help, then Paula's stereotype of
inner-city African American men might have been forever altered. Of
course, this does not always happen when people's stereotypes are contra-
dicted, but it does happen. For reasons of cognitive ability to "break set"
or of emotional responsiveness to inconsistency, some people may be more
open than are others to altering their schemas in the face of schema-
contradictory information.
In addition, in connectionist models even well-learned schemas do not
dictate an unchanging response, because the response that is suggested
depends on all the features of any given situation and situations are rarely
the same. As Paula's world changes, not only do some of her schemas
change, but even her more resistant, durable associations adapt by com-
bining in new ways with other associations to yield innovative responses.
For example, as Paula was growing up, she observed countless examples of
conversations. In each, the identity, roles, and aims of the participants
were associated with different tones of voice (varying in loudness, tempo,
and voice quality), gestures, postures, and facial expressions. According to
our model, Paula did not learn a set of rules (for example, if you are a
woman and if you want something, speak in a soft, hesitant, breathy voice)
but rather a large number of connections of different strengths among
activated by features of the contexts for talk (identity of the actors,
purpose of their actions, content of their talk) and those activated by the
paralinguistic and kinaesthetic cues accompanying talk. These multiple
connections allow her to respond flexibly to situations she may never have
been in, been taught to handle, or observed.
This is exactly what happened to Paula when she changed jobs and
discovered that her new boss was a woman. The context of speaking to a
woman activated one set of units for Paula, while the context of speaking
to someone with power over her activated another, only partly overlap-
ping set. As these units simultaneously activated still other units, their
combined influence eventually settled into one response more than others,
suggesting a way of reacting that felt right to Paula: a vocal, postural and
gestural style and a corresponding sense of herself in this new situation that
represented a compromise between the respectful manner she usually
reserved for men in authority and the easy give-and-take that was her
typical demeanor toward other women.
This experience illustrates the
"regulated improvisations" (Bourdieu 1977:11) that connectionist models
explain well and shows how this property they have allows connectionist
networks to adapt to changes in society. (Of course, with repeated experi-
ences of dealing with her boss and other women in positions of authority,
100 Implications for a theory of culture
the new connections activated by Paula's improvisation would be
strengthened as well as supplemented by connections to other features of
such interactions, and Paula would eventually have a strong, stable schema
for this new social phenomenon.) Paula's first experience with a woman
boss also illustrates especially well the way in which schemas can be
learned and activated independently of language. When she found herself
adjusting to the unfamiliar interaction with her new supervisor, she had no
ready way of describing how she was feeling and behaving, so it did not
occur to her to talk to anybody about it.
Paula's possibilities for innovation are not limited to passively reacting
to changed situations, such as getting a woman boss. By making a deliber-
ate effort, Paula can accustom herself to thinking and acting in ways
different from those that come automatically to mind as natural, given the
pattern of her past experiences. This conscious effort is an example of what
some cognitive psychologists call a "controlled" process, in contradistinc-
tion to involuntary "automatic" ones (Schneider and Shiffrin 1977; Shif-
frin and Schneider 1977). Well-learned, strong associations suggest auto-
matic responses, but we have seen that these responses can be overridden
through conscious control. As we pointed out in chapter 3, controlled
processes are not readily explained within a connectionist framework. Of
course it is also possible that once-novel responses, initially under con-
scious control, can themselves become habitual with practice, and hence-
forth activated like any other network of connections. For example, many
of us at first found substituting "he or she" for "he" foreign to our tongues
and remembering to do so a struggle; now most people think "he or she"
effortlessly and use it automatically.
We have already seen an example, in all the things that Paula undertakes
in order to be what she considers a good mother, of deliberate effort to
maintain a traditional social role. Conscious effort can be put to centrifu-
gal as well as centripetal effect, depending on the circumstances, as a
different example will show. While parts of Paula and Michael's feminist
agenda may have been undermined by structural constraints in the world
and social pressure from other people, they have been able to effect some of
that agenda by retraining. In the beginning of their marriage, Paula and
Michael fell into a pattern: Even though they took turns shopping, cook-
ing, and cleaning up after meals, Paula was the one who planned the
weekly menus, noticed what groceries they were running out of, and put
them on the grocery list, and in other ways took primary responsibility for
their meals. When Paula took her new higher-level and more responsible
job, however, she told Michael that between home and work she simply
had too much to worry about and he agreed. Michael volunteered to take
over managing the kitchen. At first he had to be reminded to sit down on
Two properties of culture 101
Sunday mornings and make up the menus for the week; and more than
once Paula found herself in the middle of cooking dinner without a key
ingredient for her recipe, which Michael then had to run out to pick up.
But Michael soon mastered the job of kitchen manager and was exceeding-
ly proud of himself for doing it well.
After awhile it became unremark-
able and nowadays, making up weekly menus and a shopping list,
monitoring the supply of staples, and checking Paula's recipes for needed
ingredients are as routine for him as doing the Sunday crossword puzzle.
Now, only when a surprised guest comments on his role in running the
house is Michael reminded that it is a bit unusual. The change is not just
behavioral. Michael and Paula do not even realize it, but, for them, the
meaning of "masculine" and "feminine" has shifted slightly, in line with
their reassignment of kitchen responsibilities.
In any case, deliberate cognitive effort of the sort that goes into a more
equalized division of household labor is more likely to be fueled by
well-learned schemas than by shallow, fleeting beliefs. It is not sufficient to
explain Paula's attempts to share kitchen work and child rearing more
equally as a product of the feminist discourse and practices of the present
time. That discourse and those practices are not imprinted on a tabula rasa;
they appeal to Paula as a result of the durable schemas she brings from her
past, like her memories of her mother working in the kitchen. It is to these
schemas that we next turn, to examine the affectively and motivationally
charged associations that her early experiences of her mother working in
the kitchen have for Paula.
Motivational force
Some of the same processes that make people's understandings and feel-
ings more or less durable, as well as some other processes not yet consider-
ed, give these schemas more or less motivational force. Were cultural
schemas to stay with the individuals who share them, but never have any
discernible affect on how these people behaved, that would be inconsistent
with what we know about culture. On the other hand, the observations
that individuals enact the schemas they share selectively, some motivating
a given individual more than others, and that different individuals can
share the same schemas but not necessarily the same motivation to enact
them, explain much of the behavioral variability that is evident among
In this half of the chapter we address, first, the general process by which
emotion and, through it, motivation, are incorporated into schemas as
part of the experiences from which those schemas are formed. We go on to
consider how teaching exploits some of the same techniques discussed in
102 Implications for a theory of culture
connection with the durability of learning, to impart motivation to what is
being learned. Then we turn to the special role of social evaluations in
motivating cultural understandings - whether these evaluations are ex-
plicitly taught, or are learned from the approval, disapproval or modeling
of those who matter. These generalizations about how cultural schemas
are motivated are next illustrated by the case of American self-reliance.
This case also enables us to discuss how motivation is deliberately im-
parted to learning, not only by explicit teaching, but also by the design of
the environment; and by the establishment of a foundation, in early
learning, for further training. Pursuing the example of self-reliance, we
next consider the psychodynamic basis, in infantile experience, for the
powerful motivation attached to some cultural schemas. The final section
of chapter 4 asks why cultural understandings do not always motivate
people to act. One possibility we explore is that the goals of different
schemas conflict. Another is that some schemas are learned with little or no
motivation attached to them.
Centripetal tendencies toward motivational force
As Paula was growing up in the 1950s, she observed that her own mother,
her friends* mothers, mothers she read about in stories, and mothers she
saw on television were more likely to be in the kitchen than were the men in
those families. It is crucial to our account to emphasize that Paula's
observations of her mother in the kitchen were not cool or detached.
Certain feelings were part of that experience. As she grew older she became
more aware that her mother did not like being so confined to the kitchen.
Paula's unsympathetic reaction was to feel impatient with her mother for
not insisting on a different division of labor in the household. As time went
on, the mere sight of her mother in the kitchen was sometimes enough to
annoy Paula, and also reminded her how important it would be for her to
insist on a different domestic arrangement when she was married and in
her own home.
We have been assuming that feelings and motivations are incorporated
into schemas, just as other experiences are,
and in the last section this
assumption contributed to our explanation of how schemas are durably
learned. It is central, as well, to our explanation of their motivational
force. Clearly, Paula learned more than associations between observable
features of her childhood experiences; she also learned associations be-
tween what she observed and certain feelings she had. Those feelings, in
turn, engendered certain motivations in her - in this case, the determina-
tion not to be "stuck," as she thinks of it, in the kitchen herself.
Connectionist modelers, most of whom have been concerned exclusively
Two properties of culture 103
with the problem of getting computers to think intelligently, have not been
much interested in how to explain emotion and motivation. On the other
hand, connectionist models can readily be extended to include them.
make the extension we need to assume that inner subjective states of feeling
and wanting, like observable behaviors, are activated through associative
networks (Fiske and Pavelchak 1986). Granting this assumption, Paula
could develop links between ideas about her mother in the kitchen, feelings
of annoyance, and the aim of doing something different. Her memories of
Thanksgiving at her grandparents, on the other hand, are suffused with the
warmth of a happy occasion, free of the tension in her parents* house. So
Paula associates Thanksgiving, not only with turkey, cranberries, and
stuffing, but also with warm happy feelings, and now, with the intention of
recreating that experience for her children. Her ideas about Thanksgiving,
like those about her mother, have motivational force (D'Andrade and
Strauss 1992).
Implicit in this last example - as in some of the examples of durable
learning in the previous section - is a further assumption about the relation
between emotion and motivation: While feelings may be elicited by, and
hence associated with, a wide range of experience, motivations are me-
diated by the inner experience of feeling. How this works precisely, and
how it can be understood in connectionist terms, has been detailed by
Drew Westen:
[PJeople are motivated to seek pleasurable states and avoid painful ones. When
they experience painful events, they attempt to alter the situation to alleviate the
feeling. If this cannot be done behaviorally, they may use conscious coping mech-
anisms, such as praying or telling themselves things will be better soon, or uncon-
scious defensive processes, such as denial or rationalization. To the extent that
these procedures work, they will be associated with regulation of the aversive
feeling and hence reinforced, that is, made more likely to be used again in analog-
ous situations. And to the extent that they are repeatedly used under certain
conditions, the weights connecting the units representing these procedures and
those representing the circumstances in which they are elicited will increase.
(Westen, in press)
Conversely, of course, people also attempt to regulate pleasurable feelings
in order to bring about or maintain them, and the behaviors that are
successful at doing so are reinforced.
As we have already seen in the last section, social evaluation is one kind
of experience that is highly likely to be emotionally laden and, hence, to be
motivating. Just as they can make learning very durable, social evaluations
can often play a general role in adding motivational force to a wide
spectrum of experiences with which they become associated. When social
evaluations are part of an experience, any emotions associated with these
104 Implications for a theory of culture
ideas and the motivation to act on these emotions attach more force to that
experience than it would otherwise have.
Thus, Paula does not recreate
Thanksgiving for her children only because she loves them and wants them
to have the same kind of happy time she had at her grandparents* on
Thanksgiving, although that is certainly the case. She also believes that
children need happy childhoods, that an important ingredient of a happy
childhood is happy memories of family holidays, that mothers are respon-
sible for making these memories, and that good mothers fulfill this respon-
sibility. And she not only thinks of herself as a good mother, as we saw in
the previous section of this chapter, but has a considerable emotional
investment in being such a mother.
Paula's desire to be a good mother and other culturally shaped desires
were learned in a variety of ways. Some such motivations are deliberately
inculcated by socializers. The same techniques used to make learning
durable work for making learners want to enact what they have learned,
and much teaching is dedicated to this dual end. Thus, parents, teachers
and other socializers engineer repeated occasions for the learner to ob-
serve, attend to, and practice the desired behavior. We have already
pointed out, in the last section, how repetition strengthens the networks
of connections being learned, increasing the durability of learning. Rep-
etition also intensifies the learner's motivation to enact what has been
learned, in two different ways. First, behaviors repeatedly observed and
practiced become habitual; in the absence of any strong motivation to
behave otherwise, they will always be the behaviors of choice because
they are the only familiar and perhaps the only imagined possibility.
Their very habituation may make them seem so natural that ways of
acting differently become not just difficult to imagine but unlikely even to
be noticed, let alone learned. Secondly, the more occasions learners have
to attend to and enact behaviors, the more often they can be rewarded for
attending to them well or doing them right, and punished for not paying
attention to them or doing them wrong. For, another universal way in
which socializers insure that their charges not only learn, but are moti-
vated to do what they have learned to do, is by connecting the desired
behaviors directly to rewards and punishments. About these processes
behaviorists were certainly right. Learners are then motivated to experi-
ence the pleasure of a reward, such as an approving smile, praise, or food,
or avoid the pain of a punishment, such as being shamed, frightened,
hurt, isolated, or deprived.
Socializers frequently fortify their teachings in still a third way, motivat-
ing learning with the same evaluative discourses they employ to make this
learning durable, by attaching the behaviors they want enacted to ideas
about learners' own goodness or badness. As we discussed in the last
Two properties of culture 105
section on durability, these self-evaluations have already been learned, and
charged with emotion and motivation, in other contexts. That is, wanting
to be good and not bad not only makes learners learn well what it takes to
be good and not bad, but also makes them want very much to do what they
have learned it takes.
Of course, learning social evaluations of these kinds, like other learning,
can and does occur apart from the context of explicit teaching. One of the
most important ways in which we come to want to perform certain
behaviors, even apart from being explicitly taught to behave that way, is
from the social evaluations of particular others who matter to us, whether
our elders or our peers. Their approval and disapproval, with all the strong
emotions engendered by these, make us want to be what those important
to us care about being and want us to be, which accounts for much
conformity to the dictates of our family, our peer group and the other
groups to which we belong. Paula can not remember ever being explicitly
taught that she would grow up and be a mother; it seems to her that she
always wanted to be one. Her desire to be a mother grew, in part, out of a
sense of what her own mother, her aunts and other older female relatives,
and later in adolescence and early adulthood, her high school and college
friends expected of her. If she could have analyzed the divergent sources of
this motivation, she might have said that wanting to be a mother was
partly something innate in her, and partly tied to feelings of finally becom-
ing an adult woman in the eyes of her relatives, and of being a whole person
in the eyes of her woman friends. For Paula, as for most of us, both full
adulthood and whole personhood are charged with positive evaluation.
Another way in which we learn without explicit teaching is by modeling
(although, of course, teachers can and do deliberately model behavior as
one technique for teaching it). Modeling involves the tendency to imitate
other individuals, such as parents, who are available, the focus of atten-
tion, and perceived to be powerful or prestigious, skillful, and nurturant;
and the further tendency to pattern our behavior after the prototypical
behavior of important, valued categories of others (Maccoby and Martin
1973). Of course, we do not automatically model our behavior after
another's behavior just because the person is a parent or is otherwise
important in our lives, as we have already seen in the case of Paula's
determination not to be like her mother in certain ways. Her ideas about
the good mother she wants to be are drawn from many representations and
actual instances of mothering she has encountered, and include some
things she admires about the way her own mother brought up her brother
and her. Modeling is closely related to social evaluation, as Eleanor
Maccoby and John Martin (ibid.:9) point out: While very young children
often simply imitate what they see, older children begin to select persons
106 Implications for a theory of culture
and prototypes as models who exemplify the behaviors that the children
have come to value and incorporate into their own self concepts.
Let us return to the example of Paula's self-reliance. Paula is typically
American in believing in the importance of being self-reliant (e.g., Bellah et
al. 1985:55-65; Harkness, Super, and Keefer 1992; Hsu 1972:248-252;
Wierzbicka 1993:223). It is no wonder that she is highly motivated to be so.
Her parents found many opportunities to encourage her attempts at
self-reliance. They rewarded her for these efforts by showing approval if
she did something "all by yourself like a big girl!" and punished her for
lack of self-reliance by chiding her disapprovingly if she asked for help too
often. She was their oldest, and they expected her to be very grown up.
Beginning when she was quite little, the pleasure she felt when she got
approval for doing something on her own made her want to do other
things on her own; her feeling of unhappiness when her parents were
disapproving of her for being dependent made her want to be less depend-
ent the next time. When she was a little older, too, playmates and school-
mates made fun of children who acted babyish, and she did not want to be
the butt of such ridicule. When she was young the reward of approval and
the punishment of disapproval motivated her directly to be self-reliant; as
she got older, this approval and disapproval were gradually attached to
ideas about herself as a person that she was learning in a multitude of other
contexts. As a teenager, she wanted to grow up to be like her mother's
extremely independent younger sister, who did not come to visit very often,
but who maintained a special relationship with Paula, writing encouraging
woman-to-woman letters to her from all the places she traveled as a foreign
correspondent. As an adult herself, Paula has come to be proud of her
achievements and of her competence when she accomplishes something on
her own; disappointed in herself and ashamed to admit it when she needs
We do not suppose that the force of the desire to be self-reliant, a
preoccupation so striking that US Americans as a people are widely
characterized in its terms, can be explained entirely as a product of shaping
by approval and disapproval. For one thing, when values like self-reliance
are especially significant to them, parents and other caretakers are likely to
go further to inculcate them and begin doing so earlier in children's lives.
They will deliberately design the environment and structure their own
behavior in it, to give the child experiences that will motivate these values.
This is a particularly useful teaching technique with infants, who would
not be able to fully understand verbal instructions or admonishments
about what they were expected to do, and might not be able to do it, in
order to win approval and avoid disapproval. We can use a West German
example supplied by Robert LeVine and Karin Norman (1994) to make
Two properties of culture 107
this point about teaching infants self-reliance. As the example also testifies,
US Americans are not the only ones who care about this value, nor even
the most assiduous about teaching it.
The German parents in this study,
for example, not only put their babies to sleep in a separate bedroom, as
Paula did hers, but follow other practices that Paula never would have
considered: They may leave an infant alone in its room for an hour or more
after it awakes, and alone in the house while they go shopping; they refrain
from comforting it immediately when it is distressed, and worry more
generally about "spoiling" it with excessive attention or too much accom-
modation to its demands (ibid.:5-7). These authors also observe that
American childrearing between 1920 and 1950 resembled the German
childrearing practices they report. "American mothers were told and often
believed," they remind us, "that infant schedules were all-important, rapid
response to crying could spoil a baby, it was necessary for babies to play
alone and control themselves as early as possible" (ibid.:9).
As LeVine and Norman (1994:8) point out, once engineered, such infant
experience establishes motivation to be self-reliant and a baseline of habit-
ual self-reliant behavior on which further self-reliance training can be
built. This early motivation should be a particularly effective foundation
for later learning because it will be not only particularly lasting, as we
asserted in the last section, but also, as we will argue momentarily, particu-
larly powerful. The example of self-reliance provides an opportunity to
indicate still another source of motivational force that gains its intensity
from its roots in earliest experience.
The explanation for American self-reliance might end, for some theor-
ists, with the description we have given so far of self-conscious sociali-
zation techniques. We are led to ask, however, why US Americans think it
so important to teach their children self-reliance so early, and, correspond-
ingly, why they have such a deep aversion to behaving in any way that feels
to them like not being self-reliant - so that they sometimes cannot bring
themselves, for examples, to ask good friends for help even in dire circum-
stances, or to apply for food stamps even when they qualify. A full
explanation for the strength of Americans* need to be self-reliant, we are
persuaded, requires attention to the psychodynamic processes that charac-
terized the infantile experience of people like Paula. A product of this early
experience, she would have already been strongly predisposed to self-
reliance by the time her parents urged it upon her.
It is a fundament of psychoanalytic theory,
and one that we believe
must be incorporated into any more general account of the motivational
force of culture, that infantile experience has a powerful effect on moti-
vation (and, as we pointed out in the previous section, a lasting effect on
learning). The neoassociationist model we are developing here supports
108 Implications for a theory of culture
this assumption. As we observed in the last section, the feelings activated
by very early experience will be strong ones, and to the degree that these
early feelings are tied to emerging ideas about being good and being bad,
these ideas, the emotions attached to them, and the motivations to be good
and not to be bad that these emotions activate, will also be exceptionally
powerful. Jessica Benjamin (1988) makes this case for self-reliance. She
argues that US Americans' earliest evaluations regarding goodness and
badness are gendered, being tied for both girls and boys to the idealization
of freedom and its epitomization in the father and in men, and the
devaluation of nurturance and its attribution to the mother and to women.
In conclusion to a rich and compelling argument, Benjamin writes:
the deep source of discontent in our culture is not repression or, in the new fashion,
narcissism, but gender polarity. Many of the persistent symptoms of this discon-
tent - contempt for the needy and dependent, emphasis on individual self-reliance,
rejection of social forms of providing nurturance - are not visibly connected to
gender. Yet in spite of the fact that these attitudes are almost as common among
women as they are among men, they are nevertheless the result of gender polarity.
They underlie the mentality of opposition which pits freedom against nurturance:
either we differentiate or remain dependent; either we stand alone or are weak;
either we relinquish autonomy or renounce the need for love. (Benjamin
In Benjamin's (1988:169-171) fuller argument - to which we cannot poss-
ibly do justice here - pervasive gender polarity is in turn the complex result
of a cultural configuration in which the liberating, exciting father is op-
posed to the holding, nurturing mother, the possibility of separation being
split apart from that of connection and valued over it (ibid.: 169-170), and
the issue of separation and individuation being recast as one of gender
(ibid.: 104). One side of the resulting opposition between separation and
oneness, the mother and femininity elicits helplessness and fear of engulf-
ment, feelings that can only be countered by adherence to an idealized,
masculinized self-sufficiency. In our terms, these strong feelings and the
splitting that is the defense against them form the original core of US
Americans' schema for self-reliance. Positing such a predisposition does
not contradict the explanation of self-reliance we have already given in
terms of parental approval and disapproval and parental engineering, but
suggests that this conscious teaching has deeper roots. These deep, uncon-
scious feelings may enhance people's determination to teach their children
to be as self-reliant as possible, and their attraction to social institutions
that reinforce this lesson. These feelings may also be what motivate enact-
ment of public policies in the United States that ensure people have to
learn to be self-reliant, even for such basics as food, shelter, and medical
Two properties of culture 109
Centrifugal tendencies against motivational force
As deeply as Paula learned to want to be a self-reliant person and a good
mother, like so many women her age she finds that the two goals not
infrequently come into conflict in her adult life, and that, unhappily and
uneasily, she must then choose between them. Her greatest sense of self-
reliance comes from her job, requiring, as it does, much independent
decision making and autonomous problem solving, and providing her with
a source of income that she can call her own. But in order to maintain her
involvement in ongoing projects, and be considered seriously by her bosses
for new projects and for eventual promotions, she had to go back to work
earlier than she wished to after each of her children was born. On the other
hand, she is acutely conscious that the accommodations she had to make
to her children's needs in the course of their childhoods have adversely
affected advancement in her career; for example, she has had to pass up
some prize assignments that would have taken her out of town too often,
and it is always her colleagues, not she, who volunteer to put in overtime at
crucial junctures in project completion. Such cruel if inevitable conflicts
between culturally shared ideals mean that otherwise strong motivations
to pursue these ideals cannot always be acted on by those who have learned
In contrast to her fierce self-reliance and her determination to be a good
mother, Paula has built up other complexes of associative links that are
relatively unattached to feelings and desires in her. For example, Paula's
parents grew up during the Depression and, as a result, often told stories
about their near-poverty in those days and impressed upon Paula the need
to "save for a rainy day." From these stories and exhortations Paula built
up an associative network relating money to the importance of saving. Yet,
though Paula's parents' concern about saving for a rainy day was linked
closely to their traumatic Depression experience and their anxieties about
ever going through that again, it did not hold more general importance for
them as a value. If they had been as concerned to transmit the value of
saving as they were to transmit the value of self-reliance to their children,
they might have contrived ways for Paula to practice saving, and resorted
to additional measures to make Paula feel good about herself when she
saved money and bad about herself when she frittered it away. As it was,
they did nothing more than exhort. Indeed, Paula tended to shut out these
exhortations and the accompanying litany about how her parents had to
sell soda bottles for grocery money and otherwise scrimp and save during
the Depression, because she grew so tired of hearing it all. Moreover,
because Paula herself grew up in the comfortable 1950s, she never learned
through hard economic experiences of her own to associate strong emo-
110 Implications for a theory of culture
tions (for example, of shame or insecurity) and motivations to their
homilies about saving. As Arlene Skolnick (1991:83) has observed about
Paula's generation, "The comfortable way of life that seemed so miracu-
lous and hard-won to the parents was simply taken for granted by the
offspring"; Skolnick goes on to quote Todd Gitlin (1987:20) on this same
phenomenon: "Parents could never quite convey how they were haunted
by the Depression and relieved by the arrival of affluence; the young could
never convey how tired they were of being reminded how bad things once
had been, and therefore how graced and grateful they should feel to live
normally in a normal America."
Nor was there anything in Paula's earliest experience that made saving
money a psychodynamic issue. Consequently, as an adult, she never has
put much of her income into savings, although she thinks, in an abstract,
detached way, that she ought to. Moreover, countering this weak impulse
to save, and undermining any plan Paula might ever have to act on it, she
confronts constant, persuasive public messages, explicit and implicit, that
encourage unnecessary spending, and many actual examples of such
spending that seem to her to have positive outcomes. Any number of
further examples could be given of social evaluations to which Paula has
been exposed but which never acquired motivational force for her (for
example, her parents' double standard about premarital sexual behavior,
and her mother's elaborate rules of etiquette and fastidious standards of
More generally, not all cultural schemas acquire affec-
tive and motivational force for people; to understand which do and which
do not, we need to learn the particularities of a person's experiences.
Three further properties of culture
The durability of culture in the individual and its motivational force are
not the only properties of culture to be accounted for. The historical
durability of culture, its thematicity, and its sharedness, as well as the
circumstances under which these properties do not hold, are the topics of
this chapter. Some of the arguments and assumptions developed in the
detailed discussion of the previous chapter have paved the way for dis-
cussion in this one.
Historical durability
Something else we recognize about culture is its historical durability - that
it can be, and often is, reproduced from one generation to another.
This is
its third centripetal tendency. Yet, some cultural ideas stay around much,
much longer than others, and some do not last even a generation. We next
turn to this question of what gets reproduced in the next generation and
We point out, in the next section, that historical durability results when
the public world is recreated by enactment of the schemas each generation
has learned; but also when one generation intentionally transmits its values
to the next. We go on to describe how durability across generations is
sometimes also promoted by people's deliberate efforts to preserve
cherished practices; by the representation of heretofore private matters in
public forms, which then preserve and propagate these understandings; or
by the storage of cultural understandings in books and like repositories
where they may lie dormant over long periods of time before being
retrieved and accorded new life. The chapter then turns to circumstances
under which, conversely, cultural understandings are unlikely to endure
over time. We note the possibility that, just as existing behaviors recreate
the public world from which the next generation learns, new patterns of
behavior may alter this world. We explore how changing values about
what is worth transmitting alter what one generation teaches the next. We
consider, also, how historical change is helped when given social practices
112 Implications for a theory of culture
are labeled the objects of efforts at intentional social change. And we
discuss how the same repositories that sometimes serve to perpetuate the
ideas they preserve, can also foster change, when these ideas undergo
reinterpretation (even as their historic sources lend them a sense of conti-
nuity with the past).
Centripetal tendencies toward historical durability
The durability and motivational force of schemas for individuals have
consequences for their historical durability. First of all, when people are
motivated to enact and reenact the schemas they have learned from their
own experience, they recreate the public world of objects and events that
they knew, reproducing patterns of experience from which the next gener-
ation learns. For example, when Paula and Michael drift back into the
gender roles they learned growing up, their behavior becomes part of the
observable world that shapes their children's schemas. The children's
schemas are also shaped by the behaviors they learn about in other homes
they visit, at school, and in books, on television, in movies, and so on. To
the degree that these other behaviors tend to the mode or idealize the past,
their children's understandings surprise Paula and Michael by how tradi-
tional they often are (Weisner 1990).
This point about the shape of the environment structuring successive
generations' schemas is the essence of Bourdieu's account of cultural
reproduction. At its most unproblematic, this process is akin to individuals
acquiring habits from repeatedly observed and practiced patterns of be-
havior, described in the last chapter; that is, just as any given generation of
individuals is predisposed to behaviors that are familiar to them and
appear to be the only ones available, natural, or even imaginable, so the
next generation is predisposed to those same familiar, available, natural
behaviors being reenacted by the adults around them, (There is the old
joke about the woman who always cut an end off her roast before putting it
into the oven. When her daughter asked her why, she explained that was
the way her mother had always done it. The girl pursued the matter with
her maternal grandmother, who reported that that was the way her mother
had always done it. The next time the girl was taken to visit the great-
grandmother in the old age home, she asked her why she had always cut off
one end of her roast before cooking it. The old woman answered, "My pan
was too short.") Furthermore, the world of objects, people and events
from which children learn can have emotional resonances and moti-
vational implications as well as informational content. Thus when Paula
and Michael recreate the patterns of their own childhood experience in
their childrearing practices - for example, putting their children to sleep in
Three further properties of culture 113
separate bedrooms from an early age - the patterns so replicated may carry
and reproduce lasting emotional and motivational effects as well.
Practices like this one often seem like such a natural way to do things
that they go undiscussed; as long as they continue to go undiscussed, no
language for talking about them develops. The absence of any such lan-
guage to talk about an experience makes it less likely to come under
sustained and widespread scrutiny; and, not being the object of such
reflection, it is simply continued. Social problems, thus, can be especially
persistent when they lack a language for articulating them. When they do
not even have labels, as in the famous instance of the "problem that has no
name," identified by Betty Friedan in The Feminine Mystique (1963), such
problematic experiences are notoriously difficult to communicate and
validate. Paula, who reached adulthood a decade after the publication of
Friedan's book, is eloquent on the problem that once had no name but by
then had been not only named but endlessly dissected: the right of women
to find fulfillment outside the home. This is not to say that other of her
experiences as a woman have been easy for Paula to formulate. Like many
women of her cohort she can remember upsetting experiences in college
that she never told anyone about and that she has only recently learned to
identify and condemn as instances of sexual harassment. As long as each of
these women's problems went unmentioned, it persisted unquestioned.
Living in and experiencing a world shaped by the last generation's
everyday practices is not the only determinant of cultural reproduction, as
Bourdieu's account makes it seem to be. We have already argued that the
process of teaching imparts durable cultural understandings to new
learners. Here we need only add that there is a significant tendency for the
content of what is taught to be reproduced over successive generations.
This is because members of any society have, often among their most
motivation-laden cultural understandings, certain ideas about what values
must be transmitted to the young (see, e.g., Whiting and Whiting 1960:920;
LeVine and Lun 1994:10-11 for this point made about parenting cross-
culturally) and what knowledge must be preserved for future generations.
Thus, in addition to unintentionally recreating the world they know by the
schemas they enact, people also act intentionally to pass on practices and
beliefs that they value - to insure that certain of their own enduring
schemas will become the enduring schemas of the next generation of
individuals. Because Paula and Michael have learned to value self-reliance
in themselves and others, they not only enact that value (thereby inciden-
tally modeling self-reliant behavior for their children), but also transmit it
quite deliberately to their children, often relying on child-rearing practices
like the ones through which self-reliance was developed in them.
Sometimes a changing world can prompt people to apply deliberate
114 Implications for a theory of culture
TIME Att> HAVE A ^^mT7tL-tEr ^ 2 ^
AM I &EIN& ^ K ^ B /
Figure 5.1 Sally Forth comic strip, originally published April 13,1994
effort not just to enact but to preserve valued practices learned growing up
but felt to be slipping away. For example, Paula believes that a family
should be close-knit, and attaches a great deal of importance to regular
sit-down family dinners as a way of promoting this ideal. The hectic pace
and unpredictable hours of her job and Michael's, and the difficulty of
coordinating four people's dinnertime schedules as their children grow
older and pursue independent after-school activities, make it all too easy to
slip into a style of individualized pick-up eating. Now Paula wages a
deliberate campaign to preserve sit-down meals at least a few evenings a
week. To the extent that she and other mothers like her succeed, family
sit-down dinners continue for another generation. To be sure, they survive,
not as they once did as a matter of habitus, but now as a result of conscious
Private matters like mothers' personal crusades to save the sit-down
dinner also have a good chance of attracting public notice, to be repro-
duced in public forms. The 1994 Sally Forth comic strip shown above
captures just the tension Paula experiences between the value many US
Americans put on sitting down together for dinner and the reality of their
busy lives.
As this chapter is being finished in 1996, a more aggressive approach to
the maintenance of sit-down dinners has appeared in US television ads for
a new kind of take-out food, one being marketed as healthy and substan-
tial enough to be the basis for a "real" family dinner. In each of these ads, a
mother is pictured as using fantastical means to get the family to the dinner
table at the scheduled hour. (In one, failing to get the band leader to
change the time of band practice so it does not conflict with dinnertime, the
mother dispatches him in a rocket. In another, she lays a deadly booby-
trap that is tripped by an unsuspecting caller who has made the mistake of
ringing the doorbell at dinnertime, and in still another, she rigs the TV so
that the children will have nothing good to watch at dinnertime and will
come to the table.) In the last frame of each ad, the mother says ominously,
"Don't mess with dinner." Like the comic strip, the ads illustrate how
Three further properties of culture 115
certain public products target just such widespread dilemmas - ones that,
while experienced privately by many, have heretofore gone unshared and
publicly unrecognized. Patterns like these are grist for public humor just
because these individual efforts have had no established place in shared
understanding or in public practice. They elicit "ahas!" of amused recogni-
tion that, without knowing it, we and others have been worrying about the
same problem and improvising solutions to it (and, as in the case of Sally
Forth, sometimes giving in to compromise solutions). Public representa-
tion of these heretofore private matters not only preserves them but lends
them social reality, making it more likely that they will be thought, talked,
and written about in future.
These last examples should not obscure the point that historical durabil-
ity rests as well on the fact that old ideas can be expressed and preserved
outside of minds and in the absence of any sustained effort to reenact or
recreate them - for example, in oral narratives (e.g., Ortner 1989), in
books, drama, music, architecture, and in tools, routines and the organiz-
ation of groups (e.g., Hutchins 1990b; Lave 1988). Thus it is that dormant
idea systems like Nazism or romanticism have the ability to gain renewed
However, as we point out in the next section, such external reposi-
tories of cultural understandings can also become effective vehicles of
Centrifugal tendencies against historical durability
Our account should have made clear that the understandings of Paula and
Michael's children will also diverge from their parents* because the
children have grown up observing different patterns of behavior; not only
when their parents drift back into old gender roles, but also the times when
Paula and Michael realize their new beliefs and share housework more
equally; not only sit-down family dinners but also new compromise ver-
sions of these like the sit-down take-out dinner depicted in the Sally Forth
strip. Moreover, behaviors, like sharing housework, that require a deliber-
ate effort for Paula and Michael will shape what their children come to
take for granted. When this new generation reaches their thirties, forties,
and fifties, their assumptions about what is natural and desirable will
further reshape the public symbols and institutions from which still an-
other generation acquires their schemas. These new patterns of behavior
and experience can also have profound psychic effects. For example, to the
extent Paula and Michael alter traditional childrearing practice to divide
childcare responsibility evenly between them, rather than letting the re-
sponsibility rest primarily on Paula - as it did on her mother and Michael's
- they may unwittingly alter the psychological dynamic of gender identifi-
116 Implications for a theory of culture
cation and consequently the way their children think, feel about, and enact
their genders (see Benjamin 1988:203-205; Chodorow 1978:217-218).
In addition to modeling new behaviors for its children, the parental
generation may have changed its explicit values about how these children
should be reared. Paula and Michael may not be teaching their children to
be as extreme in their self-reliance as they were taught to be (and, as we saw
in the last chapter, as West German children are still taught to be), or to
put quite the same construction on being self-reliant. This might happen
because some of the techniques through which Paula and Michael were
imbued with self-reliance as infants have come to seem too harsh to them,
due to the spread of what LeVine and Norman (1994:9) characterize as "a
new ideology of affectionate and emotionally responsive maternal care" in
the United States after World War II and especially beginning in the 1950s.
To make our point about the necessity of studying intrapersonal meaning
again, we cannot divine from studying a given generation's public practices
alone which of them will be continued in the next; we must discover which
practices people believe merit being continued, and so teach the next
generation, and which they do not.
Deliberate instruction, too, can lead to generational change just as it can
lead to cross-generational reproduction of values. Paula and Michael not
only model more gender egalitarian behaviors than they observed growing
up, they also deliberately impart such values. They refrain from imposing
on their children traditional boys' and girls' clothes, toys, activities, house-
hold roles, and expectations about what they can do and should become.
They also encourage their daughter to play sports and their son to express
his feelings, read "liberated" versions of fairy tales to their children, and
comment on sexist practices the children have observed. Whether such
instruction results in significant changes in their children's practices and
outlooks will depend in part on how cooperative the rest of the world is.
Just as we discussed regarding Paula and Michael's efforts to change their
own behavior, for their teachings to result in changes in their children's
gender-typed behavior and identities, parallel widespread changes in many
extradomestic social institutions would be necessary.
Earlier in this section we observed how experience that goes unlabeled
can persist over time. Labeling previously unarticulated experience is
another way in which historical change can be initiated. Paradoxically,
fixing in language experiences that are problematic but previously un-
articulated - like the one to which Betty Friedan gave a name and other
problems that have been labeled feminist issues - can be an important step
in social change directed to the elimination or amelioration of those
problems. This is because, as we said in chapter 2, naming can link
disparate understandings in a new schema, lending the named experiences
Three further properties of culture 117
a social reality not unlike the way representation in public media like TV
ads and comic strips does. (This is also the way in which putting psychic
experience into words is said to facilitate individual therapeutic progress.)
Having been distinguished and articulated, moreover, the problematic
experience can be more easily shared. Shared, the new social reality can
more easily become a legitimate social concern.
When troublesome experiences disappear or alter, of course, so do the
shared schemas that people derive from them. This can lead to another
paradoxical effect, one that has been widely observed and discussed in
both scholarly debate on contemporary feminism and its coverage in the
popular press. Since many women in the current generation have not
experienced many of the most trying frustrations suffered by Paula's
generation and those before it, they are not inclined to see their situation as
women as being terribly problematic. Thus, even though feminism was
largely responsible for the more progressive social climate these women
now enjoy, this very climate has made them more likely than the past few
generations to dismiss feminism as irrelevant to their lives. Paula and the
other woman manager her age at work, for example, have had a hard time
getting their younger female colleagues to join them in protesting what she
is sure is a slower rate of advancement for women managers.
We observed that old ideas can be preserved outside of minds in books,
architecture and the like. New ideas, as well, can be disseminated through
the same vehicles, which can then contribute to historical change. Indeed,
the sense of continuity conferred by the presence of the old can be decep-
tive, when interpretations of old texts take decidedly new twists - as, for
example, have exegeses of the Bible over its several-thousand-year history.
Moreover, these texts and other cultural products stay around for chance
encounters with later people from anywhere in the world, which can lead
to their incorporation into new and quite different contexts than originally
intended. Thus, for example, many Christians read the Bible today and
seek application of its messages, conceived and written in a distant time
and place, to their contemporary lives. Other examples closer to Paula's
life are the woman she knows who swears by homeopathic medical treat-
ments translated into English from centuries-old Asian pharmaceutical
texts, and another friend who was also born in the United States and grew
up a Baptist, but is now a practicing Buddhist and follows a Buddhist
regime of daily meditation.
It is also important to bear in mind that the mere presence today of these
artifacts from the past does not guarantee their continued force; unless the
appropriate understandings and feelings about them and the appropriate
motivations toward them are passed on (even if in altered form) across
successive generations, the things themselves will be as lifeless and uncom-
118 Implications for a theory of culture
municative as the metal, stone, and ink of which they are composed. A
mundane example is Paula's living room which, like those of many Ameri-
cans, has lost its function as a separate and more formal space for hosting
non-household members, and now sits empty a great deal of the time,
while company tends to join the family in the TV room or on the patio,
congregate in the kitchen before dinner, and sit around the dining room
table long after dinner is finished. It is especially hard to predict what
things will retain their meaningfutness and, if they do, what meanings they
will have across time and complex historical shifts in the contexts of their
use. These observations reinforce our point, once again, about the necess-
ity of studying intrapersonal meaning.
The fourth property of culture, one amply illustrated in the ethnographic
literature, is the tendency of some schemas to be evoked in a wide variety
of contexts. Examples would include complexes of understandings about
honor in the circum-Mediterranean area, as described by Bourdieu (1977)
for Kabyle society, or about rivalry in Sherpa society, as described by
Ortner (1989), about the Western discourse of normality that Michel
Foucault (1972a) describes, or about self-reliance in the United States,
which we will describe further below. As Foucault (e.g., 1979:23) makes
clear, a theme or "discourse" can spread even across contexts separated by
the boundaries of distinct subcultures, such as those that comprise the
separate professional disciplines of social science and penal law. This
thematicity that culture exhibits, again, depends upon a complex interplay
between properties of the culturally constructed world and properties of
the mind. As we will consider, conditions are only sometimes right for
thematicity to result: Not all cultural schemas become cultural themes.
A theme is most likely to spread across domains of experience, we note
in the following section, when these experiences bear a family resemblance.
Understandings gained from early learning and those gained from a wide
variety of contexts are more likely to be generalized, we reason. We next
give some attention to the process of "elective affinity" by which rhetoric
and products embodying themes that appeal to people are elected, so that
more such public discourse and commodities are produced, further dis-
seminating the themes that they embody. On the other hand, we point out,
there is no necessity for themes to spread across contexts, and often they
do not. Both individual minds and the public world exhibit inconsistencies.
Centripetal tendencies toward thematicity
Paula and Michael generally say "please try to do it yourself when their
children ask for help (unless the parents think the obstacle faced is some-
Three further properties of culture 119
thing that the children can not overcome by themselves). As a result their
children have come to associate obstacles of many different sorts with the
expectation that they will overcome them on their own. Nor will this
expectation be triggered only by future situations that are exactly the same
as those the children have encountered in the past. As conceptualized in
connectionist models, we have said, schemas are not rigid structures with
precisely stated rules of applicability. New experiences evoke an interpre-
tation based on overall similarity of features of the current experience to
repeated or particularly memorable combinations of features of previous
experiences. So in the future, any new obstacle that Paula and Michael's
children encounter will evoke their "do-it-on-your-own" expectation, pro-
vided that it has the right sort of family resemblance to previous occasions
when they learned to handle a difficulty by themselves.
As psychoanalytic thinkers since Freud have known, schemas learned
very early are likely to be widely generalized to new experiences. This
observation makes sense in connectionist terms, too. Since human infants
start with a grossly undifferentiated understanding of the world, the very
first associations they make serve them as models for much of their later
experience. Furthermore, since, as we have pointed out, these earliest
understandings are very durable, they are available to serve as models for
later experience. As we have been illustrating, self-reliance is one such
model available to people in our society. Another is a set of understandings
around accomplishment and success, to be considered in chapter 6. The
thematic effect of culturally patterned early experience was the focus of a
well-known tradition of research in mid-century American cultural an-
thropology, under the rubric of national character studies (most notably,
Benedict 1946; Gorer and Rickman 1950).
Schemas are also more likely to be generalized if they are learned,
initially, from a wide variety of contexts. Home and summer camp are only
two of the contexts in which middle-class children in the United States
learn self-reliance; other such contexts are day care, school, and extracur-
ricular activities like Scouts and Little League. It is not only that these
institutions reinforce self-reliance; they also broaden its definition. For
example, overnight summer camps give middle-class children the experi-
ence of living away from their families, while US schools stress the import-
ance of originality and independent work. Connectionist models would
lead us to expect that the wider the range of contexts in which behaviors
like self-reliance are learned, the greater the likelihood that any new
context will resemble one of these earlier experiences, and thus the schema
that has been built up from them, thereby evoking self-reliance as the
appropriate response.
A further dynamic contributes to the widespread recurrence and institu-
tionalization of themes like this one. Paula, Michael and their children
120 Implications for a theory of culture
have an "elective affinity" (Weber 1946) for new rhetoric and products that
appeal to their understandings about self-reliance. Because they find it
natural to be self-reliant, they are unlikely to question the assertion, for
example, that the United States must supply its own energy needs, or that
welfare dependency is bad, and they are likely to support politicians who
invoke these assumptions, and reject politicians foolhardy enough to try to
alter the terms of debate and suggest that we live in an interdependent
world or that we must help the poor. In addition, because they have come
to attach positive feelings to being self-reliant, Paula and her family enjoy,
and regularly participate in, pastimes such as backpacking that demand,
and indeed test, this quality in them. Finally, because they find self-reliance
desirable, they are attracted to stories such as the Nancy Drew books,
television programs such as Little house on the prairie
and the continuous
stream of movies such as Home alone (and more recent ones sure to occur
to the reader) that celebrate it and confirm the value they place on it.
Because many people in our society share understandings about self-
reliance, these kinds of public discourses, pastimes, and forms of entertain-
ment "sell" well. This popular response, in turn, gives messages to the
powers-that-be about future policies and products to develop and ways to
market them, playing a role in the continual shaping and reshaping of the
landscape of public objects and events from which further cognitive associ-
ations are learned.
We may surmise that it is largely through this cyclical
process of production and election that the discourses and technologies of
interest to Foucault are distributed widely across disciplinary subcultures
- although it is not popular response, but the response of various profes-
sional elites, that is relevant to his case.
Centrifugal tendencies against thematicity
Having stressed the ways in which themes can spread across contexts, we
must emphasize that they need not do so. There is no cultural dynamic
ensuring that the members of a society will come to agree upon a wholly
cohesive world view as earlier, functionalist anthropological accounts
assumed would be the case, and no constraint against the possibility of
multiple, even contradictory, cultural themes emerging. This possibility
explains why attributions of personality types to individuals or - as were
made in national character studies and other culture-and-personality stu-
dies - to cultures, have generally overstated the predictability and consist-
ency of the individual or group of individuals so typed. (Lindesmith and
Strauss 1950 and Wallace 1961:84-119 pointed out this problem with
culture and personality studies at the time.)
Thus Paula, like other Americans, is not always consistent in the
Three further properties of culture 121
schemas she enacts. Thought of in connectionistic terms her self-reliance
schema, for example, is not an overarching rule that programs her to
overcome all obstacles on her own in every situation. Instead, it is a
network of links among units representing certain types of situations in
which she was expected to be on her own, as a result of which she finds it
natural and desirable to be self-reliant in situations similar to those. She
has built up this network from her experience of particular cases. New
cases only activate the goal of self-reliance if they have enough features in
common with previous situations in which it was expected of her. (Of
course, we can be taught general rules that, applied to new situations, may
weigh heavily in our experiencing them as similar to or different than
earlier ones.) Thus, even if it might seem to an outside observer that given
situations are alike, they may be experienced so differently that each is
linked to different interpretations and responses. We know that when
Paula was growing up she learned that self-reliance was generally re-
warded. Yet she also learned that there were occasions when no one
expected women to be self-reliant: on the contrary, when a car had a flat
tire, for example, or a faucet was leaking, a feminine woman was expected
to act helpless. Paula experienced the occasions for acting feminine and
helpless very differently from the way she experienced those for acting
self-reliant. As far as she could see from both fictional and real-life
examples, things seemed to work out well for women who acted helpless in
those flat-tire, leaky-faucet situations.
Nowadays when Paula's car gets a
flat tire or the faucet is leaking or there is some other problem that involves
heavy or dirty, "male" activities, it seems natural to her to let Michael take
care of it. To an observer Paula's behavior may seem inconsistent, but
Paula is unaware of the inconsistency, because her self-reliance under-
standings and her feminine-helplessness understandings are represented in
different parts of her neural network and are triggered in different sorts of
contexts by very different features of experience, so they rarely come into
conflict. The learning from specific experiences that is posited in connec-
tionist models and the nonlinkage of some experiences that can result,
make it easy to understand how this could be so. If knowledge is organized
experientially rather than logically, there is no reason why all of Paula's
contradictory understandings about self-reliance should be interlinked.
The processes of such "compartmentalization" or "containment" of
schemas as well as their limits - we are not completely fragmented selves -
are the topic of research by Claudia Strauss, described in chapter 8. (See
also Strauss 1990 and Strauss and Quinn 1991.)
It needs to be added that there is no requirement, either, that public
culture will be consistent in the messages it conveys, as even the most
casual flip of television channels from the Trinity Broadcasting Network to
122 Implications for a theory of culture
almost any daytime talk show makes vivid. This inconsistency in what is
imparted by public institutions, rituals, and representations - even in
so-called "simple" societies- combines with the human cognitive capacity
for encapsulating information to help explain how contradictory schemas
come to coexist in a person's head.
As observed at the beginning of part II, sharedness is for most of us the
defining property of culture. As we have also indicated, cultural schemas
differ not at all from other schemas learned from humanly mediated
experiences, except in being shared. Schemas unique to individuals are
built up from idiosyncratic experience, while those shared by individuals
are built up from various kinds of common experience. Some schemas are
undoubtedly held in common by millions of people, shared by everyone
growing up and living his or her adult life in the same nation state, with its
hegemonic influence and its capability for broadcasting popular events to
the far limits of its boundaries. Other schemas are shared by only some
individuals - subgroups forming self-conscious subcultures, or experts
who have undergone the same esoteric training, or individuals who happen
to share similar experiences (cf. Rosaldo 1989:28-30), for example. At
what point in the continuum of sharedness we decide to call a given schema
"cultural" is simply a matter of taste; we may, if we wish, speak of a family
or workplace "culture," or "high culture," or a "subcultural theme."
Cultural schemas are in people's heads, but a given cultural schema need
not be in everyone's head.
Since subgroups such as households, genders,
regions, ethnicities and historical cohorts cross-cut one another, an indi-
vidual may share schemas with as many different groups of other individ-
uals as he or she shares a history of like experiences - at the same time
sharing all of his or her schemas with no one else in any of these groups. In
this final section of part II, we address the question of how patterns of
relative sharedness and nonsharedness come about.
Sharedness, we begin the next section by pointing out, requires not that
people have the same experiences, but only that they experience the same
general patterns. We observe that much of the social world is organized to
ensure just this. We go on to consider, in addition, how historical events
and trends can affect cohorts of people similarly, and how shared experi-
ence can also be enforced, through the exercise of power. Next we draw
attention to three notable sources of modal patterning - common lan-
guage; uniform child care and socialization practices; and shared task
solutions. We devote some discussion to the third of these, noting how
ubiquitous such task solutions are and delineating how they arise and
Three further properties of culture 123
spread. We consider the more general instance of cultural understandings
that become widely shared because they have wide appeal. Turning, lastly,
to the conditions that mitigate against sharedness, we discuss geographical
distance and several different kinds of social impediment to interaction,
and, hence, to like experience. Finally, we observe that individuals often
experience similar circumstances differently, bringing to them different
past experiences and, hence, different interpretations, feelings and moti-
vations, and saliences.
Centripetal tendencies toward sharedness
Interplay between the public world and private psyches, once again, helps
explain why certain schemas are so widely shared in a society. It is not
necessary that two people have exactly the same experiences (and no two
people ever will), for them to arrive at some of the same schemas. Accord-
ing to connectionist models, the most frequently encountered patterns of
associations among features are internalized as strong connections, while
random variations around the mode are represented by weaker connection
weights that have much less effect on cognitive processing. (This leaves
aside the effects of emotional modulation, which we will ignore for the
moment.) Thus, to learn about the nature and value of self-reliance, it does
not matter that Paula acquired this cultural model from parental homily
A, personal experience B, observation of model C's behavior, and movie
D; while Michael acquired it from a different assortment of experiences.
Since both sets of inputs reveal the same general patterning of associated
features, that pattern is represented by strongly interconnected units in
each of their cognitive networks. Experiencing many of the same general
patterns, people in a society will, to some extent, come to share the same
understandings, have common emotional and motivational responses, and
exhibit like behaviors. Not any public message is destined to become a
widely shared understanding. Rather, the messages that become most
widely shared will be those stored in widely distributed cultural products,
so that most people will encounter them in some form or other.
Much of the world is organized in exactly such a way as to ensure that
people in the same social environment will indeed experience many of the
same typical patterns. This modal patterning is broadly characteristic of
human social life, a requirement of many of the practices by which people
interact with each other, share knowledge, coordinate common activities,
collaborate in common ventures, play the established roles expected of
them, and otherwise conform to the laws of their government and the
conventions and values of their fellows (see D'Andrade 1987b, 1989; Spiro
1987b: 112), as well as model and explicitly teach these common laws and
124 Implications for a theory of culture
values to others. To cite two mundane but not inconsequential examples:
Paula and her friends who have children and millions of other suburban
mothers all find themselves caught up in the common experience of incess-
ant driving that is dictated by the organization of middle-class children's
activities in combination with the pattern of suburban housing and reli-
ance on automobiles that developed in the United States in the middle
decades of the twentieth century.
Another experience common to Paula
and the other mothers she knows (and other networks of mothers all over
America), throughout late winter and early spring, is the exchange of
information they have located about summer camps and other summer
activities, and the rush to sign up their children early for these programs, in
order to keep the children busy and supervised during working hours.
They find it necessary to do so because of the cultural tradition of summer
vacation from school, a vestige of long-ago days when most Americans
lived on farms and children were needed for planting, harvesting and other
intensive seasonal agricultural tasks.
Historical events and trends affecting whole cohorts of individuals are
another source of shared patterns. Paula's mother is hardly the only
woman of her generation to have resented the gendered division of house-
hold labor, nor is Paula the only daughter of such a woman who deter-
mined she would organize her household differently.
The modal pattern-
ing of experience is also, we must not forget, a by-product, not only of
social movements like feminism but also of the exercise of power from
above. Participation in certain institutions and allegiance to certain ideolo-
gies are enforced because this suits the interests of people in positions of
power. In this way, cognitive and social forces interact to determine
cultural dynamics.
Deserving special note are three other sources of modal patterning. The
first is the near constancy and minute detail of the shared experience that
the structure of a common language or dialect provides for its speakers.
While we have argued against a deterministic view of language as a screen
for perception, we certainly recognize the degree to which language shapes
what one notices and easily remembers, and hence the degree to which
speakers of a common language tend to notice and remember the same
features of experience (Lucy 1992). A second source of modal patterning
are the subset of practices that are devoted to the care and socialization of
the young, and that, to the degree that they are conducted collectively or
based on shared knowledge and values, can result in an early experience
that is uniform in ways consequential for the schemas children grow up
sharing. (Whiting and Child 1953 can be considered an early formulation
of this relationship between child socialization and the "characteristic
habits" shared by adults; the research to be described in chapter 7 evolves
Three further properties of culture 125
from this tradition.) Anthropologists are still a long way from pinning
down which aspects of this shared experience are consequential, however
(Ingham 1996:83-85).
Another important source of patterning that causes many individuals to
experience the world similarly are shared task solutions. This subject
demands a somewhat extended discussion of what such task solutions are
and how they arise and come to be shared by many people. Among the
common experiences with which human sociocultural life presents mem-
bers of any community are, inevitably, recurrent tasks, problems, chal-
lenges, contradictions, and conflicts. As solutions to these are invented, the
most useful, practical and appealing of these solutions tend to spread.
Even complex tasks are met by creating and widely sharing the corre-
spondingly complex solutions, such as human language, that they engen-
der. Such cultural solutions are what Edwin Hutchins (1995:290) calls
mediating structures
and, as he has observed, they are ubiquitous in
everyday life:
Many tasks in our culture are mediated by written procedures or procedure-like
artifacts, but even considering all of them would not begin to approach the full
range of mediated performances. Language, cultural knowledge, mental models,
arithmetic procedures, and rules of logic are all mediating structures too. So are
traffic lights, supermarket layouts, and the contexts we arrange for one another's
behavior. Mediating structure can be embodied in artifacts, in ideas, in systems of
social interaction, or in all of these at once. (Hutchins 1995.: 290-291)
Paula is familiar with and uses many, many such cultural artifacts and
procedures. Like the rest of us, and thanks to the introduction of the
low-cost hand calculator, she has forgotten through disuse a great deal of
the arithmetic she learned in school. In another everyday example, Paula,
like many shoppers, likes to go to one particular supermarket, largely
because she knows where in the store to find the items she typically
Taxiing the children, driving to work, and driving around town
on her various household errands, she not only obeys the traffic lights but
anticipates where they are on all the routes she knows by heart to all the
major stores, restaurants, cinemas, civic buildings, and neighborhoods -
the physical structure of her community and her well-learned knowledge of
it together guiding her through these daily tasks. Other people all over the
country, balancing their checkbooks or doing their taxes, repeat the same
set of procedures on their calculators to keep their finances in order. Other
shoppers and other drivers in Paula's city memorize the same grocery
displays and the same routes around town, the structure of their public
world and their shared mental models of these structures mediating shared
solutions to their everyday tasks of buying food and doing errands.
126 Implications for a theory of culture
Such examples bear out the point about the dependency of cognitive
task performance on the material and human environment that is made by
Hutchins and other of the cognitive anthropologists we discussed in chap-
ter 2, under the rubric of "cognition in practice" or "situated cognition/'
Hutchins himself insists that some task solutions are wholly or partially
internalized, as is clear in his reference, in the quote above, to cultural
knowledge and mental models, procedures, and rules. Indeed, as we noted
in chapter 2, a thrust of his argument has been that human beings oppor-
tunistically and in combination employ both in the solution of cognitive
tasks (ibid.:288-3i6,1996:64-65). Nevertheless, much of his work, like the
work of other anthropologists in this tradition, focuses on the role played
in cognitive task performance by the external world - including not just the
material world but also other people, especially groups of people across
whom cognitive tasks and the knowledge needed to perform them are
"distributed." This is an important mission. However, as we also discussed
in chapter 2, it risks neglecting the intrapersonal resources people bring to
task performance in favor of extrapersonal ones available to them. Con-
cerned as we are in this book with intrapersonal culture, we would like to
balance the picture by emphasizing the mediating structures that are
wholly internalized - those that operate without the help of artifacts or
other people's knowledge. Many of these internal mediating structures,
like many artifacts and other public embodiments of mediating structures,
are widely shared. Shared schemas of this kind are as ubiquitous as the
mediating structures embedded in widely used artifacts or widely observed
social practices, and are every bit as important in performing everyday
cognitive tasks.
Before calculators, the mental procedures taught in schools to do multi-
plication, long division, and so forth, would have been one example of a
shared internalized mediating structure. For a trivial example still in wide
use, Paula regularly has occasion, in calculating the days on which future
events will fall, and in writing checks near the end of the month, to recite to
herself the familiar rhyme that begins, "Thirty days hath September..."
until the month in question is encountered, reminding her whether this is a
"thirty-" or a "thirty-one-day" month. The point that these routines are
internalized is not contradicted by the fact that heuristics for doing arith-
metic are written and taught from books, or that the rhyme for the number
of days in the month is orally transmitted from person to person; these are
some of the ways, after all, in which such mediating structures become
shared. Once learned, however that learning is accomplished, such mental
procedures are stored mentally, and used to perform each new task inde-
pendently of external context.
For a less trivial example of such a shared internal task solution, Paula
Three further properties of culture 127
does not experience each conflict and dissatisfaction in her marriage as a
new, unfamiliar problem. Rather, she brings to these events a widely
shared mental model of how such marital difficulties are to be interpreted
and how they can be resolved. In this model, difficulties are understood to
arise from incompatibilities and are resolved through effort. For instance,
Michael, who grew up in a family of limited income, tends to worry
incessantly about debt, while Paula feels comfortable living a little beyond
their means. For another instance, Paula imagines that if Michael really
loved her he would anticipate her needs, but he typically has to be told
what they are. It is these incompatibilities, Paula understands, that have
kept their marriage from always being as happy as it could have been. They
pose difficulties that must be overcome in order for the couple to achieve a
mutually fulfilling marriage. Wanting to succeed at marriage, Paula and
Michael both exert the effort to make it more compatible: they have
worked out a compromise household budget, for example. Paula has
consciously relinquished her cherished fantasy that the man who loves her
will anticipate and meet all her needs, while Michael has slowly learned to
gauge some of them. The issues that have arisen in Paula's marriage, and
the ways she and Michael have dealt with them, differ in their particulari-
ties and emphases from the problems that their married friends have faced
and the solutions that these other couples have devised. Yet, shared by
Paula, Michael, other couples they know, and most other couples in the
United States, is the ready-made set of understandings with which they
address their marital conflicts and dissatisfactions, casting these as difficul-
ties to be overcome by effort in order to achieve compatibility and enjoy a
fulfilling, successful marriage. This cultural mediating structure for rea-
soning about marriage has been the topic of research by Naomi Quinn; in
chapter 6, where this research is described, the structure will be further
elucidated, and evidence for its sharing presented.
Memory devices like that widely used for recalling how many days are in
each month of the Gregorian calendar, and structures for reasoning, like
that Americans apply to marriage, evolve, we may suppose, in response to
recurring cognitive needs. Thus an irregular date system that is not easy to
remember creates a need for a memory trick that frees people from having
to consult a published calendar constantly. Important close interpersonal
relationships like marriage inevitably contain recurrent interpersonal ten-
sions. The possible causes, repercussions and resolutions of such interper-
sonal differences can be varied and complex, and, as we will consider more
fully in the context of shared understandings of marriage in the next
chapter, parties to such relationships are relieved of a great deal of mental
effort by cultural accounts that aid them in identifying, explaining and
knowing how to deal with these situations - tasks that are so frequent in,
128 Implications for a theory of culture
and so necessary for, everyday life. In each case the mediating structure
allows a recurrent cognitive task to be performed with considerable cogni-
tive efficiency.
It may occur to some readers that, in invoking such structures for
mediating thought, we have strayed quite some way from the connectionist
approach to cognition with which we began in chapter 3. But connectionist
processing alone cannot perform, or at least perform efficiently, all of the
cognitive tasks that humans confront. We subscribe to the view of some
cognitive scientists, so eloquently put by philosopher Andy Clark (1997:
194-200 especially), that brains perform many tasks by first transforming
the problem space to augment performance. It stands to reason that such
augmentations, when successful, would be shared. Clark's examples are
ones mediated by language, among which he would number the calendar
rhyme, which we are likely to externalize by saying it aloud to help us recall
the whole serial list of months and keep track of where we are in it. In
chapter 6, Naomi Quinn is concerned with "augmentations" like that for
reasoning about marriage, that appear to be more wholly internal and less
tied to language.
We know little about how such mediating structures actually evolve.
Talking of our ability to imagine internal representations of syllogisms and
multiplication problems and then solve these mentally, connectionist
modeler David Rumelhart and his colleagues consider the difficulty of
inventing such task solutions:
[W]e think that the idea that we reason with mental models is a powerful one
precisely because it is about this process of imagining an external representation
and operating on that. It is interesting that it is apparently difficult to invent new
external representations for problems we might wish to solve. The invention of a
new representation would seem to involve some basic insight into the nature of the
problem to be solved. It may be that the process of inventing such representations
is the highest human intellectual ability . . . [I]t seems to us that such representa-
tional systems are not very easy to develop. (Rumelhart, Smolensky, McClelland
and Hinton 1986:47)
Solutions for reasoning of this sort, and of the sort Quinn describes in
chapter 6 - in the terms Rumelhart et al. use, imagined representational
systems on which reasoners operate - are invented once and thereafter
culturally shared, rather than having to be invented anew each time they
are needed. Such systems are simply much easier to borrow than to invent.
Further, the "basic insight" that Rumelhart et al. deem necessary to the
invention of new task solutions defies explanation in the connectionist
terms they are developing, as does deliberate conscious thought more
generally, as we have seen in chapter 4.
However such solutions are invented, once in existence as cultural
Three further properties of culture 129
schemas they are spread -just as they are durably learned, endowed with
motivation, transmitted across generations, and imported into new con-
texts - in some of the same ways that we have described for other cultural
understandings. Task solutions of obvious utility like the rule for stopping
at red lights or the rhyme for recalling the number of days in the months
are especially likely to be widely taught, of course. Rumelhart, Smolensky,
McClelland and Hinton (1986:47) note, in the discussion, just cited, of
representations for problem solutions, that "one of the critical aspects of
our school system would seem to be teaching such representational
schemes." On the other hand, knowledge indispensable to everyday task
performance may simply be picked up without formal teaching. For
example, as will also emerge in the next chapter, this is the way we learn the
metaphors for marriage that are vital to communicating about it in every-
day speech on the subject.
The cognitive utility of a task solution can often be sufficient motivation
for learning it. Cultural task solutions may become widespread, however,
due to more than their sheer utility. As will be considered more fully in the
next chapter, much of the broad appeal of the structure on which Paula
and other US Americans rely to reason about their marriages comes from
its incorporation, and evocation, of their understandings, feelings, and
motives connected to that natural American inclination, trying hard to
succeed at whatever you do. Widely appealing, this way of thinking has not
only spread to marriage and many other domains of life to become a
cultural theme, by the process we described in the last section; it has also
spread to many people.
Indeed, this observation regarding the appeal and consequent dissemi-
nation of the schema Paula uses to reason about her marriage extends
beyond the case of such cultural mediating structures for cognitive task
performance, bringing us to another way in which cultural understandings
of all kinds become widely shared. It stands to reason that whatever has
broad appeal or incorporates material of broad appeal will be widely
adopted. We noted, to give one example, that some of the understandings
underlying Paula's and Michael's marital difficulties might represent more
widely shared ones. It is easy to imagine, for instance, that Paula's expecta-
tion that the person who loves you will anticipate all your needs might not
be idiosyncratic to Paula, but might be a more common idea among
women. Assuming this to be so, we might ask what is the broad appeal to
American women of such an understanding. The example, which was
borrowed from a real-life interview with a woman about her marriage,
makes sense within the framework in which Janice Radway
(1984:140-151) analyzes the appeal of popular romances published in the
United States for women readers. These women, she argues, find in the
130 Implications for a theory of culture
romances a working out of their wish to be nurtured by their husbands as a
child is by its mother. Certainly one of the desiderata of the mother-child
relationship is the mother's ability to anticipate and meet the child's
The subsequent failure of a patriarchal and homophobic society
to grant their wish for nurturance, says Radway, leaves this desire unfulfill-
ed in many women (ibid.). Not only the longing for a husband who will be
a mother, but also the practice of reading romance novels becomes wide-
spread, and hence, also, the purchase and circulation of these books that
confirm the validity of such a wish and fulfill it through fantasy
(ibid.: 150151). Earlier in this chapter we described the dynamic by which
themes with widespread appeal recur across different products and dis-
courses, as consumers "elect" these and producers rush to supply what
"sells." Here the closely related point is that products and discourses that,
like romance novels, embody these appealing themes, and hence them-
selves have wide appeal, will be widely adopted and used. In this way many
people, sharing the same needs, come to share the same kinds of things and
practices that fulfill these needs. Of course, not any policy or product has a
chance to be "elected" or "bought,"
we must remember. Just as those in
power can dictate many of people's practices and coerce them into public
expressions of allegiance, so the powerful can control what is available for
people to consume.
We should note that other kinds of products do not so much answer
shared needs, as capture and reflect them, perhaps for the purpose of social
commentary or humor. Good examples are the comic strip and the TV ads
about sit-down dinners that we considered in the context of historical
durability. These public representations of a widely shared concern not
only give social reality to it, as we have already pointed out, but expose it to
a still wider audience of people who might not have been aware of it
previously or truly understood what it was about - for example in the case
of the sit-down-dinner crusade, new immigrants from places where sit-
down dinners have never been heard of; or some husbands and children
who never before understood why their wife or mother was making such a
fuss about it.
Centrifugal tendencies against sharedness
Our account makes sense of the substantial sharing of understandings that
is observable in any society, while recognizing the wide degree of individ-
ual and subgroup variation that is equally obvious empirically. Just as
connectionist modeling, with its emphasis on learning from particular
experiences, is able to explain how understandings of different experiences
sometimes, but not always, incorporate the same cultural themes, it is also
Three further properties of culture 131
able to explain how the experiences of different individuals do not always
result in shared understandings, and how these do not always become
widely shared across individuals. The first and most obvious explanation is
that in any society, different people live comparatively separate and diver-
gent lives. When their resulting experiences diverge, so do the schemas they
build up from these experiences. The larger and more complex a society,
the more opportunities for experiences to diverge and the more radically
they may do so within its borders.
Some such variation in experience comes from the sheer distance be-
tween regions of a large country; just in the way that dialects come to
diverge geographically, so do many other cultural practices (a tendency
countered, in the United States, by that of middle-class Americans like
Paula to migrate in search of college education or job). Much other
variation in experience results from multiple social impediments to the
uniform mixing of people, even those who live in the same region or city -
notably, in our own society, people of different racial and ethnic identity,
religious affiliation, income, and occupation. Divergent cultural practices
may simply grow up in different groups or in different settings due to
noninteraction across these groups or settings, just as regional differences
do. Zoning codes, lack of local jobs, and lack of low-cost housing or mass
transit keep lower-income residents out of some neighborhoods. Among
other sources of segregation are the division of the educational system into
public and private schools, the growth and promotion of a youth culture,
and plain social unease between people from different groups. Any and all
of these circumstances can create barriers to interaction as effective as the
distance between New York and California. Thus, as Paula remembers it,
the exclusive neighborhood in which she grew up and the curriculum
tracking system at her school kept her from mixing with children from
poor families long enough to make friends with them. The segregation of
groups is heightened considerably by group identities that stress the in-
commensurability of one's own group's practices and understandings with
those of other groups. Among those separated in this way by their identi-
ties are some social classes, some racial, ethnic, and religious groups, and
some groups of men and groups of women in the United States. Acting on
such segregationist beliefs - for example, that intermarriage with someone
from a different group would not work, or that they could never under-
stand us - increases the likelihood that the separate experiences of "them"
and "us" will continue to diverge. This effect is enhanced still further when
coupled with beliefs in the superiority of one's own group over others, with
negative stereotypes of members of other groups,
and with explicitly
exclusionary practices.
Even people whose lives are necessarily intertwined in important con-
132 Implications for a theory of culture
texts (employees working in the same company, students attending the
same school, or men and women in the same family) may be segregated in
other contexts. They may then bring such different, and differentiating,
experiences from these other contexts that they learn to be and to under-
stand themselves as quite distinct, fostering a lack of interaction and
consequent lack of shared experience even when they come together. On
occasion, Paula has wanted to get to know a colleague at work better, but
has been at a loss to know how to overcome the social reserve between
herself and the other person, or how to interpret whether that person was
interested in knowing her, when the other was from a different class or
ethnic "background" (as she thought of it) from herself.
Less obviously, people may grow up and live in circumstances similar to
others, in proximity to these others, or even interacting with them, but still
have different experiences. Both Paula and Michael learned the virtues of
self-reliance as children, but there are differences in the specifics of the
schemas they acquired. Michael, like many men of his generation, did not
learn to associate self-reliance with washing or mending his own clothes, as
Paula did; Paula, like a lot of other women her age, did not learn to
associate self-reliance with doing her own home repairs, as Michael did.
Likewise, both Paula and her childhood friend, Yvonne, learned an associ-
ation between mothers and kitchens as they were growing up, but in other
respects their experiences were different. Unlike Paula's mother, Yvonne's
mother enjoyed baking and cooking, and her self-confidence and pleasure
as she did that became part of Yvonne's network of associations with
mother and kitchen. So - while Paula and Michael share a general under-
standing about the importance of self-reliance, and Paula and Yvonne
share a general expectation that mothers will be found in kitchens, there
are many specific respects in which their schemas diverge. The single-
family houses so common in the United States and our ideas about
nuclear-family autonomy and privacy make it possible for many such
differences to exist without people being aware of them. In any society,
too, shared ways of talking - labels for things, formulaic expressions - can
cover up subtle differences in practice, as can the discreet silence or the
token assent that often comes of wanting to conform or not wanting to
cause argument.
Among the differences that can exist between people, it should be clear
from these last examples that there can be differences in the feelings and
motivations that attach to the schemas they hold. Thus Michael gets a
genuine feeling of pride when Paula asks him to attend to a home repair,
and a rewarding sense of self-sufficiency when he completes it. Paula, on
the other hand, is displeased or even panicked when she finds it necessary
to undertake such a repair herself. Likewise, Paula's memory of her
Three further properties of culture 133
mother "trapped" in the kitchen is tinged with annoyance at her mother's
acquiescence to this arrangement and it makes her determined not to fall
into a traditional division of household labor herself; and Yvonne's mem-
ory of her mother in the kitchen is suffused with admiration for her mother
and makes her want to be an equally good cook.
Another, more subtle, source of individual variation comes from varia-
bility in the salience of information. Spiro (1987a: 163-164) some time ago
pointed to the differential "cognitive salience" a cultural doctrine might
have, characterizing this as a difference between cultural belief 'and cultural
We think we can explain, in general terms quite consonant with
the 'surface' and 'deep' meanings that Spiro identifies as motivating relig-
ious belief, why this difference would arise. People's schemas do not merely
copy the most frequently presented associations in their environment (see
Strauss and Quinn 1991 for examples). Due to the neurochemical process
we described earlier, the learner's emotions and consequent motivations
can affect how strongly the features of those events become associated in
memory. It is possible, of course, for different individuals to surround a
given experience with similarly strong emotions and motivations, and
hence be equally likely to develop strong and readily activated schemas for
interpreting that type of experience. It is just as possible, however, for
emotions and motivations to lead to nonsharing, because two individuals
could be exposed to the same information but care differently about that
information and so not internalize it in the same way. For example, Paula's
brother, Daniel, had the same opportunities as his sister to observe their
mother while they were growing up. Since the happiness of mothers in
kitchens was not of especial interest to him, however, the mental associ-
ations he formed of her in this context were not as strong as Paula's and
have since been overlaid with other memories of their mother, more
emotionally salient for him.
Thus the same processes we have described, by which Paula was shaped
as a cultural and historical personality, have also shaped her in all of her
unique Paula-ness.
We know that children who grow up in the same
household do not emerge with identical schemas. On the other hand, a
great deal of formative experience is shared in this environment. Consider
the opposite case of the person who immigrates to the United States as a
young adult. No matter how "assimilated" such an immigrant becomes
into US society, he or she will always be different with regard to inter-
pretive sensibilities, emotional tonalities and motivational pulls that were
the product of a childhood quite different from that typical of someone
growing up in the United States (although, as LeVine 1982:21 points out,
and as we might expect, the children of immigrants are much more
thoroughly Americanized). The indelibility of these differences may be
134 Implications for a theory of culture
likened to that of the accent that lingers in the immigrant's speech, how-
ever well he or she learns the language of the adopted country. Unlike a
foreign accent, however, differences in early cultural experience may leave
traces so subtle and inexpressible as to remain largely hidden from other
Part III
Practice and possibilities
Paula's life supplies convenient illustrations of our points, but empirically
minded readers may be impatient with fiction and eager for us to tie our
approach to real applications. While we could not possibly give full
accounts of our own research within the confines of this book, much of
that work is published elsewhere and will either be excerpted or cited as
appropriate in what follows. Naomi Quinn's current research revolves
around US Americans* shared understandings of marriage, while that of
Claudia Strauss centers on the beliefs of diverse US Americans about
welfare and success. Both bodies of research lean heavily on the close
analysis of discourse from interviews, though the interviews themselves
and the methods by which the resulting discourse was analyzed were
designed for somewhat different ends. By and large, Naomi Quinn's re-
search is designed to uncover centripetal properties of culture, particularly
the sharedness and motivating force of understandings; Claudia Strauss'
research is designed to reveal centrifugal properties of culture, particularly,
different ways people can internalize the cultural understandings they
learn, and the effect of these differences for cognitive consistency, moti-
vation, and conformity to or divergence from dominant ideologies. Thus,
while the two bodies of research in no sense confute each other (Quinn
1988), each lends itself to somewhat different observations. In addition, the
two of us have collaborated on a small pilot project investigating what
children in our society understand about marriage (work that we hope
soon to extend to other domains and larger samples of child interviewees),
and that study (Strauss and Quinn 1992) will provide a few of our
The chapters in this part carry a general import. The research they
report shows why we need to rethink culture instead of throwing out the
concept, as (we noted at the beginning of part I) some anthropologists are
advising us all to do. Our examples have been chosen to make two further
points bearing on our central argument about the need to study the
intrapersonal realm of culture. First, we will show through these examples
136 Practice and possibilities
that studies of the intrapersonal side of culture can yield results that are
nonobvious - indeed, even surprising - when set against studies of the
extrapersonal side of culture. These research results have to do with
unexpected differences between public culture and the shared understand-
ings held by ordinary people; the relatively greater or lesser sharedness of
people's cultural understandings than might be expected from the pattern
of dissemination in public culture; the relatively greater or lesser moti-
vation cultural understandings have for people than might be guessed
from public enactments and material artifacts; and, finally, the relatively
greater or lesser degree of integration of cultural understandings in
people's minds than might be supposed from the cohesiveness or fragmen-
tariness of public culture. Secondly, we will show that research on the
intrapersonal side of culture allows us to pursue questions about the
nature of culture that cannot be answered, and indeed often are not asked,
by research on the extrapersonal side of culture (which is suited to asking
and answering other questions). The questions we are thinking of have to
do with the way cultural understandings are learned; the way motivational
force is or is not imparted to these understandings; and the way diverse or
conflicting cultural messages are cognitively structured.
We decided to include these chapters, not only for the empirically
minded reader, but because we ourselves are empirically minded. At the
same time, we are acutely conscious of the disparity between the sweep of
our theory and the relative modesty of our research agendas. If some
readers welcome the change of pace from theory to application, others
may be frustrated by the inevitable shift in scope from theoretical ques-
tions that vex a whole field to research questions that can reasonably be
addressed by individuals in their lifetimes. We invite anyone who is made
impatient by detailed investigations and findings to skip to chapter 9.
Research on shared task solutions
Naomi Quinn
Cultural sharing has been the focus of my work over the past fifteen
years. When I began this work, I was agnostic about the existence of such
sharing. It is true that in my more recent collaboration with Claudia
Strauss, leading up to this book, it is typically I who remind her to
consider evidence for and examples of sharing and the other "centripetal"
properties of culture, while she has always reminded me to include instan-
ces of nonshanng and the other "centrifugal" properties of culture in our
discussion. She and I jokingly interpret this predilection as a generational
one: in my training in the sixties I was primed to think about how culture
was shared; trained some fifteen years later, she was primed to think
about how culture was not. But by the time I began the research program
to be described in these two chapters, my own early researches and
perhaps the beginnings of a new climate in cultural anthropology had
made me pessimistic about how shared culture might really be. I am
certain of this because of one incident in my career that I have never
forgotten. When I came to Duke, in the spring of 1972, to be interviewed
for the faculty position I hold today, one of my interviewers was the late
Weston LaBarre, the distinguished culture-and-personality theorist then
on the faculty. Since I thought of myself and presented myself as a
psychological anthropologist, a more contemporary label that was clearly
differentiated in my mind from that of personality and culture, I was
dismayed when he prefaced our interview by saying, "I see by your vita
that you work in the area of personality and culture." Ever foolhardy, I
reacted instantly, "Oh, no, that can't be. Because I don't believe in per-
sonality," and added, "And I'm not sure I believe in culture." While this
story illustrates that I might have been skeptical about finding it (or
maybe, says my co-author, it just exhibits my contrary streak), I cared
enough about the question of cultural sharing to spend the next many
years pursuing it. That fact, certainly, reflects my membership in an
earlier generation of cultural anthropologists.
I began the work reported in this chapter and the next by collecting and
138 Practice and possibilities
examining extensive interviews from a small but varied group of US
Americans on an everyday topic about which they had much to say.
(Because of the awkwardness of repeating the whole phrase "US Ameri-
cans" each of the many times I refer to them in this chapter and the next, I
often take the liberty of dropping the "US" and simply talk about Ameri-
cans and American cultural understandings. Readers should remember
that I am referring to the United States.*) The topic was marriage. Results
of this research on Americans' understandings about marriage appear in a
series of published articles (Quinn 1982,1987,1991,1992, and 1996). I will
begin by briefly describing the methods I used to select interviewees and
conduct interviews with them; several analyses to which the transcribed
interviews have been submitted so far are described in the pages to follow.
Student research assistants and I conducted interviews in 1979 and 1980
with twenty-two individuals, the husband and wife in each of eleven
marriages. Beyond some commonalities of cultural and marital experi-
interviewees were selected to maximize diversity with regard to such
obvious differences as their places of geographic origin, religious affili-
ations, ethnic and racial identities, their occupations and educational
backgrounds, their neighborhoods and social networks, and the lengths of
their marriages. Husband and wife in each marriage were interviewed
separately, for an average of between fifteen and sixteen hours.
"intensive interviews" (for lack of a better term in social science parlance)
were structured as closely as possible after ordinary conversations that one
might have about one's marriage with a friend or relative, one's spouse, or
Undoubtedly the resulting interviews are exactly like none of
those, but they are enough related, and the task of being interviewed was
natural enough in and of itself, so that what resulted was extensive natural-
seeming discourse
- indeed, an extraordinarily rich body of it.
Initially focussed on the cautious question of whether and to what extent
Americans might share a cultural model of marriage, this research revealed
an unexpectedly large degree of shared understanding on this topic. The
obvious next question was, how did this sharing arise? Explaining the
circumstances under which sharing arises should also illuminate the ob-
verse circumstances under which understandings do not become widely
shared and individual difference has more play. As will become apparent in
the course of this chapter and the next, this concern with the sources and
limits of sharing touches upon still other issues addressed in the last part of
this book - the circumstances under which shared understandings gain
motivational force and durability for individuals, and draw on preexisting
thematic understandings. Another issue raised in previous chapters, that
will emerge again in this one, is a methodological one: the need to study
cultural sharing at the level of individual understandings in order to
Research on shared task solutions 139
describe what is shared (and what is not) and to find clues to how this
sharing happens, when it happens.
Following the course of a multilayered analysis of the corpus of dis-
course that I collected, I will describe three different kinds of sharing that,
taken together, produce substantial commonalities - though, as we will
see, certainly nothing like uniformity - in the way most contemporary US
Americans understand marriage. The first kind is the sharing of cultural
exemplars that provide speakers with the abundant metaphors in which
they cast marriage in their talk about it, as they perform the routine
linguistic task of clarifying, for their listeners, the points they want to make
about it. Being a matter of sheer efficiency of communication, this kind of
sharing is motivated largely, if not wholly, by its utility for performance of
this clarification task. The second is another kind of shared task solution in
the form of an internalized template people use to reason about marriage -
indeed, without which they could not readily so reason. Incorporating, as
it does, a cultural schema for successful realization of difficult enterprises
in general that is itself widely appealing to Americans, the schema for
reasoning about marriage is motivating not only by dint of its intrinsic
utility for this reasoning task but also (as was already briefly noted in a
discussion of Paula's reasoning about her marital incompatibilities in
chapter 5) by the force of the more general schema on which it relies. This
chapter will be devoted to a consideration of these two different kinds of
shared task solution. The third kind of sharing, taken up in chapter 7, is a
shared constellation of powerful hopes and expectations about marriage.
While my research convinced me - and in these chapters I hope to
convince readers - that culture is indeed substantially shared, it also
changed my idea of what it is that is shared - about the nature of culture
itself. In what was perhaps an all too monolithic view of culture, I long
spoke and thought of what I was uncovering in my analysis as the cultural
model (set especially, Quinn 1987 and Quinn and Holland 1987) of Ameri-
can marriage - as if it were the one and only such model. I now think of it
as a cultural model Xhai has arisen from specific experiences US Americans
have had in common - although it is certainly one that, due to these
common experiences, most of these Americans share. Even further, I have
come to see the shared understanding implied in the term "cultural model"
as a product of variable tendencies toward different degrees of sharedness,
differentially endowed with motivating force. My imagining of the shared-
ness, the motivational force, and the other properties of this cultural model
I owe to developing theory from cognitive science about schemas, and a
recasting of these shared schemas in connectionist terms.
I should make clear at the outset what I mean by a cultural model,
expanding on our brief definition, at the beginning of chapter 3, of cultural
140 Practice and possibilities
models as complex cultural schemas. Cultural schemas may organize
domains of experience of all kinds, perceptual or purely conceptual, from
simple concepts of single objects or events to elaborate knowledge systems
(Langacker 1986:4). Schemas like the one to be described in this chapter
and the next are conceptually complex; they connect and organize an
interrelated set of elements and hence not only delineate but serve as
working models for entire domains of activity in the world, and for this
reason have been called "cultural models" (D'Andrade 1995:151-152;
Holland and Quinn 1987), a terminology I will interchange freely with
"cultural schema," as called for by the context.
Rethinking cultural models in a connectionist framework not only
makes clear the experiential basis of these, but also makes it possible to get
closer to the processes by which such complex shared understandings are
built up from shared experience. Only through studying these processes of
internalization can we discover what kinds of shared experience are se-
lected to become shared understanding. Although my research on Ameri-
can marriage certainly does not encompass all the kinds of shared experi-
ence from which shared understanding might arise, it does point to two
major, if disparate, kinds: the shared experience of having to perform a
recurrent cognitive task (the topic of this chapter), and shared experience
of the emotionally intense kind exemplified par excellence by that under-
gone in infancy (the topic of the next chapter). My future research will
continue to explore the different kinds of shared experience leading to
shared cultural models. This research question simply does not arise within
a paradigm confined to study of the externalities of experience - of public
culture alone.
In most of my research so far, as exemplified in this chapter and the next,
the strategy has been to exploit clues in ordinary discourse for what they
tell about shared cognition - to glean what people must have in mind in
order to say the things they do. This strategy depends at once on extensive
analysis of patterns in certain linguistic usages that recur in discourse, and
close analysis of the details of this use. Metaphor use, reasoning, and the
use of key words, as exemplified in these two chapters, have proven to be
especially fruitful linguistic features of discourse because they occur rela-
tively frequently in ordinary talk, and because each, for a different reason
that will become apparent, bears a heavy load of cultural knowledge.
Analysis 1. Metaphors for marriage and what they do
An analysis of the metaphors people used in their talk about marriage,
prompted by the influential and provocative work of George Lakoff and
Research on shared task solutions 141
Mark Johnson (1980), provided me with some of my first clues to Ameri-
cans* understandings of marriage. This examination of metaphors for
marriage also led me, ultimately, to rethink the role that linguists have
granted to culturally shared metaphor (Quinn 1991, n.d.). I have arrived at
a wholly different interpretation from that of Lakoffand his colleagues, for
the phenomenon to which they must be credited with drawing our atten-
tion - that is, the ubiquity of metaphors in ordinary human speech. My
explanation rests on the repeated necessity of clarifying the point we are
trying to make to others as we convey our ideas to them in speech. As will
emerge later in this part of the chapter, speakers' deliberate use of meta-
phor to clarify their points is not the only source of metaphor in ordinary
speech. It is, however, a very productive source, and the metaphors so
chosen are not only common but, just because they have a point to make,
especially noticeable. The metaphors speakers use to talk about marriage
are of this type.
Here I will present evidence for my view that speakers employ these and
other metaphors to clarify their points. Using a connectionist framework, I
will then indicate how metaphors do this clarification work and describe
how cultural knowledge of apt metaphors for doing so becomes shared.
Finally, I will offer a critique of the prevailing theoretical and method-
ological approach, derived from linguistics, to the study of metaphor. This
is an approach, I will argue, that understates the variability in the use of
metaphors while overstating their role in constructing understanding. My
critique parallels our earlier one, in this book, of the pursuit of cultural
meaning through the exclusive study of public culture. By taking culture
seriously as an integral part of cognition, and applying the theory of
culture developed in this book to my findings regarding metaphors for
marriage, I arrive at a fresh account of a phenomenon - metaphor use -
that has been the topic of long investigation and continued controversy in
several cognitive sciences.
A great advantage of my view of metaphor is that it suggests an answer
to the question of how metaphors like those Americans use for marriage
come to be so widely shared. It will emerge that metaphors like these derive
their sharedness from two sources. First, they draw upon cultural exemp-
lars of those aspects of experience that the speaker wishes to clarify for the
listener. Secondly, these metaphors fall into classes reflecting a shared set
of underlying concepts that the metaphors have been chosen to represent
and highlight. Why this set of concepts should, in turn, be shared will be
addressed in the second half of this chapter. Here I will direct my attention
to the metaphors themselves, presenting evidence, first, of the classes into
which metaphors fall and the underlying concepts that these classes reflect.
142 Practice and possibilities
Shared classes of metaphor
Strikingly, the hundreds of metaphors for marriage found in my corpus of
discourse fall into just eight classes. These are: (1) metaphors oflastingness,
such as, '"It was stuck together pretty good" or "It's that feeling of
confidence about each other that's going to keep us going"; (2) metaphors
of sharedness, such as, "I felt like a marriage was just a partnership" or
"We're together in this"; (3) metaphors of mutual benefit, such as, "That
was really something that we got out of marriage" or "Our marriage is a
very good thing for both of us"; (4) metaphors of compatibility
such as,
"The best thing about Bill is that he fits me so well" or "Both of our
weaknesses were such that the other person could fill in"; (5) metaphors of
difficulty, such as, "That was one of the hard barriers to get over" or "The
first year we were married was really a trial"; (6) metaphors of effort, such
as, "She works harder at our marriage than I do" or "We had to fight our
way back almost to the beginning"; (7) metaphors of success or failure,
such as, "We knew that it was working" or, conversely, "The marriage was
doomed"; and (8) metaphors of risk, such as, "There're so many odds
against marriage" or "The marriage was in trouble."
This discovery, of the recurrent use of a small number of metaphor
classes to describe a domain of American life, across discourse on this topic
by a diverse group of Americans, was my first evidence of widely shared
understandings about marriage. This regularity had to be discerned be-
neath a great deal of diversity in the way in which couples and individuals
within these dyads represented, spoke about, and thought of their mar-
riages. Some interviewees obviously relished their metaphors and pro-
duced novel and elaborate ones, while others were minimalists in this
regard. Most liked to talk about their marriages; a few did not, or shied
away from disclosure on certain marital topics. Different interviewees
made something a little different of the interview itself, one couple seeing it
as an opportunity to present their visionary model of marriage to the
world, another treating it as a chance to explore their own role as pioneers
in what they saw as a different kind of marriage, another to reaffirm what
they found good about their marriage, one man to relate the history of his,
and other individuals to therapize about their marriages or tell their side of
things to a neutral person. Different couples and individuals dwelt on
different aspects of their marriages, providing more discussion of and
metaphors for some concepts than for others. One wife, for example, spoke
long and eloquently, in a variety of extended metaphors, about the difficul-
ties that had been overcome in her marriage. Some - for example, the man
who maintained a running metaphor for the benefits of marriage as liquid
in a container, the level of which could rise or fall and be measured by a
Research on shared task solutions 143
mechanical gauge - returned repeatedly to one or another pet metaphor
(to use a metaphor). They had different theories about marriage, and
different marital issues, and different styles of married life, of which they
were also in different stages. Some were in happier, others in distinctly less
happy marriages at the time they were interviewed, and some subsequently
divorced. For all of these reasons, and others, interviewees had diverse
stories to tell about their marriages. The common set of concepts -
lastingness, sharedness, mutual benefit, and so forth - that all these stories
embellish is therefore all the more arresting.
Having given just two examples to illustrate each class of metaphor, I
will reinforce my point about the variety of these metaphors for marriage
within the classes to which they belong with a single more extended
example - that of marital lastingness. Speakers frequently cast its lasting-
ness in the metaphor of marriage as a manufactured product: one that is
well-made in the senses of being not only well put together - "stuck
together pretty good," in my earlier example - but also structurally sound
- for instance, "They had a basic solid foundation in their marriages that
could be shaped into something good"; one that is made out of strong
materials - e.g., "We forged a lifetime proposition"; and good parts - e.g.,
"We have looked into the other person and found their best parts and used
those parts to make the relationship gel"; and one that is made with the
necessary care - e.g., "When the marriage was strong, it was very strong
because it was made as we went along; it was sort of a do-it-yourself
project"; and effort - e.g., "Maybe that had something to do with what
was good about it. The fact that we really had to work at it." However, this
is far from the only metaphor that these speakers used to convey the
expectation of marital lastingness. Another very common metaphor for a
lasting marriage is that of an ongoing journey that married people under-
take together - "That's going to keep us going," in the other earlier
example of marital lastingness. Somewhat less frequently but no less
predictably, speakers use other metaphors of marriage as two inseparable
objects - "We knew we were going to stay together"; as a durable attach-
ment between spouses - "That just kind of cements the bond"; as a
permanent location - "I was able to stay in the marriage"; as a secure
possession - "We got it"; and as an indestructible natural object - "the
everlasting Gibraltar nature of the whole thing" - among other metaphors
- to convey the expectation of its lastingness.
At the same time that the small number of metaphor classes suggests
widespread sharing, the finding of such variability among the metaphors
within these classes suggests that the metaphors themselves cannot be the
basis of this shared understanding. This conclusion runs counter to argu-
ments made by Lakoff and his colleagues (Johnson 1987; Lakoff and
144 Practice and possibilities
Johnson 1980; Lakoff and Kovecses 1987) to which I have already alluded.
These researchers imagine that the metaphors they encounter in spoken
and written language reflect deeper conceptual metaphors that "allow'* or
"constitute" conceptualization, comprehension, and understanding.
However, the many different metaphors that speakers readily call upon to
recast the same underlying concept is evidence against a constitutive role
for metaphor. Instead, as I have elsewhere argued (Quinn 1991), in their
pattern of usage these metaphors for marriage reflect an underlying
schema that people share for thinking about marriage, and that guides
their selection of metaphors for it. This schema will be the focus of the
second half of this chapter, on reasoning about marriage.
In assuming that what is in mind is so directly manifest in language,
students of metaphor like Lakoff are reminiscent of anthropologists such
as Geertz, discussed in chapter 2 (on this point about Geertz in particular,
see Wikan 1987) and literary critics such as Fredric Jameson, to be dis-
cussed by Claudia Strauss in chapter 8, who assume that what is in people's
minds can be read directly from public culture. Perhaps it should not
surprise us that a theory placing such inferential weight on linguistic forms
should have originated with linguists. And perhaps it is only because
anthropologists lack any more developed theory of cultural meaning of
their own, that they have so quickly and uncritically incorporated into
their own formulations Lakoff s assumption that metaphor "structures"
or provides the "model for" the understanding that it captures.
language holds the clues to the cultural schema I will describe, the schema
is far from isomorphic with the language, or obvious from it; certainly, it is
not retrievable from any given metaphor speakers use. As I will return to at
the end of my analysis of metaphor, reconstruction of such cultural
schemas from the indirect clues provided by language requires a suitably
sensitive method.
Cultural exemplars as sources of metaphors
This conclusion, that our knowledge of the domains about which people
are speaking guides their selection of metaphors to describe these domains
and reason about them, points to an explanation for what metaphors are
used to do. Metaphors, I have already suggested, clarify. One important
reason they are used, and used so very frequently in everyday speech, is
that they do a good job of clarifying for listeners the nonmetaphorical
points speakers are trying to get across. And for this same reason - along
with their high rate of use for this purpose in ordinary discourse - they are
excellent clues to the cultural schemas that underlie them.
This brings us to the second source of sharing that gives metaphor use its
Research on shared task solutions 145
regularity. Not only are metaphors selected to exemplify shared concepts;
as I will next explain, they are drawn from shared exemplars of these
concepts. In this way they introduce an outstanding and unambiguous
instance of the point being made. The degree to which these cultural
exemplars are shared by speakers is no less striking than the sharing of the
underlying concepts that guide metaphor selection and account for the
classes of metaphor speakers use. Indeed, metaphors clarify precisely
because they are drawn from domains that are widely acknowledged
exemplars, with regard to whatever aspects of experience they are being
made to stand for. Because speaker and listener intersubjectively share an
exemplar, both knowing what it exemplifies, the chosen metaphor does the
job of clarifying the point for the listener. Members of a speech community
must share these exemplars in order to be able to use them to communicate
in this way.
To see how metaphors drawn from exemplars work, consider an illustra-
tion I culled from the sports page of USA Today (May 5, 1993). Third
baseman George Brett, interviewed on the occasion of his retirement from
baseball, comments on his unusually long-term relationship with the Kan-
sas City Royals: "I compare it to a marriage. We've had our problems, but
overall, we have had a good relationship. I never, ever want to put on
another uniform." Marriage is famous among Americans for being some-
thing that is meant to endure and that does so (when it does so) in spite of
its difficulties. That is why George Brett's metaphor gives us all a surer
sense of what he wants to convey about his relationship with the Royals.
As with Brett's choice of the marriage metaphor, speakers' selection of a
metaphorical source domain often seems motivated precisely by the point
they have in mind. The case of the husband who speaks of his marriage in
terms of "the everlasting Gibraltar nature of the whole thing," is illumina-
ting in this respect. The context for this metaphor is his remembering the
moment in his marriage when he realized "that my confidence in the
everlasting Gibraltar nature of the whole thing was rather naive" [5H-4].
To underscore how, in his naivete, he overestimated the lastingness of
marriage, this speaker employs a metaphor of something truly lasting -
indeed, an icon of the everlasting for many US Americans, appropriated,
reproduced, and widely disseminated as the logo of a national insurance
company. A parallel example comes from another interviewee: expressing
his surprise about a marriage of acquaintances that "suddenly broke up,"
he says, "We had no warning; if ever a marriage was nailed in cement, that
was the one" [8H-2]. Both these men draw metaphorically on things
known for being exceptionally lasting to emphasize how unquestioning,
indeed unrealistically so, were their initial expectations about the lasting-
ness of given marriages. On the other hand, for the points about actual
146 Practice and possibilities
marriages that speakers ordinarily want to make, the Rock of Gibraltar
and cement may be entirely too durable. Handier metaphorical source
domains are manufactured products that, while they might be made to
last, can indeed "break," and that are known to break as frequently as they
last; other kinds of things that, even though of a type that usually lasts, can
conceivably be flawed or - like the marriage described by one interviewee
that "just blew apart like someone put dynamite under it" [6W-6] -
destroyed by external forces; possessions that can be "up for grabs," or
"slide right down the tubes," as well as being securely held or safely kept;
or travels that sometimes end in mid-journey, rather than reaching their
planned destinations. All these are supposed to last, even though people
recognize that they do not always do so: Americans believe that manufac-
tured things should be well made; possessions are often treated as inalien-
able in their society; like people everywhere, they expect hard materials to
resist erosion and rough treatment; and they set out on their travels fully
intending to reach their destinations - even though these outcomes, like a
lasting marriage, do not always eventuate. Manufactured products, dur-
able things, possessions and journeys are drawn upon over and over again
as sources of metaphors for lasting marriages, not because they constitute
our understanding of marriage, but simply because they are the major
cultural exemplars, in our world, of things that typically last.
By contrast, there are metaphors that seem distinctly wrong when we try
them out as characterizations of marriage. Here is one case. At a Catholic
wedding ceremony described to me, the officiating priest counseled the
couple about to be wed that they should think of marriage as an ice cream
cone. You can eat it up fast, the priest explained, or you can lick it slowly
and make it last a long time. The wedding guest who told me this story
reported that members of the wedding party squirmed during that part of
the ceremony, and complained vociferously among themselves afterwards
about the priest's disconcerting comparison. This metaphor has an obvi-
ous sexual connotation that might explain some of the guests* discomfort.
But I think any reader would also agree, the metaphor simply does not fit.
It is experienced as inappropriate, I would argue, because food is nondur-
able, and ice cream is a particularly perishable food: no matter how slowly
you eat an ice cream cone, you can only make it last so long. Among all the
hundreds of metaphors for marriage that occurred in the discourse of my
interviewees, not one likened it to any kind of food.
Sometimes speakers wish a metaphor to exemplify, more than a feature
of marriage in isolation, a relation between its features - as Brett uses the
analogy to marriage to capture the relation between the lastingness, the
difficulty, and the satisfaction of his association with the Royals. Some
metaphors are favored, and hence recur with frequency, because they do
Research on shared task solutions 147
such a good job of capturing these complexities of experience. Thus one
reason well-made products are such popular metaphors for marriage is
because they are intended not only to last but to work (and can not only
break, as I have noted, but can also break down). The prototypical
manufactured product is one made to perform some function. When
Americans speak about a machine or other manufactured product that
"works'* they mean that it does what it was made to do, performs this job
successfully, and doesn't break down in the process. Indeed, a fixture on
the American scene is Consumer Reports, a nationally distributed, widely
subscribed magazine devoted to testing and rating products for their
performance and durability. Marriages should also last and "work" in the
sense of succeeding. Thus, one woman describes herself and her husband at
the wedding of friends, looking at each other and shaking their heads,
saying," It'll never work. It'll never work'" [1W-5]. (They turned out to be
right.) Conversely, a man says, "[W]e wanted the marriage to keep work-
ing." [3H-1]. In a much more elaborate and much more specific metaphor
of something that performs its function well because it was made well,
another husband explains "[t]he self-righting concept that the marriage
has enough soundness and equilibrium that it will take steps to right itself
in any kind of stormy situation" [7H-6].
For Americans, manufactured products stand as the prime exemplars of
things that not only last and work, but are built to do so in part because of
the effort put into them. Chief among the ingredients that go into a
well-made, properly-working product, most people understand that effort
is also central to a lasting, successful marriage. Thus, the man who says,
"[W]e wanted the marriage to keep working," continues, "so we made the
effort to communicate and talk things out, to change our routines." This
effort may be that of manufacturing something durable, like a seaworthy
vessel. It may also be conceptualized as the effort of keeping something
well maintained, as another interviewee conveys, with the example of a
different kind of manufactured product: "If you get a new house and just
let it sit for year after year without doing anything to it, it's going to
deteriorate and I feel that a marriage is something that you have to
continue to work at it" [11W-16]. Favored, too, is a metaphor in which
effort is the work that goes into making something function as it should,
because it captures the relation between the effort put into marriage and
marital success - and perhaps also because it conveys the kind of effort of
adjustment that, as will be seen in the next section, ordinarily goes into
achieving marital compatibility. In the words of still another interviewee,
"This is more permanent, this is something that you should make work,
you know. And not anything that you should give up on. You know, you
should always keep trying no matter what" [2W-1].
148 Practice and possibilities
Journeys, as already noted, are another popular metaphor for marriage.
They can be of long duration, continuing until arriving at some desired
destination, "a place" that stands for a happy marriage. They typically
also involve encounters with obstacles - like the "stormy weather" of the
metaphor already cited, that the ship has been built to ride out on its
voyage. These obstacles often require effort to surmount. Thus, as a
metaphor for marriage, the journey can be made to express the need to
overcome marital difficulties with effort, to attain benefit, and to last, as in
"It was worthwhile enough to struggle through those periods and move
on" [5W-1], or, "However long and stony a road it was we had agreed to
set out on it and meet each small situation as it came" [4H-7]. Journeys are
often also undertaken with travelling companions. Thus they work as a
metaphor for the relation between sharedness and lastingness, as in such
comments as this one, in which the companions go their separate ways in
mid-journey: "If it gets static in our relationship then that's when we split I
guess, or start going a different direction" [3 W-5]. Of course, the same
points can readily be made in different metaphors, or nonmetaphorically,
as speakers elsewhere illustrate. The journey is a favored one, I argue,
because it is a cultural exemplar of a protracted activity having an ultimate
objective, predictably beset with difficulties that require effort to over-
come, and that can be undertaken with another person - a cluster of
features that also characterize marriage.
How metaphors become shared
Connectionism offers a straightforward associationist account of how
given metaphors come so readily to mind that we can produce the ones we
need without hesitation in the course of talk. Whether familiar or unfamil-
iar, richly or sparsely detailed, a metaphorical source domain that has
come to exemplify a given feature of the world - like the Rock of Gibraltar,
well-made products and secure possession have come to exemplify lasting-
ness for many Americans - will have become strongly linked to that aspect
of a person's experience. In future encounters with that feature of the
world - say, in discussions of the lastingness of marriage - rocks, manufac-
tured products, and possessions, among other exemplars, will be activated
in that person's mind. Likewise, talk of marital sharedness will call to most
people's minds partnerships, travelling companions, things wrapped up
together in packages, and love birds, among other exemplars. (Notice that
whether the exemplar that comes to mind is specific or general depends on
the strength of the linkages that someone has made. It is conceivable, for
instance, that some individuals have come to associate durability more
strongly to rock in general than to the Rock of Gibraltar, or to BMWs or
Research on shared task solutions 149
other products famous for solid construction, more strongly than to
well-made products in general.) In just the same way, efforts to clarify
more complex points about the relations among elements of a cultural
model will bring to mind the narrower set of just those exemplars of such
relations. Thus the possible future dissolution of her marriage, we saw,
brought to one speaker's mind the idea of travelling companions who
come to a parting of the ways. The activation of such cultural exemplars
insures that an appropriate metaphor will almost always be identified and
selected with the speed of speech.
As they gain popularity, given metaphors for marriage become more
and more strongly linked through repeated encounter and use, not only to
the particular aspects of marriage they exemplify, but to marriage in
general. Too, we might expect metaphors like those of the journey, and the
manufactured product, that capture multiple features of marriage and
complex relations among them, to appeal to speakers and come into
widespread use in talking about marriage simply because they have be-
come so strongly associated with the marital experience. Once in currency,
moreover, these metaphors often continue to be used just because they
come to mind so readily in any context having to do with marriage.
Speakers thus tend to favor such metaphors as "rocky" for a marriage in
the throes of difficulties, or a long, successful marriage as one "made to
last," even when leaving unexploited other of the features that made the
journey and manufactured product metaphors apt ones in the first place.
In this way metaphors become conventional. Of course, many of the
metaphors that can be and are used to talk about marriage are also applied
more broadly to numerous contexts - think, for example, of "It works for
me," or, "It was rough going," both in wide use.
I have used the connectionist framework put forward in chapter 3 to
describe how cultural exemplars of various aspects of experience are drawn
upon in the deliberate production of metaphors in speech. It remains to
suggest how these shared exemplars might be learned in the first place. Over
the course of experience each of us acquires a sizable bank of culturally
agreed-upon exemplars, each exemplary for given features of the world.
Some of these exemplars are culturally distinctive; others, to the degree that
they occur widely in nature or bodily or psychic experience, recur cross-
We learn them partly from first- and second-hand experience
with the nonlinguistic world - our own experiences with, and stories we
have heard about, difficult activities, things that don't work as they should,
durable things, and so forth. Significantly, also, we learn these exemplars
from the linguistic experience of hearing (and using) metaphors in talk.
This linguistic experience serves to expunge from our bank of exemplars,
candidates based on personal experiences that are too idiosyncratic. The
150 Practice and possibilities
requirements of spoken communication foster such selectivity, forcing us
to use just those metaphors that have cultural currency, so that their
meaning- what they stand for - will be intersubjectively shared with those
to whom we talk, and hence useful in clarifying our point. Thus it is that
linguistic exchange plays an especially large role in the learning of exemp-
lars - as it does in the learning of so much else - greatly enhancing the effect
of shared nonlinguistic experience to ensure that the separate storehouses
of exemplars each of us possesses well overlap extensively. In its role in the
learning of cultural exemplars and other kinds of cultural knowledge,
linguistic communication is perhaps the paradigm case of the process,
discussed in the section on cultural snaring in chapter 5, by which the modal
patterning characteristic of human social life fosters shared understanding.
At the same time that linguistic experience culls some potential meta-
phors, it vastly enlarges our fund of such metaphors, adding many more
exemplars than we could ever collect in the course of our nonlinguistic
experience, including conventional metaphors like "forging" a proposition
and others based on obsolete practices like being "on tenterhooks" that
may not speak to any nonlinguistic experience we have had, but the points
of which we have gleaned perfectly well from hearing (and reading) them
used. Conventional metaphors like this last one are good illustrations of
the point we made, in chapter 5, about the possibility of either historical
continuity or historical change in the meanings evoked by public forms
that remain fixed. Some of the meaning of these metaphors is preserved
through their continued use as exemplars; the rest of their meaning,
attached to practices that are in disuse or are not widely known, has been
lost. Of course, the fact that the original practice from which a metaphor
was derived is no longer in use may hasten disappearance of the metaphor
itself, and this may be the case with being "on tenterhooks." In a recent
lecture to an audience of undergraduate anthropology majors,
I gave it
as an example of a metaphor still in use even though most of us do not
know to what it referred. Several students came up afterward to tell me
that they had never heard the metaphor and had no idea what it meant.
It should be clear that this explanation of how metaphors become shared
does not assume that the stock of them that speakers share is fixed and
unchangeable. Indeed, because they are typically not very tightly connected
to emotion and motivation, metaphors may be especially prone to erasure
from historical memory, as the "tenterhooks" example illustrates. Older
readers should be able to think of other metaphors that have disappeared
or altered during their lifetimes. An example that I have noticed is that,
during my life, Mother Teresa has supplanted Albert Schweitzer as an
exemplar of the saintly helper of the poor, in remarks on the order of "He's
a regular Mother Teresa." When cultural exemplars like these that serve as
Research on shared task solutions 151
sources of metaphors do change, they do so in the same way in which any
schema alters. A new connection is made. This might be between an already
known aspect of experience and a new exemplar for it, adding that exemp-
lar to those already linked to that experience. In this way, presumably, we
came to adopt Mother Teresa as a cultural exemplar of the selfless benefac-
tor. Similarly, we learned to think of the computer as a cultural exemplar of
complex computational processes and hence as a potential metaphor for
mind, so that English speakers now readily talk about, and can understand
each other's references to, our brains having "crashed" or our being unable
to "access that file" in our memories. At the same time, connections to
earlier exemplars that have now largely disappeared, like Dr. Schweitzer
and tenterhooks, cease being learned. Or, the new connection might be
between a novel or newly attended-to aspect of an experience and an
already well-known cultural exemplar. Thus the historian E. P. Thompson
(1967) tells of the way in which time, reconceptualized in terms of the
amount of work employees did for their industrialist employers, came to be
represented as money in the course of the eighteenth and nineteenth
centuries. When new metaphors like these catch on, it must be because they
both capture features of experience that have already become important to
many, and draw on what many have already come to consider good
exemplars of those features.
In the course of their use in ordinary talk, then, metaphors typically do
not give rise to new understandings; rather, they reinforce existing under-
standings by clarifying them. This is not to say that the kind of thinking
involved in metaphor usage can never lead to new understandings. Indeed,
the cultural exemplars that provide the metaphors we employ have another
use: This same fund of exemplars also supplies us with analogies with
which we can try to reach understandings of new experiences. Dedre and
Donald Gentner (Gentner and Gentner 1983) address one kind of situ-
ation in which thinkers actively seek such analogies. These researchers
gave nonexpert subjects the unfamiliar task of explaining how electricity
works, and found them reasoning about it in terms of the flow of fluid
through channels and the rush of crowds through narrow corridors -
exemplary, for them, of how they imagined electricity running through
wires. Claudia Strauss and I (Strauss and Quinn 1992) show, similarly, that
children asked to explain something beyond their ken, namely, marriage,
do so by analogy to playmate relationships (in the case of a six-year-old
girl) and friendships (in the case of a fourteen-year-old girl), which exemp-
lify dyadic relationships for them at their ages. It is likely that we charac-
teristically try to understand unfamiliar experience by analogy to familiar
in this way, using exemplars for this purpose. Thus the flow of fluid (for
some subjects) or the rush of crowds (for others) is already strongly linked
152 Practice and possibilities
to movement through constricted passageways for the naive physicist
approaching the problem of how electricity travels through wires; play-
mates (in the case of the younger girl) or girlfriends (in the case of the
older), are already strongly linked to dyadic relations for the child asked to
tell about marriage. As both these cases illustrate, thinkers are able to
reason analogically about the new domain in terms of the old. As they also
show, some of the conclusions at which these thinkers arrive are likely to
be faulty; if married adults, in the one case, and trained physicists and
electricians, in the other, do not rely on these naive metaphors it is because
they have learned more useful ways to conceive of the experience at issue.
At the same time, these are unusual encounters with experimentally con-
trived tasks; children do not ordinarily think very hard or talk very much
about marriage, nor nonspecialists about electrical flow. Hence, while the
examination of such analogic thinking opens a window into one process by
which potential new metaphors might originate, especially telling for my
argument about metaphor is that these particular analogies invented for
thinking about new situations do not survive to serve as metaphors in
speech. Were the analogies more suitable and useful, they might have
gained currency as metaphors. More generally, it is important to distin-
guish the use of analogy to the known to help us in understanding the
unknown, from the use of metaphor to make a point to listeners. While
both tasks draw on cultural exemplars, only the former task typically leads
to new understanding. Indeed, speaker and listener both need to appreci-
ate why a metaphor - even a newly minted one - fits, if the communication
is to succeed.
At the other extreme from newly invented metaphors, occasionally
something has become so famous for a given aspect of the world, that its
use as a metaphor is overdetermined. It is in such special cases of highly
popular metaphors, such as that encapsulated in the saying, "Time is
money," that Lakoffand Johnson (e.g., 1980:8) find their favorite illustra-
tions for their view of metaphor as underlying conceptualization more
generally. That metaphors are widely shared, however, does not establish
that they are any more fundamental to conceptualization than any other
widespread cultural understanding. (That money does not, in fact, con-
strain the way Americans conceptualize time is suggested by the many
other metaphors for talking about it other than as being "saved,"
"wasted," "spent," and so forth - time is also said to "stand still," for
example, or "slip away," or "fly," or "flow," or be "ripe," or "right" - and
our ability to alternate readily among these metaphors.) Thompson's
(1967) aforementioned history shows, to the contrary, that metaphorical
talk of time as money followed and reflected rather than underlying and
instigating change. Originating as it did with the industrial revolution this
Research on shared task solutions 153
phrase captured the new work habits industrialists were imposing on their
workers, and then was used by these employers to inculcate this new order,
in the face of considerable resistance from their employees. Thus the
metaphor manifested a much more extensive and multiplex shift in Euro-
American understanding of the use of time, due to a profound change in
economic life.
In other words, time conceptualized as money is part of a larger theme
concerning work discipline under industrial capitalism that, like other
themes characterized in chapter 5, has spread into other domains kindred
to that of industrial labor. The metaphor encapsulating this theme has
spread with it. Like other themes this one has not spread to domains that
are experientially unlike, or differentiated from, that of wage labor. The
temporal rhythm of family life, for example, is largely immune to being
thought about in these monetary terms. (Witness the feminist complaint
that economists' analyses have traditionally overlooked the monetary
value of housework.) Even occupational work time does not always read-
ily convert into monetary terms: there is great resistance on the part of
patients to doctors charging for time spent on the telephone with them, for
example, and artists are not expected to charge for their work according to
the time put into it.
A more challenging situation is presented by the classic case of the
conduit metaphor (Reddy 1979) for talking about meanings as transmitted
in words - as in, "I gave you that idea" or "I find it impossible to put my
thoughts into words." Here is a conventional metaphor argued to have
infiltrated such a large proportion of our available language and to have
become such an automatic way of speaking about linguistic communica-
tion (ibid,:297-299) that it goes beyond clarifying our understanding of
such communication to constrain it. According to Reddy, the conduit
metaphor has made it difficult to think about the transmission of meaning
in any other way. But by accepting the predominance of the metaphor in
linguistic expressions and everyday speech as evidence for its control over
our thinking, Reddy, like LakofT and others, makes this metaphorical
language itself seem coercive - "falling] victim," Linger (1994:290) has
pointed out, "to the very model he renounces." More reasonable, from my
position, is that the language reflects an underlying cultural model for
linguistic communication.
The conduit metaphor may be said to comprise a cultural model in that
it encompasses an interrelated set of elements - the speaker, the listener,
the speaker's meaning, the words that convey that meaning, and the
relevant actions and intentional states on the part of speaker and listener -
that model how communication works. To be sure, in this cultural model
the linguistic communication of meaning is conceptualized by analogy to
154 Practice and possibilities
the transmission of physical things. This analogy is the source of the many
metaphors for communication as transmission that Reddy reports. It is
hardly unusual, however, for cultural models or parts of these models to be
cast in analogic terms in this way - although many cultural models are not.
For example, as has already been indicated, and as will be the subject of
further discussion in the next section of this chapter, the cultural model of
marriage we are considering is partially (though only partially) delineated
by analogy to success, with the result that some of the metaphors for
marriage are drawn from the domain of successful enterprise. Such con-
ceptualization of one domain of the world in terms of another can be
considered an instance of thematic spread, discussed in chapter 5. That
domains like that of physical transmission lend themselves as analogies to
other domains, such as meaning, is of interest, of course, and worthy of
empirical study. However, it is not the analogies on which some cultural
models or parts of these models are built, nor the resulting metaphors
provided by these analogies, that make it hard to think in any other way
than in the terms of the cultural model. A hallmark of cultural models of
all kinds is their "referential transparency" (Hutchins 1980:12); once
learned, they become "what one sees with, but seldom what one sees"
(ibid.). It is the typical transparency of cultural models that gives them the
ability to constitute our reality - including our language. Thus, for
example, the mostly nonmetaphorical language given to Americans by
their model of the mind is as automatic of use and as difficult to think
without as are the metaphors owed to the conduit model of communica-
tion (D'Andrade 1995:167).
Why cultural models should have this quality of transparency can be
understood within the connectionist framework we are employing in this
book. If the elements of a cultural model are like the units of a connection-
ist network, they need not be in consciousness. Indeed, given the limitation
on what can be held in short-term memory, the many parts of such large
schemas could not all be in consciousness at the same time. Instead, it
appears, only outputs of cognitive processes, or parts of these processes
involved in cognitive "glitches," come into consciousness. (For example,
when we produce a given word, we are conscious of the word but not of
how we arrived at it. Only if - as happens to me more and more frequently
- we cannot think of the right word, are we likely to become conscious of
our search through a relevant part of our network of related words to find
it.) It is just this characteristic of structuring our understanding while being
largely out of our consciousness that Hutchins captures when he says that
we see with cultural models, but do not ordinarily see them. It is a
characteristic no less true of cultural models reliant on analogy to other
domains of experience, than of those not so based.
Research on shared task solutions 155
This argument against metaphor as controlling thought is not meant to
deny that the language in which it is cast may have some role in perpetuat-
ing and spreading a given cultural model. Again, however, this would true
of any language attached to a cultural model, metaphorical or not. More-
over, in the framework we set out in chapter 5 to deal with the historical
durability and sharedness of cultural schemas in general, language is only
one of many repositories of cultural meaning external to the individual and
one of many sources of patterning of the external world that the individual
In the context of the present analysis of metaphor, the most important
point to follow from a consideration of the conduit metaphor is the need
to differentiate it analytically - just as, I argued, individual analogizing
must be differentiated - from what I have been describing. The concep-
tualization of one domain of experience by analogy to another, exemp-
lary one, in a cultural model, has a significant relation to the use of
analogy by an individual thinker to explain the unknown in terms of the
known. We may imagine that cultural models incorporate just those
cultural exemplars that prove natural and helpful to many people in
thinking analogically about a given domain of experience. Unlike the
notion of electricity as a fluid flowing through channels or that of mar-
riage partners as playmates, these are analogies that have caught on. In
the same way as do cultural exemplars that provide individuals with
analogies between the known and the unknown, those incorporated into
cultural models facilitate comprehension of the domain so modeled (as
they inhibit alternative understandings of it, Reddy illustrates persuasive-
ly with the case of the conduit metaphor). These analogies perform a
different task altogether than do the metaphors used in speech to clarify
points that speakers wish to make. The metaphorical and other language
that grows up around the culturally shared analogies on which some
cultural models are wholly or partially based - e.g., "I gave you that
idea" - becomes the only or the handiest language available to a speaker;
it has no point to make. On the other hand, the use of metaphor to
clarify the speaker's point - e.g., "It was stuck together pretty good" - is
intentional. Instances of the first sort might helpfully be designated obli-
gatory use of metaphors, in the sense that they are so embedded in the
language that speakers can find no alternative to their use. Those of the
second type, the focus of my analysis here, might be distinguished as
selective use of metaphors, in the sense that speakers intentionally select
them. Actual usage may reflect mixed cases. (This is not the same as
distinction between conventional and novel metaphors. Metaphors inten-
tionally selected to make a point may or may not be conventional ones.
Metaphors that have become so embedded in language that they are the
156 Practice and possibilities
only ones, or virtually the only ones, available to speakers, are conven-
tional in the broad sense that all language is.) The most general point is
this: Metaphors sometimes occur because they have been intentionally
selected and produced to communicate a point, but not always. Some-
times metaphors stem from an analogy creative people have invoked to
explain an unknown phenomenon; sometimes they reflect a tacit cultural
model speakers share, one that is based on an analogy between one
domain and another. Failing to distinguish among these cases, investiga-
tors have failed to recognize the actual functions that metaphors serve,
used in each of these different ways.
My analysis of the use of metaphors for marriage suggests that the task
of clarifying what we mean to say provides one circumstance - a univer-
sally human, experientially common one - under which understandings
will become shared, or cultural. Of course, the circulation of cultural
exemplars in metaphor is only one way in which these metaphors them-
selves can become widely shared. I have suggested, for example, that
metaphors like that of marriage as a manufactured product, encapsulat-
ing, as it does, complex relations among elements of the cultural model for
marriage, may gain particularly widespread usage in the context of mar-
riage because they conjure up the experience of marriage in its near
entirety. We should now be able to see that metaphors like this one are
prone to popularity for another reason as well. The manufactured-product
metaphor, for one, reflects and encapsulates a prominent cultural theme, a
theme that colors Americans' understanding, not only of marriage, but of
the many domains of life to which they apply the values of hard work,
pride in one's work, Yankee ingenuity, and other American virtues asso-
ciated with making things that last and work well. This metaphor, like that
of time as money, has spread to new domains like that of marriage for the
same reason that, as we argued in chapter 5, any understanding becomes
thematic - because of the multiplicity of similar-but-not-identical experi-
ences through which it is learned. Having colonized not only marriage but
many domains in this way, the manufactured-product metaphor now finds
itself used and reused in many contexts, and through its repeated and
varied use, ever more widely shared.
Retrieving cultural understandings from metaphors
There are good and bad methods for reconstructing the shared under-
standings people must have in mind from the metaphors they use. I have
suggested that metaphors are good clues to the cultural understandings
that lie behind them because of what they do: In drawing on cultural
exemplars and using these exemplars metaphorically to clarify the
Research on shared task solutions 157
speaker's meaning for the ordinary listener, they also spell out this mean-
ing, and the cultural understanding behind it, for the analyst. But in order
to exploit this analytic potential of metaphors it was necessary for me to
attend to their usage in actual discourse. I could not have discovered that
the metaphors for marriage fall into a small number of classes had I not
systematically examined extensive discourse on the subject. Nor could I
have gained insight into the way in which these varied metaphors are used
to highlight the shared conceptualizations of marriage reflected in these
classes, had I not looked closely at metaphor use in this discourse, which
provided me with a window into this process.
Close analyses of discourse in which metaphors are used gives evidence
that, indeed, speakers do select their metaphors to match the points they
are developing or have already made, as we would expect if they were using
these metaphors to clarify these points - rather than the points arising de
novo from the metaphors. For one piece of this evidence, consider the way
in which the speakers quoted in the last section selected just the metaphors
of marriage as "nailed in cement" and "everlasting Gibraltar" in nature,
that would do a good job of conveying the point of the story they were
intent on telling: not just their expectation that marriage was lasting, but
their having overestimated its durability. For example, the husband who
uses the "Gibraltar" metaphor starts out the passage in which it occurs by
observing that his marriage had changed, due to an extramarital relation-
ship his wife had. That relationship had caused him, he continues, to
question his own lack of understanding of his marriage, who his wife was
in it, and how they interrelated. Finally, he says, "And I think also it raised
for me kind of the whole idea that I really didn't know who she was very
much. And that my confidence in the everlasting Gibraltar nature of the
whole thing was rather naive." It is difficult to imagine that the point
extracted from the metaphor governed the topic of this story from its onset
through its development.
To summarize quickly other such evidence that I have elsewhere (Quinn
1991) provided in detail: in this discourse about marriage speakers also
sometimes articulate, in nonmetaphorical language, the point they are
trying to make with a metaphor. Usually this explication of the metaphor
comes immediately after its use. Speakers' ability to amend such non-
metaphorical interpretations with the speed of speech suggests that they
had the point in mind already, rather than being led to it by a previously
unrealized entailment of the metaphor. Often this nonmetaphorical re-
statement seems to be made in order to disambiguate the point made by the
metaphor, especially in the case of metaphors that highlight marital shar-
ing - a concept susceptible to multiple interpretations. A nice example of
such disambiguation occurs in the following brief passage: "We present a
158 Practice and possibilities
front of sorts together. Not a couple front, but a united front. A common
value system that we share and things that we want to show to other
people. Where we're at." [3H-4] Occasionally, also, these speakers make
their point nonmetaphorically first, and only then restate this point meta-
phorically. If speakers* ability to restate immediately a metaphorical point
nonmetaphorically is suggestive, cases in which they do the reverse demon-
strate quite unequivocally that the metaphor did not give rise to the point.
Just as tellingly, speakers may string together two or even three different
metaphors, seemingly for emphasis as well as clarity, to bring home a given
point. A single example containing both these kinds of evidence will have
to suffice by way of illustration here: "But my feelings are, our relationship
is permanent. It's no - we know too much about each other. There's no
getting away from each other, you know. 'Cause I feel like she got me, and
I feel like I got her." [2H-8] In this passage, the lastingness of marriage is
first described nonmetaphorically - "our relationship is permanent."
Then, this point is restated in a metaphor of spouses who are unable to
escape from one another, and that metaphor expanded upon in another
one, of spouses as each other's possessions.
Another kind of evidence comes from metaphor use in reasoning. For
both Lakoff(Lakoff and Kovecses 1987:219) and Johnson (1987:130), a
key sense in which metaphors constitute understanding is that they gov-
ern reasoning, allowing us to draw just those inferences that follow from
metaphorical entailments. But in my close analysis of reasoning about
marriage in actual discourse, I have elsewhere (Quinn 1991:84-87) shown
that far from consistently adhering to metaphorical entailments, reason-
ers are just as likely to switch metaphors in the middle of a piece of
reasoning in order to reach a desired conclusion, discarding the meta-
phor that does not carry them to this conclusion for one that does.
Indeed, a scrap of reasoning introduced in an earlier section of this
chapter, "If it gets static in our relationship then that's when we split I
guess, or start going in a different direction," exemplifies such a shift
from one metaphor to another in mid-argument. Here a metaphor of
stasis stands for lack of marital benefit while a second metaphor, of
traveling companions, characterizes the separation of the married couple
that signifies the end of marriage, resulting from the lack of benefit. I
have shown, furthermore (ibid.:87-88), that other speakers conduct
identical reasoning about marriage without relying on metaphors at all
to do so. For, as will be taken up in the next section, not only are the
concepts underlying metaphor use shared, but the causal relations link-
ing these concepts and underlying the reasoning people do about mar-
riage are shared as well. Again, it would appear that speakers have this
shared causal schema for reasoning about marriage already in mind, and
Research on shared task solutions 159
that they select just those metaphors having entailments that roughly
match the reasoning they aim to do.
Previous analyses of metaphors by LakofT, Johnson and others have not
been based on systematically collected discourse. In none of their writings
do they tell us exactly where and how they have collected their examples. It
seems likely that their assemblage of metaphors relies, in some combina-
tion, on their own and other peopled memories, on the perusal of written
texts, and on cases unsystematically drawn from overheard speech or
accidentally encountered in text. They take and analyze even the meta-
phors they collect in this way out of the context of their use. This method
has several serious consequences.
First, the corpus of metaphors used within a given domain that is
collected by this method is bound to be not just incomplete, but biased as
well. The metaphors most likely to be remembered or encountered are the
most common ones. Linguists who study only the most popular metaphors
in this way are reminiscent of anthropologists of public culture who focus
on the most standardized features of the public world. In both cases, the
resulting analyses are likely to be misdirected. In the case of metaphor, the
consequence is misidentification of these most popular metaphors as com-
prising the unique system of basic-level metaphors in that domain. In a
rhetorical strategy these students of metaphor favor, moreover, their
accumulation of examples that fit this assumption about unique systems of
basic-level metaphors, distracts readers' attention from all the metaphors
that do not fit and that might otherwise occur to us - as I have indicated in
the case of metaphors for time. Another example in which the uniformity
of metaphorical usage is similarly overestimated is LakorT and Zoltan
K.dvecses' (1987) claim that heated fluid in a container is the "central
metaphor" for anger. I would argue from their own material that this
appears to be one of a class of equivalent metaphors - anger as insanity, as
a wild animal, and so forth - that capture an underlying cultural concep-
tualization of anger as dangerous and unpredictable because uncontrol-
lable. Lakoff and Johnson make the same error when they claim that an
ongoing journey is the central, and hence defining, metaphor for love
relationships (LakorT 1990:49) and for marriage in particular (Johnson
1993:53-57, drawing selectively on my own material, no less). As I illus-
trated earlier in this chapter, the journey metaphor is only one of a number
of metaphors, certainly not the one, used to talk about marriage. To
recognize that there is not one central metaphor, but a number of meta-
phors that draw on different cultural exemplars to call attention to differ-
ent features of a given domain, requires the more systematic analysis I have
Secondly, considering metaphors out of the context of their actual use in
160 Practice and possibilities
discourse, as these researchers do, means they are not in a position to verify
their analyses against the actual details of this use. Again, this approach is
reminiscent of the work of anthropologists who believe they can analyze
public symbols out of the context of their use and apart from their
meanings to the people who use them. Neither group of researchers gets
close enough to the processes by which metaphors, in the one case, or
symbols, in the other, are internalized and used, to reveal how they work.
Thus the method linguists have used to study metaphor is blind to subtle-
ties of usage that, as I have demonstrated, show how metaphors are being
used to clarify the points being made in speech, rather than constituting or
entailing these points. Such an analytic approach disallows any cognitive-
ly-informed investigation of the cultural understandings reflected in the
metaphors, how shared they are, or how they have come to be shared.
Analysis 2. A shared schema for reasoning about marriage
Up to now we have taken as givens the features of marriage that account
for the eight classes of metaphor I have identified, focussing on the way
metaphors are selected from these classes to exemplify these features and
the relations among them, and thus clarify the points speakers wish to
make about marriage. The classes of metaphor for marriage and the points
they exemplify, I will next show, make overall sense in terms of a more
encompassing schema for marriage. This is a schema for performing a
conceptual task of a different (though just as commonplace) sort from that
of finding apt metaphors to communicate one's point. This other task is
the reasoning about marriage Americans are called upon to do in the
course of daily life.
Reasoning is not limited to the problems with which teachers vex their
students in formal logic classes.
Thus, people use the schema I am about
to describe to reason about their own marriages, the marriages of others
they know, and marriage in general. There is much about marriage to
consider. People may feel that their marriage is not as successful as it should
be, for example, and wonder what to do about it. A person may want to
know, more specifically, how to respond to a spoused unhappiness and
dissatisfaction. Sometimes these questions are hypothetical, sometimes, as
we can tell from some interviewees' observations, critical to immediate
action. Or, like the husband who thought his friends' marriage was "nailed
in cement" if ever a marriage was, a person may be trying to make sense of
the unexpected and unsettling divorce of a relative, friend, acquaintance, or
colleague. Contrariwise, people sometimes wonder why some couple they
know, their own parents perhaps, never divorced, given the kind of mar-
riage they had. People may be concerned or curious about what they read in
Research on shared task solutions 161
newspapers and magazines about the rising divorce rate, for instance, or
the increasing number of cohabiting couples who never marry.
In this half of the chapter, I will delineate the schema used to perform
these reasoning tasks, show how it works, and bring evidence for its
widespread use. Then I will turn to a consideration of why such a schema
should have become widely shared. I will end by considering the advantage
of my close analysis of the understandings that inform people's reasoning
about marriage over an approach that pays less close attention to these
intrapersonal understandings, for characterizing developments in contem-
porary American marriage.
Two concise examples will serve the immediate purpose of conveying
how commonplace is the reasoning that we do about everyday concerns
such as marriage, and how readily and automatically we perform such
reasoning tasks as they arise. An interviewee, talking about the observable
conflict and evident dissatisfaction in her sister-in-law's marriage, con-
cludes, "I'm sure they must have something good in their marriage or they
wouldn't still be together" [1W-3]. This entirely ordinary piece of reason-
ing rests on this woman's general understanding that if a marriage is
mutually beneficial, it will last; it's having lasted she treats as proof that it
has been beneficial. The analysis to come in this chapter will show that this
snippet of natural reasoning is not singular. Rather, it follows a predict-
able path, defined by a schema this reasoner shares with other interviewees.
Indeed, the alert reader will have recognized that the interviewee who said,
"If it gets static in our relationship then that's when we split," was
reasoning to the negative variant of this same conclusion. The ease with
which they arrive at conclusions like these about marriage depends upon
reasoners' use of this culturally shared solution to this reasoning task, as
we may suppose that other recurrent reasoning tasks rely on other such
cultural solutions. Thus everyday reasoning is another context in which
cultural sharing arises.
The shared schema
Before showing how reasoning about marriage depends on a shared
schema, I will outline the schema itself, and suggest the problem to which it
is the solution. The classes of metaphors for marriage are linked in the
following way. Americans expect marriages to be lasting, shared and
mutually beneficial. I will return in the next chapter to these three expecta-
tions and their common basis in experience. Taking them as given for the
moment, the three together set in motion a well-understood sequence of
Marriages must be shared in order to be mutually beneficial. This is
162 Practice and possibilities
because the benefit of marriage, in contemporary US Americans' under-
standing of it, is psychological need fulfillment - an equation made evident
in interviewees' statements like the following one (to be quoted more fully
later): "[I]t's not worth sticking together because life's too precious to
waste your time, with another person. Unless they're really fulfilling you
on an emotional level." A physically absent or socially separate or emo-
tionally disengaged spouse cannot fill one's needs; all of these states of
affairs are thus violations of Americans' expectation of the shared life that
marriage entails.
Marriages must be beneficial, in turn, in order to last.
This is because, Americans understand, parties to any utilitarian exchange
will agree to it only if and when it benefits them, feeling free to terminate it
if and when it fails to do so. And marriage, in the United States in the latter
part of the twentieth century, has become very much a contractual rela-
Lack of fulfillment, then, leads to the lack of expected marital
benefit that results, in turn, in divorce. This lack of fulfillment is inter-
preted by interviewees as a matter of incompatibility, defined as one
spouse's inability to fulfill the needs of the other. Given the way in which
Americans marry - that is, for love (as will come under scrutiny in the next
chapter) - they often do not know enough about each other to gauge how
well-matched the two will be in terms of their ability to meet each others'
needs. More than one husband spoke of the moment, early in his marriage,
when he realized that he did not even know what his wife's needs were,
never mind how to fill them. As should be clear, this cluster of understand-
ings about marital benefit, the fulfillment that constitutes it, and the shared
life and compatibility it requires are distinctively American; in other
societies marriage may be understood quite differently (as it was under-
stood in American society in other times).
There is an inherent contradiction between the expectation, on the one
hand, that a marriage should last and the expectation, on the other, that an
unfulfilling and hence nonbeneficial marriage should be ended. The incom-
patibilities between spouses that initiate the chain of events exposing this
contradiction are hence viewed as difficulties for the marriage. These
incompatibilities that are the stuff of marital difficulty being virtually
inevitable, this difficulty, in turn, is understood to be the marital norm
rather than the exception. As one interviewee put it, "I guess I really don't
deep down in my heart think that probably most people's marriages are
any easier than mine is. That everybody has enormous difficulties" [4W-3].
Some social commentators have argued that the conflict between these
two expectations of marital lastingness and marital fulfillment represents
an historical shift from one view of marriage, family, and relationships
more generally, to another. Robert Bellah and his coauthors (1985:101),
for one notable example, have called this a shift from "the traditional view
Research on shared task solutions 163
of love and marriage" as an obligation to sustain committed, enduring
relationships to the newer "therapeutic attitude" toward it that privileges
individual self-actualization. What these authors do not note, but what my
interviewees demonstrate abundantly, is that they and other Americans
living at this historical juncture share a cultural solution to the cultural
contradiction they face between an older and a newer construction of
The solution is quintessentially American. Difficulties are to be over-
come by effort. In the case of marital difficulty what is called for is largely
the effort to achieve spousal compatibility so that the marriage will be
more beneficial. This requires learning what the other person's needs are
and learning how to fill them; sacrificing some of one's own desires, if
necessary, to accomplish that; figuring out one's own needs and com-
municating those to one's spouse; and coaching the spouse in their fulfill-
ment. The early part of marriage is often described as being dominated by
these difficulties, which eventually taper off as spouses learn to fill each
others' needs (and, usually also, to accept that some of these will remain
unfulfilled within the marriage). However, these married people are cogni-
zant of the difficulty of responding to new needs that may arise as each
spouse grows and changes in the course of a marriage - as they and other
Americans assume all people do over time. A marriage in which the major
foreseeable difficulties have been successfully overcome is spoken of not
only as a lasting, but also as a successful one. The risk, of course, is that
difficulties will not be overcome and the marriage will fail. This set of
shared understandings, in its entirety, comprises a cultural model of mar-
How the schema works and why it is efficient
Why should we find people reasoning according to a common schema? In
the last chapter it was suggested that schemas like this one for reasoning
about marriage become shared in large part because of their efficiency for
performing the cognitive tasks they address. In order to evaluate this claim
in this case, I must first demonstrate how the schema works as an efficient
solution to this reasoning task.
In reconstructing how the schema works, I have found it particularly
helpful to think in the terms, discussed in chapter 5, of Edwin Hutchins'
(1986, 1995) characterization of culturally snared mediating structures
which, coordinated with the world in which a task takes place, allow actors
to perform given tasks. A simple, clear example is a written procedure such
as a checklist (Hutchins 1995:290-312), which controls the order in which
components of a task are to be executed. A difference between the equally
164 Practice and possibilities
commonplace lists, scales, dials, charts and landmarks - the cultural
artifacts - that mediate performance of the navigation tasks Hutchins
analyzes, and mediating structures of the sort I will describe for reasoning
about marriage, is that the latter are not externally mediated at all. (The
words in which reasoning is expressed in the course of speech are a partial
record of that reasoning, but do not mediate it.) As was noted in chapter 5,
Hutchins emphatically intends that the notion of mediating structures
encompass internalized cultural models for task solution; and others (e.g.,
D'Andrade 1989a:822; Rumelhart, Smolensky, McClelland and Hinton
1986:46-47) have speculated that the task of reasoning, in particular, must
be mediated by culturally provided mental models invented for this pur-
pose. But until now we have not had good descriptions of such internalized
cultural models in actual use, of the sort I provide here, so that we can see
how they actually work as mediating structures for performing reasoning
Viewed as such a mediating structure, the cultural schema for reasoning
about marriage takes the form of a prototypical or idealized sequence of
marital events.
In this structure, one event is linked to the next by a
relation of causality, the concatenation of such relations forming a causal
chain. The resulting structure is used to reason about a common marital
quandary. It is idealized in two ways important for the performance of this
reasoning task. Reflecting as it does these two forms of idealization,
natural reasoning such as that interviewees do about marriage - like the
metaphors they use - provides the analyst with excellent clues to the
cultural assumptions they share.
First, the events that can occur are highly circumscribed, being of a
limited number and following one another in an invariant order. These
events, as we have seen, are: lastingness, sharedness, mutual benefit, fulfill-
ment, compatibility, difficulty, effort, success, and risk of failure. The
sequence is: marriages are ideally lasting, shared and mutually beneficial.
Marriages that are not shared will not be mutually beneficial and those not
mutually beneficial will not last. Benefit is a matter of fulfillment. Spouses
must be compatible in order to be able to fill each others* needs so that
their marriages will be fulfilling and hence beneficial. Fulfillment and,
more specifically, the compatibility it requires, are difficult to realize but
this difficulty can be overcome, and compatibility and fulfillment achieved,
with effort. Lasting marriages in which difficulty has been overcome by
effort are regarded as successful ones. Incompatibility, lack of benefit, and
the resulting marital difficulty, if not overcome, put a marriage at risk of
A second idealization is that the causal relations between pairs of events
in the sequence are themselves highly simplified (see Quinn 1987). A
Research on shared task solutions 165
successful marriage is a lasting one and a lasting marriage successful,
always. If a marriage is beneficial it will last, and if it has lasted it is
beneficial, categorically. The benefit of marriage is fulfillment, pure and
simple. If two people are compatible, their marriage will be beneficial, and
if they are incompatible, it will not be, period. Incompatibility causes
difficulty, without qualification, just as a compatible couple will have no
difficulty. Effort overcomes difficulty, absolutely. And so on.
Of course, the logical relations involved here are not that simple, as
shown by some examples taken from an impromptu analysis of my ma-
terial by Roy D'Andrade (personal communication).
Lastingness is a
property of a successful marriage, so that all successful marriages are
lasting ones but not all lasting marriages are successful ones; in fact, there
are plenty of unsuccessful old marriages around. Mutual benefit is a
sufficient cause with respect to a lasting marriage. That is, while mutual
benefit always makes a marriage last, a lasting marriage need not always be
a mutually beneficial one, and we all know such cases; we call them
"unhappy" marriages. Mutual benefit is actually a property of psychologi-
cal need fulfillment; as such, it can also be a property of other marital
benefits, such as money or social status, and sometimes people marry for
them. Effort is a requirement, but only one requirement, for overcoming
difficulty - a necessary but not a sufficient cause; we all know that the best
efforts cannot always save a marriage. What I am asserting is that in the
course of ordinary reasoning these complexities drop away. People do not
follow strict inference. Rather, they appear to substitute something like
plausible inference (see Hutchins 1980:56), that is, treating likely events as
if they were always true. Next we will consider why this should be so.
A telling example of plausible inference is the previously quoted com-
ment, "I'm sure they must have something good in their marriage or they
wouldn't still be together." The speaker can cite the marriage having lasted
as proof that it has been beneficial only because she is treating benefit as
both a necessary and a sufficient cause of lastingness. Even as she makes
this assumption for the purpose of this piece of reasoning, she immediately
amends the statement, continuing, "Who knows? They might be staying
together for their little boy's sake but they - she doesn't seem to be as
happy as she could be." What this fuller passage illustrates is that at the
same time they reason by means of the prototypical event sequence I have
described, interviewees can also recognize, and incorporate into their
reasoning when called upon to do so, marital events and complex causality
not contained in this prototypical sequence of events.
Still it was unusual
for this woman to go back, having finished her brief piece of reasoning and
arrived at the conclusion she intended to draw - and fill in some of the
complexity of the situation that suggested an alternative interpretation of
166 Practice and possibilities
events. In the course of ordinary discourse, reasoners do not typically do
this; they seem bent solely on reaching their desired conclusions.
This reduction of all causality to plausible inference allows the event
sequence I have described to work like a template in producing chains of
actual reasoning about marriage. As D'Andrade (1989a:822-823) ex-
plains, such an imaginary model of the world must be brought into
coordination with the real world, the reasoner deciding that some state of
the world aligns with a given point on the model - in this case, an event in
the prototypical event sequence. Once such an alignment is achieved, the
reasoner can reach conclusions about the world by reading them from the
model. Thus, if a person observes that a real-world marriage is fulfilling
and hence beneficial, he or she can use the model marriage to predict
outcomes of that state of affairs - for instance, that it will last and succeed.
If a person determines that a marriage is lacking in benefit, he or she can
infer that it will be difficult or that it will not last. Like the wife concluding
that her sister-in-law's marriage is beneficial because it has lasted, they can
reason as readily in the other direction, from effect to cause. And, like the
wife who imagines leaving her marriage if it gets "static" - that the
marriage will not last if it is no longer mutually beneficial - they reason just
as readily to negative conclusions. Speakers reason in this way from any
event to any other event in this causal chain, concluding, for instance, that
compatible marriages will last, that unfulfilling marriages will be difficult
ones, that unsuccessful marriages were not sufficiently shared, and so
forth. In sum, people can reason across causal links from any event or its
negation to any other event or its negation, and they can reason either
forward or backward across these events. They do so readily and rapidly.
People could not accomplish the reasoning they do about marriage
without some such mediating structure as I have described. In the first
place, a limited sequence of events in a fixed order makes it possible to
reason to conclusions about marriage without confronting an unmanage-
able number of potential outcomes. In the second place, a simplified
causality makes it possible to reach these conclusions without getting
tangled up in endless complications and limitless shades of probability.
Imagine, for example, what would happen if a reasoner wanting to infer
whether a marriage was ending because it had been unbeneficial had first to
examine the causal relation between lastingness and benefit to decide
whether there were any extenuating circumstances and whether this rela-
tion was transitive and could bear the inference, and further had to subject
the linking relations between benefit and compatibility, compatibility and
difficulty, difficulty and effort, and effort and lastingness to the same
That is precisely the kind of cognitive morass that the medi-
ating structure I have described is designed to circumvent.
Research on shared task solutions 167
The connectionist framework developed in this book suggests one other
important way in which the schema, once well learned, makes reasoning
about marital events more efficient. Within this framework we can explain
why the relations between events distant from each other in the causal
chain - say, effort and lastingness - appear to be no harder, and take no
more time, for reasoners to process than the relations between events that
are directly causally linked - say, effort and difficulty. While the causal
relations among these marital events are serial, the schema itself eventually
becomes a network of strong links connecting each event with every other
event in the sequence. This is because the more occasions a reasoner has to
reason between any pair of these events (or observe the pair linked in
another context than reasoning about it), the stronger the association
between the two events will become. The same is true, of course, of the
associations between these events and the causal relations that pertain to
them. Eventually, all the events and relations in the sequence become
strongly linked to one another in a densely connected schema. Paren-
thetically, once overlearned in this way the schema is largely out of
consciousness most of the time, and thus has gained referential transpar-
ency in our thinking.
Evidence for use of the schema
I next present interviewees* reasoning about a subset of everyday marital
dilemmas, as evidence that they do indeed share the schema for doing so
that I have delineated. Early in this chapter I demonstrated a striking
regularity in the classes of metaphors speakers draw upon in their talk
about marriage. I attributed this regularity to a small number of shared
underlying concepts, understandings of marriage that speakers select these
metaphors to exemplify. Now I will show an equally striking regularity in
the way these underlying concepts are related to one another in the
reasoning people do in ordinary talk about marriage.
Finding this structure was a methodological challenge. Underlying, as
it does, the varied language of different speakers* reasoning about diverse
marital experiences, the pattern is not immediately discernible. Structure
must be recovered from verbal reasoning by the rather painstaking ana-
lytic technique I next apply. Because they are internal and not readily
observable, and can be reconstructed only with difficulty, shared medi-
ating structures, such as the one for reasoning about marriage, have been
overlooked. The identification of such widely shared, yet wholly internal
and heretofore-unsuspected structures for the performance of cognitive
tasks, such as this schema for reasoning and the store of cultural exemp-
lars described in the first part of this chapter, attests to the need for what
168 Practice and possibilities
might be called "the ethnography of inner life," along the lines I have
At the same time, somewhat paradoxically, US American readers (but
not readers from elsewhere who bring with them distinctly different under-
standings of marriage) may find the excerpts of talk on which this analysis
is based so mundane as not to merit analysis, and the assumptions about
marriage extracted from them so commonsensical as not to have deserved
retrieval. How distinctive Americans' cultural model of marriage actually
is, from a comparative cultural and historical perspective, should be kept
in mind; I will touch on what is distinctive about this model further at the
end of this chapter and in the next. If, in this section of the chapter, I
provide more than a single excerpt to prove each point I adduce about this
schema, I hope readers will not view these as illustrations belaboring the
obvious. The multiple excerpts are evidence (of which I have documented
much more than I publish here) confirming that the schema I am descri-
bing is not just the singular invention of one or another interviewee, but is
used by all of them to reason about marriage. Readers who do not need to
be convinced of this should feel free to skip over the detailed analyses of
successive passages.
For reasons of brevity, I will provide evidence for the subset of assump-
tions about the relations between the effort required of marriage, the
benefits expected of it, the fulfillment into which marital benefits translate,
marital lastingness, and the success that a lasting marriage is considered to
be. I begin with a set of eight more passages in which speakers invoke the
causal link between benefit and lastingness or success, just as they do in the
two short excerpts already considered, about the relationship that could
become static and the sister-in-law with "something good" in her mar-
riage. These examples should both familiarize readers with my mode of
analysis and convince them of the pattern exemplified by all ten cases. I
then turn to briefer evidence: four passages in which speakers make the
link between effort and benefit, and finally, six other excerpts that rest on
the link speakers assume between effort and lastingness or success. I will
end with a single instance of a somewhat rarer type, in which a speaker
spells out almost the entire causal sequence between effort and lastingness.
A considerably longer, more fully documented piece of reasoning than
"I'm sure they must have something good in their marriage or they
wouldn't still be together," but one exhibiting the very same form, is the
And there isn't any signs right now that it's not going to be a very sthat our
marriage isn't going to be a successful marriage in terms of lasting. Not only just
lasting but our wanting it to last and enjoying each other. I think Rich is very
Research on shared task solutions 169
happy. I think Rich is getting affection, and having affection expressed, probably
the first time in his life, really. Really feeling accepted just for being Rich, you
know. And he acts it. I mean he says things - he'll say, you know, how happy he is,
and I really believe so. And, I know that - 1 think I might have said this - and some
people have said to me that it's really obvious that he feels for me, in the way he
acts, sort of thing. And certainly back on my part, I mean it's very nice knowing
that I'm loved that much. That is just a tremendous feeling. I - and he spoils me
rotten, at times, like when I was sick and on the couch, you know, he always
changed the channel on the TV, or turned it up or down, and he still does. I mean -
like he - after the seven o'clock news - at seven o'clock the news are over - he'll get
up, wrap my - the afghans around me, turn down the game shows, turn off the light
and kiss me good night for an hour. I mean, he really does spoil me, and I really like
it. So I don't see any reason at this point that anything is going to happen. And I
can't imagine it happening. So, maybe we'll be a success. [9W-10]
If this woman offers extended testimony to her conclusion,
perhaps it is
because at some level she is uncertain of it. There is the cautious "there
isn't any sign right now" and the telltale "maybe we'll be a success" at the
close, both suggesting the speaker's awareness that success is only a
likelihood. (In the event, she was right to reserve judgment; her marriage
ended in divorce several years after this interview.) Yet the structure of the
argument itself is just as clear in this passage as in the earlier ones: a
fulfilling marriage is a lasting marriage is a successful marriage. As in the
case of the not-so-happy sister-in-law and other examples to follow, an-
other's fulfillment is inferred from the evidence that that person acts happy
and says they are happy ~ happiness being the emotion that one ordinarily
feels when one is fulfilled. Rich is very happy, we understand, because he is
getting affection for the first time.
The context in which speakers most often invoke the assumption that
mutual benefit causes a marriage to last is one in which, like Rich's wife
and the following interviewees from my study, they have occasion to assess
the status of their own marriages:
You know, it's like I know everything that she need. That - you know, and the
things that I don't know are things that will come up in the future, I feel like she will
tell me, you know, maybe perhaps our relationship will allow her to get it, you
know, and vice versa. So I feel like, you know, we're here to stay, you know. I think
we got it - you know, I feel like - well I really feel it's, you know, concrete, you
know. [2H-8]
But I feel pretty mutual about, we both have as much at stake in the relationship as
the other person does. We both express to each other the same desire to keep things
going. [4W-7]
This last woman relies on a chain of only partially stated assumptions that
we all share about the role motivation plays in marriage and other human
affairs. The mutual benefits that she and her husband are deriving from
170 Practice and possibilities
their marriage are cast here as some unspecified assets which each has "at
stake" (and hence risks losing); marital lastingness is a matter of "keep[ing]
things going" on some, again unspecified, kind of path. Because she and
her husband are both benefiting from the marriage, we understand from
the order of the two sentences, they want it to last. What is left unstated,
and what we fill in for ourselves, is that wanting it to last, they will make
the effort to make it last (and because they make the effort it will do so). So,
we all understand, what they have at stake will, in fact, keep things going.
Similarly, since they are enjoying each other and feeling loved, Rich and
his wife not only expect but want their marriage to last.
Rich's wife and these last two speakers are offering assessments of the
chance their marriages have of lastingness and success. Such running
assessments can certainly be an important guide to the status of marriage
as to other realms of life, telling us when we need to take corrective action
or reassuring us that we do not. (Of course, these assessments can also be
mistaken.) The next woman, however, describes a somewhat more critical
situation in which she has been trying to convince her husband that she is
not going to leave him for another man - a case in which the relation
between mutual benefit and marital lastingness is far from academic:
Like what I tried to explain to Dan was that one person can't be expected to fulfill
everything because they're not exactly the same. You know, fulfill everything that
one person needs. And that Ron fulfilled something for me that Dan couldn't, you
know. And, it wasn't as much - like Dan fulfills so much for me that I would never
want to leave him for Ron, you know. Because Ron just fulfills this one added little
block that Dan doesn't. I'm not going to leave thirty for one, you know, that's just
- 1 mean, you know - 1 mean, I can't put a number on what he fulfills for me, but
you know, that kind of ratio. [3 W-4]
As does this same interviewee when she says elsewhere, "If it gets static in
our relationship then that's when we'll split," other speakers just as readily
reason to the complementary conclusion - that lack of mutual benefit or
fulfillment will cause a marriage to end:
When a marriage gets to the point where you're really holding down the other
person, you're really restricting them, it's not worth sticking together because life's
too precious to waste your time, with another person. Unless they're really fulfill-
ing you on an emotional level [3H-1]
I'm a firm believer in divorce if things are not going well. Life is too short to spend it
with someone you're not happy with. [7W-6]
If I wasn't happy. If there came a day-to-day thing and neither one of us were
progressing in any way or getting anywhere in the marriage I wouldn't see any
reason for going on with it. But I don't have any particular cause of what would
bring on divorce immediately. [1H-13]
Research on shared task solutions 171
I think it costs me a lot and I don't think he's measuring that cost. And I'm scared
it's going to cost too much and leave me without being able to stay in the
relationship. [4W-12J
Again, this reasoning ranges from the purely hypothetical to the all-too-
real; while the speaker before her seems complacent about the current
status of his marriage, the last woman presents herself as being in the
throes of internal debate about when and whether to end hers.
Next I turn to a set of examples that illustrate the assumption that effort
makes a marriage mutually beneficial:
I think it's a whole lot easier to ride along, each person taking their position and
not moving very much but being - you know, "It's good enough to stay in it."
You've got your kids, you've got your house and that kind of thing. And I think
that the emotional - an emotional relationship that is really meaningful and a
relationship that is sustaining that affirms both people as individuals and allows
those people to be individuals and yet, you actually like being together, I think
that's terribly hard to achieve, that kind of relationship. And I feel like Dave and I
have consciously worked toward doing that, achieving that kind of relationship
and maintaining - 1 think there are periods in our life and in our marriage when it
was very questionable, when I don't think either of us was certain that we wanted
to do that, that we wanted to put the energy and emotional effort into making that
happen. And yet I think it has, so somewhere, some part of each of us must have
felt it was worthwhile enough to struggle through those periods and move on.
This is an interesting passage because it begins with an argument about
how much and what kind of benefits make a marriage last. Some people
stay in a marriage that is just barely good enough, but the true benefits of
marriage go beyond kids and a house, to a meaningful emotional relation-
ship. That kind of real benefit takes effort, as do the benefits described by
the next speakers:
In a sense I think we do work at it. I mean Reynold has to meet some of my needs
and I have to meet some of his needs. And, you know we both get tired of the kinds
of things that we don't like in the other person and yet we work at it, I think.
Because you have the man, I don't think you can stop there and say, "Well, this is
it," you know, "I can just let everything go." Because you ca - you'd be surprised at
the little things that really bug men, I mean that you wouldn't think about it. He
say, "Well you're a lousy housekeeper," you know, and this type of thing. So you
keep working on the things that you think would, you know, enhance - well not
enhance him, but, that would make him happier, you know, as much as possible
without making yourself miserable. [11W-16]
It's not something that comes natural either. It's something that we both work at.
I: Oh, really?
Yeah, that - you know, it's just that our relationship is extremely important to each
172 Practice and possibilities
of us and, you know, we want to work hard at making it so and making it better.
And so, you know, the sharing and giving to each other is part of - you know, a
major part of that. So, you know, we go out of our way to do things for each other.
This last example makes explicit a key motivational assumption that is
more often left implicit, as it is in passages quoted earlier. Wanting a
beneficial marriage motivates people to make an effort to achieve such a
marriage; therefore, they do so.
If mutual benefit makes a marriage last, and it takes effort to achieve
mutual benefit, it follows that effort is required to make a marriage last. All
of the next six passages show speakers making this last connection. Not
surprisingly, given the more general understanding Americans carry into
their model of marriage - of human enterprise as requiring effort in
overcoming difficulty to be successful - lasting marriage is rephrased in
terms of marital success more often in the context of the effort required of
it, we will notice, than in the context of the benefits expected of it.
The assumption that a lasting, successful marriage takes effort informs
the global expectations people bring with them into marriage, as in the
following quotes:
I felt like it [marriage] should be about the same as any other relationship. Any
other good relationship but that you were - you know, this is more permanent, this
is something that you should make work, you know. And not anything that you
should give up on. You know you should always keep trying no matter what.
If you have arguments and if there are problems, you have to say to yourself, "I
came into this with my eyes wide open. It's my responsibility to try to help make it
work." [7W-5]
Contrariwise, the next young husband quarrels with the idea that a lasting
and successful marriage requires more than casual effort:
I guess you can say I really don't think about it, as far as it breaking up or whatever,
you know it just - 1 think it's a waste of time to think about it. I mean why look for
problems? But I think she's always kind of got it like she's got to work harder to
make it work. And I don't really look at it that way. I figure if it's going to work it's
going to work, you know, there's no need going out of your way to do it. I mean if
there's no problems there there's no need to try to make it any better or knock
yourself out trying to do a little more when there's no problem to start with. I just
kind of do it day to day. If a problem or whatever comes up that's when I worry
about it. [1H-S]
Taken at his word, this man is an interesting exception, someone who
rejects the unqualified cultural assumption about the effortfulness of a
successful marriage that is shared by the two interviewees quoted before
Research on shared task solutions 173
him and most other Americans. Yet he not only demonstrates by his
remarks his familiarity with this piece of the cultural model of marriage,
with which he has had to reckon in his own marriage, he also reveals his
awareness that his rejection of it is a minority view and therefore requires
defending to the interviewer. His case provides a glimpse, as well, into the
causes and consequences of such nonsharing. Wed only a year, this man
made clear in his interviews his feeling that he had not been ready for
marriage. His otherwise odd-seeming rejection of the expectation that
marriage takes effort can be understood in terms of his overall resistance to
it, also evidenced in his behaviors.
Not surprisingly, of those I inter-
viewed who later divorced, this couple was the first to do so, suggesting
that rejection of the cultural model of marriage is one certain way to get it
Like the expectations that benefit makes a marriage last and benefit
takes effort, the expectation most people have that effort makes marriages
last and succeed has clear implications for action. Not surprisingly, then,
this understanding plays a part in the conscious decisions people report
making, as well as the less-consciously planned
actions they report taking,
in the course of married life:
[We talked about] Do we want to make this thing stay? You know, living together
for a long time. Because it took some work to do it. [3H-6]
What in the world chance would I have of finding anybody else who would be any
easier to be married to and I wouldn't know that person any better when I got -
married him than I knew Tom.
I: Right, right and that would be the whole thing all over again.
Exactly and never having learned or worked through what actually you need to
learn and work through to make the first marriage stick. [4W-1]
I think all the work and effort we went through in the first, I'd say three years or so
of that marriage making it work was completely subconscious, you know. I don't
think I had any idea - that's why I say I had no idea what I was getting into. I don't
think I had any idea at the time we got into it that it was necessary to work at it.
I should mention that part of the difficulty of finding the pattern
underneath linguistic variation is the necessity, in interpreting passages
like these, of decoding the metaphors speakers use to talk about benefit,
effort, lastingness, success, and so forth, as well as the causal constructions
they use to link these terms. Notice, for instance, the varied metaphors for
lastingness in sentences such as "Do we want to make this thing stay?" and
". . . the same desire to keep things going"; and "Never having learned or
worked through what actually you need to learn and work through to
174 Practice and possibilities
make the first marriage stick"; and the common metaphor for marital
success, discussed in the first half of the chapter and repeated here by
several speakers, as "making it work." Notice also the differing causal
constructions in these passages, ranging from the mere juxtaposition of
clauses - X, Y - to indicate a causal relation between the two in "this is
more permanent, this is something that you should make work"; to a
simple because in the comment, "Do we want to make this thing stay?
Because it took some work to do it"; through such constructions as ifX,
not- Y and the more amplified must do X to make Y happen and need X to
make Yhappen in other statements.
Some readers may enjoy identifying
and deciphering other metaphors and other causal constructions in the
passages that have been presented.
Let me now add the one final passage in which the speaker fills in almost
all the causal links in the entire sequence connecting effort to marital
I guess what I was saying is that since that's [our age difference and my weak self
image and lack of confidence at the time we married\ the way we started out, what
has happened since then in some ways seems to me quite miraculous. That there -
that I have changed so much and that we have changed so much and that we have
been able to work through so many basic struggles in our marriage and be at a
place now where we trust each other, we love each other, we like each other. We
appreciate each other. And feel pretty confident about being able to continue that
way and continue working out any other stuff that comes up. Just seems pretty
amazing to me. It could have gone in so many different directions and that it didn't
is incredible. But I think both of us take a whole lot of credit for the direction it
went in, that we worked ex - really hard. [5W-1]
This passage incorporates further steps in the causal chain linking effort to
a lasting marriage, steps specifying that effort overcomes the difficulty
required to achieve the compatibility necessary for a marriage to last. How
much this woman and her husband have changed can be interpreted as a
reference to their becoming more compatible, as reflected in their now
trusting, loving, liking, and appreciating each other. Just as happiness is
the feeling we expect of those in a fulfilling, beneficial relationship, feelings
of trust, love, liking, and appreciation are ones that arise from, and signal,
a compatible relationship that is working smoothly. The change in the
marriage that brought all this about came only at the cost of substantial
difficulty and effort, put metaphorically as "so many basic struggles" that
had to be "worked through," and were surmounted because, in a variation
on the last metaphor for effort, they "worked at this really hard." Over-
coming this difficulty with this effort enabled the couple to "feel confident
about being able to continue that way," an ongoing journey metaphor for
a lasting marriage.
While these links between compatibility, difficulty,
Research on shared task solutions 175
effort, and lastingness are made explicit in this passage, mutual benefit and
its relation to compatibility and to effort are here left implicit; we are
expected to understand, although the speaker does not say so outright,
that people who have worked hard to overcome their difficulties so that
they trust, love, like and appreciate each other are not only positioned to
fulfill each other's needs, but are doing so. These critical links in the
argument are made explicit in other passages from these interviews. We
have already seen the relation between benefit and effort explicitly stated,
as in the excerpt quoted earlier, in which this same speaker rejects a
good-enough relationship, one "it's a whole lot easier to ride along" in, in
favor of one that is emotionally meaningful and sustaining because of the
work and struggle put into it.
Given space, I could provide examples of speakers arguing not only that
what makes a marriage mutually beneficial is compatibility, but also that
incompatibility causes difficulty, that a marriage must be shared to be
mutually beneficial, and linking the various other pairs of events that enter
into the story of how to make a marriage both lasting and fulfilling. It is
noteworthy that minimal causal chains of the sort I have illustrated most
extensively, in which the reasoner establishes just one causal link between a
pair of events, occur much more commonly in discourse I have examined
than do longer chains of reasoning, like the one with which I ended. This is
probably because speakers ordinarily assume listeners to share knowledge
of the idealized event sequence that surrounds the point they want to
make, comprehending how compatibility is linked to marital success,
effort to marital difficulty, and so forth, without being told. Their audience
shares with speakers the densely-connected schema in which every marital
event and relation is strongly associated with every other. On the other
hand, when speakers do take the trouble to fill in intermediate links in the
causal chain they are tracing, it may be because they judge the real-world
state of affairs to be unusually complex and hard for listeners to follow, or,
as seems to be the case for the last speaker, because they wish to be
especially forceful and persuasive.
How this schema becomes shared
Now we can turn to the question of how this cultural schema for reasoning
about marriage has come to be shared. I have already argued that the sheer
usefulness of some such schema for reasoning, in terms of the cognitive
efficiency I have demonstrated, recommends it for adoption, in the way
any useful idea or artifact would be likely to be adopted. This argument
leaves unsettled how people encounter and learn this schema, however,
and why this particular schema rather than some other is an appealing
176 Practice and possibilities
solution. I will address the first of these questions in this section and the
second in the section to follow.
We encounter this schema, first of all, in other people's talk about
marriage, and learn it from their use of it in such talk. Just as we employ
metaphors to clarify the points we wish to convey, so we reason, not only
to reach our own conclusions, but also to persuade others of our argu-
ments. In the reasoning we do in ordinary talk about marriage and other
subjects we may have more or less of a stake in our ability to persuade, but
even interviewees have some stake in persuading interviewers, if just to
demonstrate that they know what they are talking about.
Just as we
draw upon intersubjectively shared exemplars to do the metaphorical job
of clarifying what we wish to convey, so we must rely on an intersubjective-
ly shared schema for reasoning about marriage if we are to persuade others
of the pertinent conclusions we reach. The cultural schema I have de-
scribed provides such a set of shared assumptions. Just as the repeated use
of appropriate metaphors increasingly familiarizes us with them until they
become wholly conventional, so the repeated use of this schema makes it
seem wholly natural. Violating it, on the other hand, is likely to earn
disbelief, if not outright challenge or simple incomprehension. Consider
the young husband's comment on how to "do" marriage: "I figure if it's
going to work it's going to work, you know, there's no need going out of
your way to do it." As with usages of inappropriate metaphor like the ice
cream metaphor used by the priest, such an argument appears anomalous.
It is no surprise that the interviewee felt he had to go on to defend it, and no
wonder that I (the interviewer) responded to it with incredulity (and, given
the neutrality demanded of my role, a strictly mental eye-rolling).
The idealized story about marriage I have described also has a cultural
life beyond its application to reasoning in ordinary everyday talk about
marriage. Some of the very properties we have been discussing - the
overlearning and the simplification of this schema for marriage - as well as
another property we will next consider - the motivational force it gains
from being about success and failure and efficacy - that give it its appeal
for deciding what to do about one's own marriages and making inferences
about the marriages of people one knows, make it equally appealing to
those who create more public images of marriage. No one should be
surprised, then, if the cultural schema for reasoning about marriage finds
its way into the marriages of TV sit-com couples or those of storybook
princes and princesses, or legal decisions bearing on marriage and domes-
tic partnership, for a few examples. This is one important way a cultural
model "for" becomes a cultural model "of* (Geertz 1973g:93; Kroeber
and Kluckhohn 1952:357). And, of course, these public forms then,
through people's experience of them, play their own role in spreading this
Research on shared task solutions 177
same idealized event sequence and reinforcing it in people's minds, making
it more widely known and more often internalized. Naturally, these public
forms can also distort the original schema in various ways and to different
degrees, as is evident from the ideas about marriage children glean from
the storybook marriages of princes and princesses. And these public
representations can and do change over time. Each in its different way has
an independent opportunity to change, to elaborate, and to distort
people's understandings of marriage, with the possible effect that Ameri-
cans will then come to reason somewhat differently about it.
If, however useful this reasoning device for solution of a given task, the
task itself were infrequently encountered and solving it were of concern to
few, then talk utilizing the schema would be limited to narrow contexts and
would be rare, as would be public forms incorporating it. But this task, like
much other reasoning that people have to do about the perceived di-
lemmas of daily life, is recurrent and widespread: The particular contradic-
tion we are considering between lastingness and fulfillment is confronted
by many people, over and over again. And, caring deeply about the success
of their marriages, people care about resolving this contradiction. So
caring, they also have a lively intellectual interest in marital dilemmas,
even when these affect marriages other than their own, or when they are
posed as general problems concerning marriage. The recurrence of this
issue in people's lives and its real import to them ensure that a satisfactory
cultural resolution of it will be widely used, hence widely talked about in
public as in private, widely overheard, and widely learned.
No matter how many encounter a cultural contradiction or how fre-
quently, and how much they all care about and talk about resolving it -
think, for example, of the contradiction between women's wholesale entry
into the workplace and women's traditional place in the home - it may not
have found a culturally agreed-upon resolution. Such contradictions can
become widespread over very short periods of time, and they may be so
novel and unanticipated that their cultural solution continues to elude
those who experience them. Or, as may be the case with ideas about
femininity, motherhood, and "woman's place" that women working out-
side the home summon up for many, these contradictions may evoke
profound inner conflict that, for a time, stands in the way of their resol-
ution. Cultural contradictions do not have automatic cultural solutions;
however, barring some truly unresolvable psychic conflict, a serious, wide-
spread, longstanding dilemma like the marital one we have been consider-
ing will be solved.
And this dilemma has been around. We can trace the decades-old
history of this contradiction between the lastingness and the fulfillment of
marriage. The older expectation of lasting marriage, predating the import-
178 Practice and possibilities
ation of European marriage practices to America, collided with newer
expectations of marital satisfaction as, in the course of the nineteenth
century, the authority of husbands and the deference of wives declined in
favor of a companionate ideal for marriage. Robert Griswold concludes an
analysis of California divorce cases from 1850 to 1890, saying:
[TJhe trail from the early stirrings of the companionate family in the mid-eight-
eenth century to "no-fault" divorce today is faint, but it can be traced. As marital
expectations and demands steadily rose, so, too, did possibilities for failure; as men
and women expected each other to act in more complex ways, more people fell
short of such expectations. By the late nineteenth century, wives complained to the
courts about cold, aloof husbands, of husbands who did not spend enough time at
home, who failed to check their sexual desires, or who ignored women's emotional
needs. Men countered with complaints of unloving, peevish, quarrelsome wives, of
wives who were poor housekeepers or insensitive mothers. Tensions like these
arose as a logical, even necessary consequence of the companionate family: mar-
riages predicated on emotion required a safety valve when affection wanted, hence
the emergence of the "divorce crisis" that began in the 1880s. (Griswold 1982:174)
By the early part of the twentieth century, says another historian of
divorce (May 1980:156), the pursuit of happiness had taken on a new
urgency and a new definition in terms of personal fulfillment within
marriage. "[DJivorce" in 1920, she observes (ibid.: 162), "did not indicate
a rejection of marriage; rather it reflected the increased personal desires
that matrimony was expected to satisfy, especially for women." In the
course of the late twentieth century, the Victorian-derived understanding
of marriage as an obligation to perform complementary roles has been
largely superseded in legal discourse (Regan 1993:34-67) - even if it has
never disappeared entirely from American understanding - by a new idea
of marriage as a voluntary contract between two individuals. This new
contractual understanding of the marital relationship better accom-
modated, and became intertwined with, a recasting of marital happiness
as complementary need fulfillment (see n. 12), This particular formulation
has been with us since the 1960s. But the contradiction in public discourse
between lastingness and fulfillment, cast in older terms of marital happi-
ness, goes back to the previous century. Why these two expectations
about marriage, along with the expectation that it involve a shared life
together, have been and continue to be so enduring is the topic of the
following chapter.
Why the schema has appeal
What makes this particular schema such an appealing choice for task
solution? In chapter 5 we made the obvious observation that cultural task
Research on shared task solutions ^g
solutions and cultural schemas more generally that have broad appeal will
become more widely shared. We suggested that this particular solution to
the task of reasoning about marriage has such broad appeal, and hence is
widely adopted, because it incorporates a cultural theme - success - itself
having enormous appeal to many Americans. What makes this theme, in
turn, appealing?
It should not surprise us that this familiar American theme has been
incorporated into the solution to this widely shared task. When David
Rumelhart and his coauthors, quoted in the last chapter, consider the way
new solutions to reasoning tasks are invented and conclude "that such
representational systems are not very easy to develop," they then continue:
"Usually they are provided by our culture. Usually they have evolved out
of other simpler such systems and over long periods of time. Newer ones,
when they are developed, usually involve taking an older system and
modifying it to suit new needs** (Rumelhart, Smolensky, McClelland and
Hinton 1986:47). The discussion of thematicity in chapter 5 suggests where
older cultural contributions to such new task solutions are most likely to
originate. We said there that schemas recommend themselves to new
contexts with a family resemblance to the old, and that schemas learned
early and from a wide variety of contexts will be more general and hence
resemble a wider range of new contexts. Success is just such a schema.
Having learned it early and in a variety of different contexts, Americans
regard success as a general way to think about many situations. Thus the
contradiction between attaining a lasting marriage and having a fulfilling
one can be readily reconceptualized as a matter of marital difficulties to be
overcome in order to achieve marital success. Connectionism may have
little to say about how new inventions evolve, but the connectionist
explanation given in chapter 5 for the spread of cultural themes has
something to say about the shape they will take when they do.
Further, general schemas such as this one for success carry the strength
of their appeal with them, lending this appeal to each new schema in which
they become incorporated. Once it has been invented, the subsequent
spread of a schema for reasoning about marriage will owe much to this
broad appeal. That is, the lesson about succeeding is neither an unfamiliar
nor an unwelcome one to newlyweds, who bring with them into marriage a
more general orientation to succeeding from the many earlier contexts in
which they have learned it. Most Americans thus want success, and find it
natural to seek success, in marriage as in so many other endeavors.
It is
so natural for Americans to think of marriage in terms of success and
failure, indeed, that some of them experience divorce as the most acute of
personal failures.
Having said this, it is important to add that not all Americans are
180 Practice and possibilities
equally motivated by economic success. In chapter 8 Claudia Strauss
points out that the working-class Rhode Island men she interviewed pay
lip service to the value of monetary success, and know that others judge
them by it, but do not act on it. Yet, it is interesting that these men not only
use the word "success" but employ the concept, extending their use of it
beyond the domain of strictiy economic success. Thus, in one of the
passages Strauss analyzes (one key to her analysis elsewhere; Strauss 1992)
the interviewee remarks, "I feel so, that with my education, which was only
to the ninth grade, Fve been fairly successful, urn. . . raising my family."
Strauss (1992:203) interprets the especially long pause in this sentence as
reflecting the interviewee's realization "that others might judge him to
have been a failure by the goals of the success model" (ibid.:203). There-
fore he goes on to finish the sentence differently than he originally meant
to, citing his success at something else about which he is secure and in
which he is invested. In chapter 8 Strauss shows that the interests of the
family (the "breadwinner" model) come ahead of the interests of the
individual (the "success" model) in this man's life and those of other
working-class interviewees. My point would be different. However weak
this interviewee's urge to act on his model of monetary success, he knows
the model well enough to transport it into the realm of family, concep-
tualizing that domain in its terms. And he finds it entirely natural to do so.
Furthermore, he cares about being a success, if not in monetary terms,
then in his. It may well be that the theme of success - the ability of the
success schema to define many domains conceptually and also motivation-
ally - is more widely shared than the motivation to be successful economi-
What, then, makes success itself so natural a way of thinking about
many domains to so many Americans? That this should be so is, in fact,
overdetermined. To begin with, the early learning of it that makes success a
durable and motivating schema for many Americans also makes it avail-
able as an American cultural theme. In part II, we illustrated our argu-
ments about how schemas gain thematicity and motivational force with
the instance of self-reliance. Success is closely related to self-reliance in that
the two co-occur in many of the same contexts, including contexts for their
early learning. When small children are encouraged to tie their own shoes,
they are rewarded for trying to do so on their own, and doubly rewarded
for succeeding in the attempt. The same could be said for schoolwork,
team sports, Scouting, camping, and numerous other activities for children
and young adults.
Self-reliance and success do not always or automatically come together
in experience, of course.
However, whether or not success is being
learned in the same contexts that teach self-reliance, both are learned early.
Research on shared task solutions 181
How very early both are learned, and how assiduously taught, by Ameri-
cans was brought home to me on a climb up Hanging Rock, North
Carolina, with my two daughters and my then three-and-a-half-year-old
granddaughter. Of three trails up the rock, we chose the one intermediate
in difficulty, but we would hardly have taken my granddaughter up that
way had we realized, before setting out, how precipitous it was. Near the
top, where the grade became almost vertical, we were assisted by a helpful
stranger who stood above and directed my granddaughter where to put her
feet, while her mother and I climbed behind her in case she slipped. The
gentleman refrained from lending my granddaughter his hand, explaining
to us, "I'm not helping her up because I don't want to take away her sense
of accomplishment."
Like this man, many American parents and other socializes instill a
sense of accomplishment in even very young children because they con-
sider it necessary to success in life, and because they value success so
greatly. Not only do they want to succeed themselves, they want their
children to do so. For this reason, Americans find many and varied
contexts in which to bring home the importance of accomplishment and
the value of success to children and young adults - even on occasions that
might objectively be deemed unsuitable, like a rock-climb potentially
dangerous to a three-year-old. The many differing contexts in which it has
been learned and is daily reinforced, make it not just a ready interpretive
theme but the most obvious available way to think about new difficulties
The same early socialization makes success motivating. Trying to suc-
ceed becomes a valued way to behave in the face of difficulties when
socializes and, later, peers reinforce successful achievements with praise
and meet failures with derision or disappointment or, perhaps more subtly,
by withholding their esteem. The motivation to succeed is enhanced when
the rewards we receive from others for succeeding and the penalties for
failing are internalized, in the way described in chapter 4 in the discussion
of motivational force. In this way, many (though not all) Americans come
to take pride in their successes and feel shame for their failures, making
each new success enjoyable in and of itself, and every failure painful. And
they come, by the processes also considered in chapter 4, to associate
success and failure with the kind of people they are - "a success," as a
person might consider him- or herself, or "a loser." Thus, what can make a
major "failure" such as divorce so very painful is the meaning it has for the
way people think about themselves.
As they find failure upsetting and try to avoid it, so do they enjoy success
and, as adults seek opportunities to succeed. Indeed, US Americans -
especially, perhaps, middle-class ones - turn many experiences into new
182 Practice and possibilities
chances to succeed. Not only does this interpretation of these experiences
seem to them an entirely natural way to confront difficulty, but, to the
extent that in the course of experience they accumulate a record of success,
it comes more and more to seem an efficacious way to overcome difficulty.
American consumers also like to experience success and triumph over
failure vicariously, in their fiction and their fantasy. Purveyors of public
culture respond to the appeal of success by marketing opportunities for it
in games, sports, stories, advertisements, and popular psychologizing.
Thus it does not surprise Americans at all - once again drawing on the
topic of the research described in this chapter as an illustration - when they
encounter articles in women's magazines with titles such as "Why some
marriages fail" (Kavaler 1969), or, in bookstores, what are sometimes
called "marriage success manuals" with titles such as Guide to successful
marriage (Ellis and Harper 1977) or Making marriage work (Campbell
1987; Bishops' Committee for Pastoral Research Staff and Practices 1990);
or textbooks called Making marriage successful (White 1983). These public
forms in which marriage is cast in terms of success or failure are additional
opportunities for more people to learn that it is natural to think of
experience that way.
More than seeming natural, engendering a keen sense of pride and self-
esteem, and warding off an equally unpleasant experience of failure and
sense of self-condemnation, success may be particularly compelling to
many Americans because rewarding in a different way: as a culturally
available reaction formation.
That is, taught not just to value indepen-
dence but to demand it of themselves, Americans can be said to dread
helplessness. In place of this upsetting feeling, they substitute its opposite,
a sense of efficacy. By this interpretation, the anxiety about not being able
to do things that these Americans feel when faced with demanding situ-
ations of all kinds, is refigured into its opposite, the idea that they can
indeed overcome any difficulties in their way by their own efforts. This is
not just a local response to marital difficulty, then, but a broadly character-
istic American approach to life. Steven Marcus (1984:257), remarking on
US Americans' willingness to accord uncommon social esteem, economic
rewards, and authority to medicine and its subdiscipline, psychoanalysis,
refers suggestively to "the poignant and apparently incorrigible American
sentiment that there is nothing finally that cannot be 'fixed up' - including,
it sometimes seems, life and death themselves."
To the list of that which
can be fixed we can add marriage. Its generality raises the possibility that
this reaction formation in response to feelings of helplessness precedes and
infuses the rewards for achievement and accomplishment that come, say,
with learning to dress oneself or doing well in school, being implicated in
the very early psychodynamic events that Jessica Benjamin (1988) de-
Research on shared task solutions 183
scribes, and that we summarized in the context of the motivational force of
self-reliance in chapter 4.
Of course, the generalizability of success to marriage and the consequent
naturalness and motivational force this theme lends marriage are not
unalterable. For example, changed socialization practices, perhaps a les-
sening of the assiduousness with which US American parents and schools
teach their children the value of hard work and accomplishment, could
also change, more or less profoundly, the way successive generations think
and reason about marriage and many other matters. Or, more likely, long
enough experience of high divorce rates, while not making successive
generations any less desirous of lasting marriages, might make divorce
seem more unavoidable and acceptable to them, and less appropriately
understood in terms of individual striving and individual success or failure.
Changes such as these may already have begun to happen to marrying
Americans and to American marriage since the husbands and wives I
interviewed spoke about their marriages almost two decades ago. (The
atemporal language of the last paragraphs should be understood as a
rhetorical simplification.)
Reasoning versus theorizing about marriage
It is important to appreciate that scrutiny of public renderings of marriage
alone can yield quite a distorted view of how people think about marriage
and enact it in their everyday lives. The need to complement the study of
extrapersonal cultural forms with the investigation of evidence for in-
trapersonal cultural understandings like those I have analyzed here is a
point that we have emphasized in this book. The point can be illustrated
for marriage by returning to the case of Habits of the heart.
It is noteworthy that two of the very few actual quotes from their
interviewees that Bellah et al. provide in their chapter on "Love and
marriage" reveal these Californians (from the San Jose area, and inter-
viewed in the same time period as my Durham, North Carolina-area
residents) to be reasoning about the same relation between effort and
marital lastingness that I have illustrated above in the discourse of six of
my interviewees. Here are Bellah et al.'s interviewees:
Melinda Da Silva: When I married him, I said that he was the person, not that I
have to spend forever and ever with, but at least I'm going to try to work things out
with this person, have a family with him, and be a family with him. If we hadn't
been married, I don't know that I would have gone through counseling, marriage
counseling, or couple counseling. (Bellah et al. 1985:102)
Ted Oster: You can't have something as good as a love relationship without
putting a lot of effort into it. It's a wonderful thing, but it's not going to keep going
184 Practice and possibilities
by itself just because it's wonderful. That person is not forever just because you
found that special person. (Bellah et al. 1985:104)
These quotes suggest the generalizability, beyond my own material, of the
model of marriage I have reconstructed from it.
Yet, although it is there
to be found, the effort marriage requires to be successful is not a theme that
the authors of Habits of the heart recover from their own interview
If these authors do not recognize elements of the success schema in their
interviewees' talk about their marriages, it is probably because their analy-
sis ignores the reasoning about marriage that these speakers do in favor of
passages in which the interviewees deliver more abstract, philosophizing,
discourse on the subject. Bellah et al. describe what D'Andrade
(1995:150-181) has called people's cultural theories, which he distinguishes
from their cultural models. In D'Andrade's (ibid.: 151) terms, parallel to my
own usage in this chapter, a model "consists of an interrelated set of
elements which fit together to represent something." "Typically," he goes
on to say, "one uses a model to reason with or calculate from by mentally
manipulating the parts of the model in order to solve some problem." To
the degree that the problem that it is used to solve does not require or
otherwise involve language, a cultural model will remain unstated. A
cultural theory, on the other hand, "consists of an interrelated set of
propositions which describe the nature of some general phenomena"
(D'Andrade ibid.: 172; see also D'Andrade 1987:114), propositions that
can be stated by those who hold this theory. Claudia Strauss (1990: 314)
explains this difference between "explicit theory" and "implicit knowledge
of what goes with what" in terms of the different kinds of experience from
which schemas are learned - from explicit propositions, in the first case, or,
habitus-like, from less-theorized experiences, in the second. Even further,
however one has learned a schema like this one for reasoning about
marriage, once overlearned, it will remain largely out of consciousness, I
speculated. Hence it will be difficult for speakers to articulate as a whole.
As Strauss (ibid.:314-315) goes on to observe, the less theorized knowledge
typical of a cultural model is therefore much more difficult to identify in
speech, being "presented more sketchily, with greater difficulty, and with
more context variability," She rightly cautions that this difference between
what is explicit and what is implicit is only a matter of degree: "For
socialized human beings, of course, experience is always somewhat theor-
ized, just by virtue of being linked to labels" (ibid. :314). Nevertheless, the
distinction has been useful to me in pointing to the fact that I have singled
out discourse in which my interviewees are using a cultural model, while
Bellah et al. appear to have singled out discourse in which their inter-
Research on shared task solutions 18 5
viewees are theorizing. Perhaps they did so because of the difficulty of
identifying less theorized knowledge to which Strauss calls attention.
Here, for example, Bellah et al. summarize the cultural theory that they
call "the therapeutic attitude," and which they find dominates the thinking
of the mainstream US Americans they interviewed:
The therapeutic attitude reinforces the traditional individualism of American
culture, including the concept of utilitarian individuals maximizing their own
interests, but stresses the concept of expressive individuals maximizing their experi-
ence of inner psychic goods. (Bellah et al. ibid.: 104)
For the classic utilitarian individualist, the only valid contract is one based on
negotiation between individuals acting in their own self-interest. For the expressive
individualist, a relationship is created by full sharing of authentic feelings. But both
in hard bargaining over a contract and in the spontaneous sharing of therapeuti-
cally sophisticated lovers, the principle is in basic ways the same. No binding
obligations and no wider social understanding justify a relationship. It exists only
as an expression of the choices of the free selves who make it up. And should it no
longer meet their needs, it must end. (ibid.: 107)
Therapeutic understandings fit many aspects of traditional American individual-
ism, particularly the assumption that social bonds can be firm only if they rest on
the free, self-interested choices of individuals. Thus even Americans who do not
share the quest for self-actualization find the idea of loving in spite of, not because
of, social constraints very appealing, (ibid.: 109)
Here the relatively more "public" - in the sense of being explicitly stated in
propositional form - thoughts of the interviewees are distilled in the really
"public," in the sense of published, discourse of the authors. Indeed, since
the authors' summary of this credo is far more extensive than any state-
ments by their interviewees that they supply, it is never clear how much of
the therapeutic attitude they are said to espouse has actually been articu-
lated by the interviewees themselves, and how much is the authors' theory,
inferred from fragmentary evidence in interviewees' discourse.
In any
event, Bellah et al. go on to claim, by contrast to interviewees' language for
talking about the therapeutic attitude toward relationships they are "with-
out a widely shared language of obligation and commitment" (ibid.: 106) in
these relationships. The authors take this to be evidence for the historical
decline of the latter view and its contemporary defeat by the former.
Because their interviewees cannot enunciate a theory of marriage as an
enduring relationship equivalent to that Bellah et al. summarize for the
therapeutic attitude toward this relationship, they mistakenly conclude
that these Americans have no good model for thinking about marriage as
1 hope to have shown, using my interviewees' reasoning about
marriage as a window into their intrapersonal cultural understandings,
that these people (and, the two excerpts suggest, those interviewed in
186 Practice and possibilities
Bellah et al/s study as well) do have such a model. Their hope and concern
for lasting marriage, far from losing ground to their expectation of marital
fulfillment, is as significant, in this model, as the latter. That marriages
should last is a belief more resilient than Bellah et al. credit it with being.
This expectation that marriage should last, moreover, appears to be
resistant to evidence that increasing numbers of marriages do not. Certainly
the demand for marital happiness, heightening at the beginning of this
century, has meant a greater willingness of Americans finding themselves in
unhappy marriages, to divorce. Certainly, too, new strategies for hedging
against the likelihood of divorce grew apace with its increase (Weitzman
1988:240-246) in the seventies and the first half of the eighties. But the rising
rate of divorce also occasioned, according to a 1987 Newsweek cover story
('How to Stay Married: The Divorce Rate Drops as Couples Try Harder to
Stay Together'), a renewed seriousness about marriage and commitment to
working out marital problems - accompanied by a rash of new advice books
about staying together, formal courses for newlyweds, and therapist-
devised exercises to help couples identify and work through their problems.
Indeed, it is to these developments that the magazine attributes the leveling-
offand slight drop in the divorce rate beginning in the mid-eighties. Despite
prognostications to the contrary by Bellah et al. and more conservative
social commentators, the expectation that marriage should be lasting - as
reflected in the effort Americans are willing to put into making it so
(especially in the face of any trend to the contrary) - is itself an enduring one.
Equally, despite the proselytizing of the New Right against what is deemed
to be Americans' selfish preoccupation with self-realization, the expectation
that marriage be fulfilling is no more likely to disappear.
Not only do Bellah et al. see the traditional obligation to sustain
enduring relationships as giving way, historically, to the therapeutic atti-
tude privileging individual self-actualization; they see the two as "counter-
vailing aspirations" (ibid.: 102) in contemporary relationships. Individuals
are portrayed as being "caught" (ibid.) in the contradiction between the
two aspirations, as "oscillating" (ibid.: 109) between them, as being led to
"paradox" (ibid.: 107), and as being in a situation in which the therapeutic
language "undermines" (ibid.: 106) one of commitment, and in which
"being without a widely shared language of obligation and commitment"
causes "confusions" (ibid.:109) and "difficulty" (ibid.:106, 109), In inter-
views that otherwise bear striking resemblance to the interview excerpts
provided in Habits of the heart, I find no evidence that the married
Americans I heard from experienced any such conflict in the course of their
daily lives.
1 maintain, instead, that they share a cultural solution to the
historical contradiction Bellah et al. identify.
These writers' tendency to
counterpose the two themes of lastingness and fulfillment as conflictful,
Research on shared task solutions 187
like their perception that the former is giving ground to the latter, may
have more to do with their own anxiety about what is happening to their
society, including marital and other family relations in it, than it has to do
with their interviewees' understandings of marriage.
Parenthetically, Bellah et al.'s book has itself become an object of public
culture par excellence - a national bestseller - its discourse on American
marriage and other alleged sites of a growing American individualism
reshaping the understandings of all of us who read it, and who read the
work of other social scientists through whom it is refracted. Arlene Skol-
nick observes:
Speaking to widely shared anxieties about social and cultural change, the book has
played a surprisingly large role in the newly pessimistic discourse about the family
on the part of social scientists. It is not uncommon to find works presenting hard
statistical data on family trends that cite Habits of the Heart as evidence for a
corrosive new individualism that can explain the trends. (Skolnick 1991:203)
Skolnick goes on to remind readers that the analyses in this book are
variations on themes in an ongoing social critique of modernity that
"contrasts a romantic version of the past with a jaundiced view of the
present" (ibid.).
She calls attention to the conservative use to which the
book's argument has been put by at least one scholar, David Popenoe, in
relation to marriage in particular.
If, as Skolnick (ibid.: 202) comments,
"Habits of the Heart presents a far gentler and more complex picture of
Americans" than had best-selling social commentary before it, then Pop-
enoe's Disturbing the nest (1989) redraws a simpler, harsher picture. It
implicates self-realization in the rising divorce rate and the more general
decline of the traditional family with its well-defined gender roles. Family
decline, of course, in a view promulgated by much other conservative
critique, is responsible for a further slew of contemporary social woes.
Thus, two views of what is happening to American marriage compete in
the public arena. One is founded the assumption that individual self-
realization is necessarily destructive of marital commitment. The other is
the more pragmatically and less ideologically motivated understanding of
marital fulfillment and lastingness as reconcilable through hard work.
While for most Americans the second view resonates with the everyday
reasoning about marriage from which it has been taken and exported, the
first view resonates with the deeply felt moral anxieties of some.
We began our section on cultural sharing in chapter 5 by noting that much
of the world is organized in exactly such a way as to ensure that people will
188 Practice and possibilities
have the same experiences. Certainly language in general, and speech in
particular, are prime instances of practices that organize experience in this
way. Language learning and linguistic communication demand virtually
all of the activities that we named in chapter 5 as sources of the modal
patterning of social life: social interaction, knowledge sharing, coordina-
tion, collaboration, role taking, conformity to rules, modeling, and explicit
teaching. My goal in the first half of this chapter was to show how the
shared metaphors that pepper everyday speech arise from a mundane
requirement of knowledge exchange, the need to clarify what we mean for
those to whom we speak. This requirement is met by use of a widely
accepted fund of cultural exemplars that stand for various agreed-upon
features of the world and index those features for the speaker's audience.
In the second half of the chapter, I introduced a different example, one just
as ubiquitous in human experience. Among the shared patterns people
encounter in the world are contradictory ones, raising the potential for
inner conflict among the contradictory expectations (with their attendant
intentions, desires, and goals) that have been learned. If such a learned
contradiction should be significant in people's lives, and one they frequent-
ly encounter and have to reason and talk about, and if it has been
longstanding historically, a mediating structure for its solution - one that
thereby resolves the inner conflict it might otherwise have generated - is
likely to evolve and spread. Borrowing, as task solutions are inclined to do,
from existing cultural themes that recommend themselves just because
these themes are so widely available and so natural seeming, the resulting
mediating structure gains the broad appeal as well as the motivational
force of the theme it borrows. In this way solutions to cultural contradic-
tions become shared solutions and ones motivating to individuals; are
valued and hence transmitted across generations; and in turn, play their
own role in perpetuating and spreading the general cultural themes that
they incorporate.
7 Research on the psychodynamics of shared
Naomi Quinn
The two cases of sharing examined in chapter 6,1 have argued, came about
as solutions to everyday tasks that all of us find it necessary to perform
repeatedly. The kind of shared understanding I turn to next results from a
different order of shared experience. This experience is not one repeated
daily but one experienced early - that is, infantile experience of the sort we
pointed to in chapter 4 as being indelible and unusually motivating be-
cause learned so very early and in the context of the exceptionally strong
feelings, related to their survival and security, aroused in infants. In the
case to be presented in this chapter, as in the cases of becoming self-reliant
and being a good person that were developed in chapter 4, understanding
appears to have become shared because crucial elements of the early
experience that formed it are shared. Of the three kinds of shared under-
standing about marriage I am describing, then, this is the most highly
It should be clear from the last paragraph that in arriving at a reason-
ably full account of how this cluster of understandings about marriage has
come to be shared and motivating, I found myself drawing on
psychodynamic as well as cognitive theory. Some readers may find the
resulting juxtaposition of two such distinct theoretical traditions jarring.
My combining them, however, is a considered move. It is an actualization
of the research stance Claudia Strauss and I set out at the beginning of part
II: our assumption that in order to adequately explain the complexities of
cultural understanding and meaning we will have to draw on multiple
theoretical paradigms. I hope to foster a spirit of inquiry (see also Hutchins
1987; Nuckolls 1996; Paul 1990), in psychological anthropology, in which
the mutual usefulness of two such paradigms, from cognitive anthropol-
ogy and psychoanalytic anthropology, can be explored.
Analysis 3. The psychodynamic basis of marital love
To introduce the next level of analysis, I must pick up the thread of the
story about marriage that weaves this chapter together with the last. There
190 Practice and possibilities
I detailed an event sequence that follows from a contradiction between the
expectations that marriage be lasting and that it be mutually beneficial in
the sense of fulfilling. I treated these two expectations, along with a third
that marriage be shared, as givens. A clue to the genesis of this triad of as
yet unexamined ideas about marriage is their distinctive expectational
status, which sets them apart from the other expectations about marriage
that follow from them. Like the woman who commented, "I guess I really
don't deep down in my heart think that probably most people's marriages
are any easier than mine is," realistic people expect their marriages to be
difficult. Difficulty, with the risk of failure that it entails, is an inevitability
that, like this last speaker, they accept with resignation. Other expectations
about marriage - that it will require compatibility, and effort to attain that
compatibility - are means to the ends of overcoming marital difficulties
that stand in the way of marital fulfillment and lastingness. In contrast to
these more realistic and pragmatic ones, the expectations that marriage be
lasting, shared, and fulfilling are desired ideals.
I will argue that the ideals of marital sharedness, lastingness, and
fulfillment form a complex schema rooted in the early experience of
contemporary US Americans. To anticipate, my argument will be that
married love, and hence marriage as Americans know it, is a "refinding" in
just the sense that Freud (1962:88) viewed adult love as a refinding of the
early love relationship. The expectations Americans have that marriage is
to be shared, lasting, and fulfilling match their understandings that people
who love each other should be together, stay with each other always, and
fill each other's needs. These ideas about love, in turn, revert to the infant's
earliest anxieties about being one with the caretaker, not being abandoned,
and being cared for. It is from this psychodynamic complex that these
expectations gain their motivational force as well as their durability for
individuals. And it is because the early experience resulting in this
psychodynamic complex has been substantially shared, as are later experi-
ences further shaping their ideas about adult love and linking it to mar-
riage, that Americans share an understanding of marriage in terms of love.
I will begin by showing how, and how closely, Americans' ideas of
marriage are structured by their ideas of love. This evidence will be crucial
to the argument to follow about American marriage as a realization of
adult love and that love itself as a refinding, supporting my claim that this
extension of love to marriage has an intrapsychic reality. The demonstra-
tion of how love structures marriage will also lay the groundwork for a
fuller consideration of how this schema for marital love might have
become shared.
In addition, the evidence to be presented counters an opposing interpre-
tation that has been given for the association of love with marriage in US
Research on the psychodynamics of shared understandings 191
Americans' way of thinking about it, an interpretation which argues that
marriage structures love. In this latter explanation, causality is the reverse
of that I posit. Offered by Steve Derne (1994), who attributes it to an
unpublished manuscript by Ann Swidler, it says that the American ideal of
love as exclusive, all-or-nothing, and enduring, "is compelling because the
institution of marriage requires that Americans choose only one person to
be married to" (ibid.:285) at a time, that they do so decisively, and that
they expect the resulting marriage to last. That is, these ideas about love
are plausible to Americans and they "continually return to this view
because it fits with the strategies that they build around the structural
reality of marriage" (ibid.:284). The psychodynamic explanation I will
offer recommends itself over this structural one on two grounds.
First, the structural interpretation Derne gives hinges on three isolated
features of love, in Americans' ideal picture of it: its lastingness, its
exclusivity, and its all-or-nothing quality. The psychodynamic explanation
draws upon the fuller array of features - individually quite striking and
taken together very distinctive - of the love experience, making sense of all
these features of love and, in turn, marriage, in terms of infantile experi-
ence. For example, although Derne (1994) calls attention to the American
idea of love as a feeling for a unique, special person who is the single
"right" person, that monogamous marriage requires exclusive love does
not really explain these ideas about the specialness or lightness of the one
who is loved. That this person is experienced as a refinding of an idealized
infant caretaker, as psychoanalytic thinkers have proposed, does make
sense of this experience. Nor is the structural explanation able to account
for another experience, "falling in love," which can be understood
psychodynamically as a reentry into the dependent infant's felt state of
extreme helplessness. On the other hand, the extremity of this dependency
and this helplessness, conveyed in the precipitousness of "falling" in love,
might be thought to contribute to the all-or-none quality of love to which
Derne points. Below I will interpret other key features of marriage and
marital love in terms of the infant-caretaker relationship.
The structural interpretation, which accords beliefs about love the status
of strategic responses to persistent structural dilemmas (Derne, 1994:285),
fails to explain why these "strategies" should be so emotionally laden and
motivationally compelling - as are, for example, falling in love and idealiz-
ation of the loved one. In the psychodynamic explanation, love and hence
marriage are permeated by the motivation for "refinding" a lost state of
blissful infantile dependency and by the heightened emotions surrounding
survival and security that also attend that early experience. I would not
want to argue, of course, that US marriage practices play no role in
reinforcing ideas, motivations, and emotions Americans have concerning
192 Practice and possibilities
love. These practices, after all, are just such public forms as, we have
argued in this book, reinforce our inner understandings. More, to the
degree that such practices are institutionalized, these institutions help to
prolong the historical life of such understandings. But public practices and
institutions alone can hardly account for either the particular shape or the
emotional and motivational intensity of culturally shared ideas about
marital love.
The analysis undertaken in this chapter subjects the same corpus of
discourse to the same search for pattern as the analyses of metaphor and
reasoning described in the last chapter. The focus of this analysis, however,
is the word "love" which, it goes without saying to US American readers,
labels a significant aspect of the marriage experience, surfacing in inter-
viewees' talk about marriage over and over again.
Hence, like the meta-
phors for marriage that draw on cultural exemplars for their meaning and
the culturally idealized schema for marriage that facilitates reasoning
about it, this culturally agreed-upon label for a salient emotion in the
context of marriage suggests itself as an obvious analytic clue to shared
understandings about marriage.
I will demonstrate two kinds of detailed correspondences between love
and marriage, in Americans' understandings of these. The evidence from
which I recovered these correspondences comes, once again, from reams of
talk by interviewees, this time about marital love. Again, I can reproduce
here just a portion of this evidence and use these examples to provide only
a broad outline of these speakers* understandings of marital love. The
details omitted, which elaborate that story in interesting ways, do not
change it. Even my sample evidence may be too extensive for some readers
who, once again, are invited to skip quoted interview passages when they
have read enough of them.
The first kind of correspondence is an alignment in Americans' under-
standing between the emotional state of love and the social status of
marriage. In the second set of correspondences, these ideas about love fill
in the motivational structure we attribute to marriage. These ideas them-
selves will not surprise American readers and may seem all too obvious to
them: love, after all, has been defined by countless scholars in various
academic literatures, and rendered repeatedly in popular culture. Never-
theless, the analysis of my interview material has the advantage of showing
that these are the ideas of ordinary, contemporary US Americans. Fur-
thermore, a systematic analysis like this one avoids focusing on one or a
few features of love to which undue theoretical emphasis is then attached,
while overlooking others, equally integral to Americans' understanding of
it. Finally, and key, this description of how people see love as motivating
the way they feel, think, and act will enable us to appreciate the close
Research on the psyche-dynamics of shared understandings 193
correspondence of this motivational structure to the triad of ideals these
same people hold for marriage, independently recovered from their meta-
phors for it and presented at the beginning of the last chapter. I will then be
in a position to return to the issues of how a cluster of motivations deriving
ultimately from infantile experience comes to be embodied as it is in
Americans' shared understandings of adult love, and through the latter, in
their cultural expectations of marriage.
Love and marriage
To US Americans, love and marriage go together. The object of this
section is to show how systematically and with what precision these two
conditions are aligned in American thinking. This demonstration should
denaturalize and make remarkable American readers' normally taken-for-
granted assumption that love and marriage go together, calling for its
When I say that marriage is aligned with love, I mean that being in this
social status is contingent on and temporally coterminous with this emo-
tional state in Americans' expectations of how things go: that is, falling in
love marks the beginning of marriage, being married means being in love,
and falling out of love signals the end of marriage. In my interpretation,
this contingency of marriage on love and the boundary shared by the two
are crucial evidence that marriage, for Americans, has historically taken
shape as the institutional realization of love.
Let us look at the evidence
for this alignment one point at a time.
1. If things go the way they are supposed to, Americans believe, two
people fall in love and get married. Most of the time, the assumption that
people marry for love is taken so for granted that it usually only emerges in
the course of narratives having other points, like the following: "Though
the Catholic church says you can't marry a divorced man in the church, I
really hadn't reconciled that, I wasn't going to be worried about it. That if I
found a man that was divorced and I loved him I would say, 'God
understood,' and marry him." [9W-4] This assumption may emerge, as
well, on the rare occasions on which it is explicitly contested:
I'd certainly hate to see people - everybody in the world jumping into marriage
immature. But maybe there's really something to the idea that it is not so much how
you feel - exactly how you feel before you get into a marriage but what you can
make of life, you know, in the marriage that really counts. Having thought it out -
in other words having thought it out beforehand and coming to the conclusion that
you really are in love, might not be as good for a marriage as having gotten
married, looked into what was worthy of being in love about, found it, identified it
after awhile - because it does - it takes awhile - and then made that the corner-
stone. [6H-4]
194 Practice and possibilities
While arguing that love should follow marriage, this husband reveals (by
arguing against it) the assumption he attributes to most people, that love
should precede marriage. Discrepant stories, too, in which people marry
people they do not love, or, as in the next case, people who love each other
do not &nd up marrying, warrant telling, and in this telling also expose the
expectation that has been violated:
I had a friend - this is when we were in college - and she was very much in love with
a guy she had been going with in high school and they were both good friends of
mine and she dropped out of college to go live with him and they loved each other
very much but were absolutely incompatible. And they didn't know that before and
it was a wonderful thing that they found out and they're still good friends and now
she's married and he has another girlfriend but it was something that came as a
shock to them particularly and to me as a very interested onlooker that you can
love someone and yet still be incompatible with them and that love doesn't conquer
all. [7W-6]
Like this woman, people recognize, and often caution, that love alone
cannot hold a bad relationship or a troubled marriage together. In so
cautioning they reveal their otherwise implicit assumption that love, while
not sufficient, is a necessary condition of marriage.
2. Once married, two people will or should continue to love each other
(though that love can certainly grow and change in quality). Again, this
expectation that spouses are supposed to be in love with each other is
nowhere revealed more sharply (and, in the next case, painfully) than when
it fails:
But in a sense I always feel like I'm cheating Tom, you know. That he has a right to
have a wife who has the sort of emotional commitment to him that many couples
seem to have, and one part of me says, "Well you probably do have at least as much
emotional commitment to him as m - almost anybody would have but you don't
have the mindless 'I'm in love' kind of losing yourself, pu - surrendering your
judgment and perspective and stuff like that." And some part of me must keep
saying to me, "That's the way you're supposed to be, in love, that's why you're
supposed to be married, that's the ideal," you know, "That's what you owe to your
husband." I know some part of me is saying that and another part of me is saying,
"Well," you know, "that's not what I've got and I just - 1 am what I am." [4W-12]
If they do continue to love each other, people want and try to stay married.
Thus, this normally implicit assumption that married people love each
other surfaces again in difficult or insecure times when the continuation of
a marriage is in question. This can be seen in one husband's affirmation
that, even though he and his wife took some time apart at one difficult
juncture in their marriage, "I loved her, I never stopped loving her, and I
don't think she ever stopped loving me" [3H-1]. Therefore, he concludes,
they resolved to come back together and work at their relationship.
Similarly, other interviewees report, "I knew I didn't want to leave her [my
Research on the psychodynamics of shared understandings 195
wife], you know. I knew that I still loved her" [2H-8] or "And I always say
to him [my husband], 'I would never leave you/ Because I really love him
too much" [6W-6].
Under this assumption that people who love each other stay married, a
case like the next one, in which people divorce, even though one does
continue to love the other, becomes anomalous - and hence worthy of
narration: "My brother and his wife were divorced in August. He's still not
real sure. He's thinking about remarrying someone else but he can't decide
because he really still loves his ex-wife" [1W-5].
3. Conversely, interviewees make clear their assumption that when two
people don't love each other anymore, the two do not stay married:
If you ever split up it's not going to be done because you like each other and you're
still in love and you're still happy, you know, with the thing. If the split would ever
come it'd be at a time when you really didn't like each other. [9W-1]
But she could say that, if we were going to stay together and that's what she wanted
- that she wasn't sure she loved me anymore - the only way would be to stay and
work it out, because to get apart would not answer anything,
I: She did say she wasn't sure she loved you anymore?
Yeah, yeah. At one point. Which was very painful. I mean that was about when I
just wanted to forget the whole thing.
I: So what she was saying was. . .
. . . contradictory. [4H-7]
Note that the structural explanation Derne (1994:285) offers, that Ameri-
cans believe love should endure because marriage is supposed to be lasting,
does not account very well for the widespread expectation, expressed by
these interviewees, of marriage ending because love has.
My interviewees worried perennially whether their spouses still loved
them. This worry is understandable, once again, given the assumption that
falling out of love spells the end of marriage, and it provides another
window onto that ordinarily tacit assumption. One woman whose hus-
band tried to tell her about his existential crisis reacted this way:
He said, "I don't even know what happy is anymore." He said, "I'm just here. I'm
just living day to day." And it kind of made me feel bad because I felt like, "God,
does that mean that he doesn't love me anymore? Does that mean he doesn't want
to be married or what?" [1W-7]
And another woman, during a long illness, queried her husband:
"Would you still - are you going to love me forever?" And I think it was really a
surprise to him that I was worried so much more about what he was going to think
about it [than about my deteriorating health itself]. [9W-1]
Other interviewees describe finding themselves especially unable to take
their spouses' love for granted, as they find it necessary to interrogate and
196 Practice and possibilities
affirm their own love, during periods of marital discord. Because of this
anxiety, these married Americans put emphasis on whether partners show
their love, to reassure them that it still exists. Purely symbolic expressions
of love, such as gifts or words, are not deemed to be enough, or may even
be distrusted - while small undramatic acts of caring are especially valued,
as is mutually fulfilling sex, as authentic and hence believable expressions
of love.
While there may be different reasons why one spouse stops loving the
other and seeks divorce, one scenario is described repeatedly by my inter-
viewees: falling in love with another person. Though one may love many
people, one can only be in love with one person at a time (see also Dern
1994). Therefore marriage, too, is necessarily exclusive. As one interviewee
spells out:
Well I guess it goes to what I believe in marriage. I just feel like when you're ready
to marry somebody then you're ready to give up everybody else as far as ever going
out with anybody else - any other men. I mean you're ready to just dedicate your
life to loving one man, you know. I mean I can love somebody else as a friend but
not romantically, you know, physically, romantically. [1W-2]
Because of the exclusivity of "romantic" love, falling in love with someone
new means that one stops loving the person one has loved up until that
time. Since people marry the one they love, this means, as well, divorcing
the old love and marrying the new one - marriage, once again, aligning
with love:
As far as divorce would go yeah, you know, if she came and told me she was in love
with another man and wanted to marry him, sure fine. [1H-13]
Certainly [if he started having an affair and\ if he developed, you know, emotional
attachment or if he loved the other party then I would think that he should leave
and go with the other party. [9W-8]
Predictably, this assumption about divorce and remarriage becomes es-
pecially problematic for couples like the next one who are attempting open
marriages. The wife, the woman quoted in the previous chapter as saying
she would never want to leave her husband for another man she was seeing
because she would never leave "thirty for one/' has trouble explaining this
to her husband:
It [my husband's attitude) was almost like, "Well if you love him it's got to be the
kind of thing where you're going to want to leave me and be with him." But it
wasn't that at all, you know. And I still have a hard time explaining in terms - in
words how I felt. [3W-4]
Beside the determined optimism that most interviewees express about
working out all kinds of difficulties in their marriages to make them last,
the fatalism they express and the obsessive anxiety some of them verbalize
Research on the psychodynamics of shared understandings 197
about the possibility of being left for another person is noticeable. Because
these Americans believe that sex is thought to lead to the kind of emotional
involvement that, in turn, can develop into love, the two are linked in their
minds - sex, as we have seen, being regarded as a sign of love. Concern
about losing one's spouse to another person - or about facing a decision to
leave one's spouse for another person - centers therefore on the question of
extramarital sex. Some interviewees think that a person who loves his or
her spouse very much may be less susceptible to falling in love with
someone else even in the face of sexual temptation or actual involvement:
I didn't have any emotion for the people that I was dealing with because I was still
in love with my woman - my baby. And so it didn't prove anything to me. I didn't -
I mean I felt I got a chance to feel like I was a desirable person. And then in time,
the thing weren't very successful because I really didn't - 1 was more or less doing
something. I wasn't - it wasn't like I really wanted these people, 'cause I didn't
want them. But I wanted to feel like somebody wanted me. [2H-8]
This man's adventures notwithstanding, the people interviewed in my
study are in broad strategic agreement that the surest way to prevent
falling in love with someone else is to forego extramarital affairs. They
differ only on tactics: Some, like the woman who said "when you're ready
to marry somebody then you're ready to give up everybody else," hold that
any extramarital sexual liaison violates the marriage vows, while others
like the next woman distinguish, whether as a matter of pragmatism or of
denial, between casual affairs and those that are likely to end a marriage
because they are emotionally involving:
There may be things that are reasons for other people splitting that I just don't even
think about. That I - 1 wouldn't want them to happen so I don't think about them.
I: You don't think about...?
Well the biggie that comes to mind is other relationships. I don't think that would
really devastate me. Knowing that Tim and I both, before we were married - before
we were dating each other, we both had numerous other people that we cared a lot
about, I think that I'm more liberal in that. If he were really in love with someone
else then it would be - 1 think I would - it would be different. But if it were just
physical and if he were just having an affair with someone and it was - that would
not really - that wouldn't do it. No. I wouldn't [inaudible].
I: You say you think you're more liberal. Do you mean in general or. . .
Than - than I think - Well, perhaps I'm more - well not liberal, but perhaps more
tolerant of physical affairs than other people. Perhaps who have not had them. And
it's something that I - 1 don't know, I just - it would hurt but it wouldn't kill me
and I think that it wouldn't really permanently affect our marriage. Unless Tim fell
in love with someone else, [beginning to laugh and adopting an ironic tone] and that
might affect our marriage somewhat.
I: An outside party that was really serious.
Yeah. If there were emotional attachments it would be harder for me to go on.
198 Practice and possibilities
To summarize: not only the expectations that people fall in love and get
married, that married people love each other, and that people divorce
when they fall out of love, but various further details in which these
expectations are spelled out, evidence the alignment of marriage with love.
Circumstances that violate this alignment, such as people falling in love
but not getting married, staying married when they are not in love, or
continuing to love someone they have divorced, earn reportability. That
people are supposed to love the person they marry is an assumption that
interviewees reveal even as they caution that love may not be enough. That
people are supposed to stay married as long as they remain in love explains
why interviewees are prone to consult their own feelings to verify whether
they still love a spouse, in making decisions whether or not to leave them.
Equally, this expectation explains why these interviewees look to authentic
acts of love on the spouse's part as proof that the spouse still loves them
and hence will not leave them; and why they view a spouse's extramarital
entanglements as potential evidence to the contrary, that, having fallen in
love with another person, the spouse will necessarily fall out of love with
them, and will leave them.
The motivational structure of love
Love is a feeling, as reflected in the remarks already quoted, "It is not so
much exactly how you feel before you get into a marriage," and "I didn't
have any emotion for the people that I was dealing with because I was still
in love with my woman," and countless other statements interviewees
make about it.
In the American cultural model of the mind that
D'Andrade (1987:120-124) has outlined, no less than in psychologists'
theories, like that of Westen, described in chapter 4, feelings give rise to
desires - which, in turn, engender intentions, which lead to actions. We
have seen, for example, interviewers assuming that if people are happily
married they will want their marriages to last and that, wanting them to do
so, they will try to make them last. Love is no exception to this relation, in
American understanding, between feeling and motivation. The moti-
vations engendered by love are clearly identified and well articulated by
interviewees: in the following passages and others, they say that they do
not wish to leave, or lose, the person they love; that they want to be with
that person; and, finally, that they want to fulfill that person's needs. It is
this constellation of motivations that matches so strikingly a triad of core
expectations about marriage - evidence for which was brought in the last
chapter. What follows is evidence for the parallel motivational structure of
1. If you love someone you don't want to lose them. As with the two
interviewees quoted as saying, "And I always say to him, *I would never
Research on the psychodynamics of shared understandings 199
leave you.' Because I really love him too much," and "I knew I didn't want
to leave her, you know. I knew that I still loved her," this means that you
are strongly disinclined to leave the one you love. More anxiety-provoking
is the possibility that the one you love may decide to leave you, as in the
case of the husband who reported, "Because I love Beth, and I was a little
bit torn up and I was worried the relationship was in trouble, the marriage
maybe was in trouble," and the next woman who, like him, was negotiat-
ing an open marriage:
And I felt terribly anxious about Bob, terribly anxious about losing him. And he
still persisted pretty much in the same thing that, in his feelings about Terry, which
was loving her and caring for her but not romantically loving her and not being
able to respond to the hopes and dreams or wishes that she had. And yet I didn't
trust that. I was real afraid that wasn't true and that any time it would click and
then I would be, you know, out. [5W-2]
It is not surprising that open marriage arrangements, which violate the
expectation that love be exclusive, should provoke anxiety about losing the
person you love. When that person experiences love toward another, their
spouse is left to worry whether, as we have already seen it is prone to do,
that other love will eclipse the love for the spouse. In the first of these two
open marriages, the wife (quoted earlier) found no amount of explaining
likely to dissuade her spouse of the belief that "If you love him it's got to be
the kind of thing where you're going to want to leave me and be with him."
Extra-marital relationships, even those that have the prior sanction of
both spouses, thus become occasions on which not only the expectation of
losing, but also the desire not to lose the person you love becomes explicit.
2. You want to be with the person you love, as still other interview
excerpts make clear:
And I've even slept with people since I have slept with Bobby when we first started
dating. When I was at school and again I would think the whole time, "I wish I was
with Bobby, I wish I was with Bobby." So that's some of what convinced me I
really must love him if I kept wishing I was with him instead of somebody else. That
was physically attractive but yet. . . [1W-6]
You're not in love all the time but there's some times when you really get into the
feeling of it. And it happens sometimes singularly and individually and sometimes
both of us at the same time. When we really feel good about each other and we
spend a lot of time talking and we just want to be with each other. [6H-4]
I: What is falling in love like?
It's hard to say - 1 would - 1 can't...
I: Put it into words?
No. There's an emotional feeling to really care for a person and you do things for
them and help them and make life easier for them. Wanting to spend time with
them even if you're not talking or interacting, you know - reading, studying,
computer work, whatever. [9H-2]
200 Practice and possibilities
3. Finally, as this last excerpt and the next ones illustrate, if you love
someone, you care about them, and therefore you want to do things for
them - make them happy, help them, give to them, share with them, fulfill
their needs, and if necessary, sacrifice for them:
Mostly if I want something very much or if I just even express it, Rich will do it. I
mean, people have said, "It's obvious that he loves you very much." I mean Rich
really loves me very much. And he's very good to me. [9W-5]
I like to do things for him that please him or - 1 don't know, again, how do you
explain love? Except that you just - you care for somebody and that you want to do
things for them. You know that you want to make them happy. I just enjoy his
being, his person. [1W-3]
I: And what do you mean by love?
Ah. Essentially the - well I think the sharing, the togetherness, the giving. Ah - and
emotional attachment, caring, that kind of thing. [2H-1]
But now I'd say that love is - to me, is the desire to give more to the other person
than you're giving to yourself, at times. You know it's the - not only willingness to
accept the part of marriage where you have to change and adjust, but a strong
desire to do so. Not only the willingness to accept that there are times when the
other part - person - partner might need some help. But the desire to give help. To
that person. The sort of love of the fulfilling the other person's needs. [6H-2]
These and other interviewees go on to explain that this desire to help
extends to a willingness to see the other person through difficulties: "If you
love a person you stick by them, for better or worse" [3H-16], one man
says. It includes various kinds of emotional support. This can mean not
harming the other person psychologically, particularly when they are
vulnerable: explains another man, the love of a spouse is "never used
against you" [6H-2]. And it can mean accepting the other person: "If you
love someone enough," a woman asserts, "you should accept them as they
are" [2W-1], The desire to fill the other person's needs is reciprocal: each
person in a love relationship cares for and wants to help the other,
and, as
at least one interviewee describes it, the idea that someone loves you can
enhance your desire to behave lovingly toward that person, making love
mutually reinforcing: "Love," he says, is like "a little extra sparkle of
energy" that "makes you want to do things" [2H-2] to please the other.
If the alignment of marriage with love, described in the last section, is
one kind of evidence that the latter has its institutional realization in the
former, then the second kind of evidence that love is realized in marriage is
the match between the motivational structure of love, described in this
section, and the expectational structure of marriage, described in chapter
6. The three desires that the loved one not leave the person who loves them,
be with that person, and fulfill the person's needs, illustrated above, match
Research on the psychodynamics of shared understandings 201
the expectations that marriage be lasting, shared and fulfilling, respective-
ly, and, I argue, are the derivation of these expectations.
Love and early experience
Following psychoanalytic theory, I have suggested that the motivational
structure of love, in turn, relates to matching infantile concerns about not
being abandoned by the caretaker, being one with the caretaker, and being
cared for by that person. It is from infantile experience, then, that love
acquires its motivational force, including its power to shape our under-
standing of marriage. This being so, the case of love exemplifies, not only
the incorporation of emotion and motivation into a cultural schema, as
addressed in chapter 4, but the primary role that may be played by these
emotions and motivations in the formation of such a schema.
Supporting this interpretation of marriage and married love in terms of
infantile love are the reports of some interviewees, corroborated by general
knowledge, that husbands and wives not uncommonly talk baby talk to
each other, and look to each other for "mothering" or "fathering." Indi-
viduals did not seem to be particularly self-conscious about these practices.
This is probably because they were unaware of the infantile look of these
practices, as obvious as it seemed to me. If they do not experience baby-
talk and being "mothered" as infant-like, one reason is that the infantile
complex at the root of adult love has been overlaid and elaborated with a
great deal of subsequent experience and resulting understandings, includ-
ing those about marital love and other adult love relationships. Another
reason is that coming to see oneself as a competent adult - a valued and
rewarded status in all societies - requires one to defend against infantile
wishes and behaviors of all kinds. Further evidence that early experience is
reinvoked in adult love relationships and marriage in particular comes
from clinical observations. Walter Gadlin (1988), for example, describes
how marital relationships can become the sites for the expression of certain
transference themes such as feelings of oneness and, conversely, fear of
being overwhelmed; of neediness and the desire for unconditional caring;
and of demands for exclusive attention and anxiety about being aban-
While Sigmund Freud was the first to argue the relation between infan-
tile experience and the form of adult love, subsequent theorists in various
psychoanalytic traditions, who place its origin in the pre-oedipal experi-
ence of the infant, have enriched Freud's original description of this
experience substantially. Not wanting to get sidetracked here by predict-
able theoretical disagreements of emphasis and detail, I offer instead the
following composite version of these writers' accounts of the origins and
202 Practice and possibilities
development of adult love, a version on which I think most could agree,
although each might want to embellish it differently: in earliest experience,
infant and mother are one. The mother cares for the infant exclusively and
meets all of the infant's needs. However blissful is this stage for the infant,
the course of normal development requires it to end, for, in order to
become individuated, the infant must separate from the mother. This
separation is experienced as loss. As one psychoanalyst concludes this
story, "Depending on the solidity of this achievement, the search for love
more or less seeks to undo this state of separateness. It also seeks to restore
to the self the sense of perfection believed to have been enjoyed in the
original mother-infant unit" (Ireland 1988:25).
By their nature the issues of separation, loss, and care that are central to
this preoedipal experience impinge upon the infant's psychological wellbe-
ing and its very sense of safety and anxiety about survival. If we are to
accept the psychoanalytic description of it, this infantile experience is of a
kind that is certainly early enough and arouses feelings strong enough to
result, in turn (as may be seen in our discussions of durability and moti-
vation in chapter 4) in an indelible, motivationally charged schema. More-
over, the connectionist account we have put forward in this book provides
an explanation for how this durable, motivating schema comes to be
reinstantiated, in its entirety, in adulthood. In this schema-theoretic ac-
count, adult intimate relations that activate one component of the early
schema, whether feelings of closeness, or fears of separation, or neediness,
say, or particular memories of the caretaker - will activate the entire
complex of ideas, feelings, and motivations (although, as we have seen,
they are overlaid with the results of subsequent experience that is now part
of this schema, and that disguises its infantile basis).
To summarize and fill in my argument: an interpretation of interviewees'
expectations regarding marriage and their understandings of love in terms
of early infantile experience, as psychoanalytic theory characterizes it, fits
the details of these interviewees' descriptions exceptionally well. Specifi-
cally, in adulthood, infantile closeness with the caretaker is expressed as
wanting to be with the person one loves and, correspondingly, expecting to
share one's life with the person one marries. The issue of individuation is
echoed in adult concerns about not being submerged by love and preserv-
ing one's autonomy in the marital relationship (about which my inter-
viewees talked extensively), and it is telling that psychoanalytic thinkers
generally regard this balancing of dependency with autonomy to mark the
attainment of mature love. Infantile anxiety about separation and loss of
the caretaker is expressed as wanting to be loved forever and, in turn,
expecting marriage to be permanent. Infantile concern about the endur-
ingness of maternal love resurfaces in adulthood, when everlasting love
Research on the psychodynamics of shared understandings 203
and the permanence of marriage are threatened by the possibility that the
adult loved one may leave, with divorce as a consequence. Particularly
preoccupying, due to the infantile fear of being replaced and the wish for
the exclusive love of the caretaker, is adult anxiety about losing the person
one loves to another. The infant's desire for the caretaker to fulfill all of its
needs is expressed as the desire for the person one loves in adulthood to fill
these needs, and the corollary expectation that marriage be fulfilling. These
needs now include sensuality and sex, which, in the psychoanalytic picture
of mature adult love, are joined with caring. There is also the proviso that,
now, one must reciprocate, simultaneously being dependent and caretaker.
Psychoanalytic theorists regard this ability to take both needful and need
fulfilling roles in a relationship as a hallmark of mature love, and indeed, it
is for them perhaps the very measure of adulthood in our society.
Although psychoanalysts are broadly agreed on the relation between the
two, these and other students of the subject are not agreed upon a theoreti-
cal account of how the infantile experience of Americans like these inter-
viewees translates into their adult understandings of love (cf. Gaylin and
Person 1988:x-xii; Ireland 1988:28; Jankowiak 1995a: 13; Lasky and Silver-
man 1988:12). Again, although they have now assembled compelling
evidence that something such as we experience as adult love is widespread
in human societies, anthropologists have not yet produced a theoretical
account, on which all can agree, of what about this experience and the
understandings that it engenders might be culturally distinctive and what
might be cross-culturally shared.
It is possible, on the one hand, that the
experiential gestalt I have described is universal in human societies, in
which case the explanation I have given, of an early and powerfully
motivating schema reactivated by adult experience, stands on its own.
That is, when children everywhere grow up and become capable of adult
intimate relationships, these trigger the experience of the infant-caretaker
relationship; everywhere, too, the unavoidable loss of that first significant
relationship lends the later one a quality of "refinding."
It is possible, on the other hand, that the psychodynamics underlying
adult love, while universal in their occurrence, are quite variable, cross-
culturally, in the way they unfold, and hence in the version of adult love
that they engender. Specifically, the emphasis on refinding the lost infant-
caretaker relationship in adult love, assumed by Freud to be a human
universal, might end differently elsewhere:
this early relationship might,
for example, be dampened or altogether repressed, or split into compo-
nents reexperienced separately in separate contexts. It might be renounced
- as Stanley Kurtz (1992) has proposed that Hindu Indian children are led
to renounce the original mother-infant relationship in favor of connection
to the larger kin group. It might be retained as an unrealizable ideal, as
204 Practice and possibilities
Holly Mathews (personal communication) suggests Oaxacan Mexican
men do, or sublimated in the role of caretaker to one's own children,
as Mathews believes Oaxacan women do.
Or - as Anne Allison
(1994:111-112 and personal communication) suggests for the case of
middle class Japanese boys tied to "education mothers" who prepare them
for their school examinations - it might never be given up in the first place.
This second possibility - that the story of refinding is indeed culturally
distinctive in its shape - would require an explanation of the circumstances
under which this story has emerged. In the terms set out in chapter 5, this is
a question about the shared learning (at least partially shared, as well, with
Freud's Viennese patients and other Europeans, past and present, as well
as earlier US Americans) that has accompanied and followed on the
infant's earliest experience with its caretaker, to shape a distinctive under-
standing of adult love. I will offer one such possible explanation - even
though I regard it as premature speculation.
Explanations of the distinctive character of love in the United States
have tended to focus on the idealization of the loved one and the love
relationship, sometimes tracing this idealization to European origins. One
suggestion (Endelman 1989:112) has been that US Americans put an
exceptionally high value on providing untraumatic parenting (even though
they do not always succeed at doing so).
Another idea has been that
American mothers create an attachment between themselves and their
infants of exceptional affective intensity (Chodorow 1978:86; Kurtz
1992:264), fostered, in turn, by the isolation of nuclear households and an
emphasis, by contemporary American childrearing experts, on maternal
responsibility (Chodorow 1978:212). While either of these conditions
might lead to idealization of the infant-caretaker experience and the care-
taker herself, it is unclear to me how such childrearing and the idealization
that it ostensibly produces can alone account for the idea of refinding an
ideal that had been lost. In addition, there is mounting evidence that
idealization of the loved one is hardly distinctive of American or Western
understandings of love (see Jankowiak 1995).
The admittedly speculat-
ive account to follow builds on these observations that US American
infant care is especially nurturant - at least ideally so - and the relationship
between infant and mother especially intense. My account draws attention
to an additional widespread characteristic of infantile experience, a feature
that may be somewhat more distinctive of US American childrearing.
Even as they are being loved, American infants begin to learn that they
will lose that perfect love. For examples, many Americans wean their
infants from breast or bottle early, by world standards, and, even more
unusual cross-culturally, require them to sleep in separate bedrooms. US
American parents interviewed in a recent study of cultural variation in
Research on the psychodynamics of shared understandings 205
their infants* sleeping arrangements (Morelli et al. 1992) explain this
practice, as we might expect them to (and as the West German parents
described chapter 5 also do), in terms of the value of teaching infants
independence. (The researchers [ibid.:611] suggest that the practice may
have the paradoxical effect of making it more difficult, rather than less so,
for these children to separate from their parents during the day.) Training
in self-reliance is not only early for American children, but also unrelent-
ing. If an overriding goal of American infant care is to provide the infant
with a sense of being loved, a secondary theme that emerges as the
paramount goal after infancy is to teach the child to be self-reliant. In
other words, US Americans are extremely nurturant to babies (even as
they are requiring them to sleep apart and beginning to pressure them to
become self-sufficient in other ways) but they are not very indulgent of
older children acting "babyish." What I propose is that the extreme
demand for self-reliance is experienced as a sharp departure from the
earlier experience of total nurturance. I further surmise that to the degree
that independence is adult-imposed before children come to it themselves,
the child may be more likely to experience the earlier unconditionally
nurturant relationship with the mother as something they have been
deprived of, and hence as something lost, rather than something voluntari-
ly relinquished.
This is one possible explanation. In placing such explanatory weight on
the psychodynamics of the earliest relationship and its unfolding in subse-
quent experience, I do not wish to unduly neglect or underplay the role of
public culture in fashioning, refashioning, and then reiterating a distinctive
cultural story out of these psychodynamic materials. It should be clear that
I do not think that this curious, yet powerful, set of beliefs about love is a
fabrication of public culture alone. Nevertheless, the role of popular song
and like media in shaping adolescent ideas, in particular, about love and
marriage should not be underestimated (Schlachet and Waxenberg
1988:53). Not only does such public culture shape these beliefs and expec-
tations, but it also plays an important role in broadcasting them. If
southern French love poetry enshrined the story of courtly love and
troubadours spread it throughout twelfth-century Europe, the twentieth-
century version of this story is encoded in popular song and circulated, not
just in live concerts but by electronic means, throughout the world. And we
have seen, public culture such as this preserves, even as it may reconfigure,
these beliefs and expectations over relatively long periods of time (as
Schlachet and Waxenberg 1988 document, love has been reconfigured in
popular song even since the 1920s).
Left unaddressed in my discussion so far is why adult love, whatever its
source, should come to be realized, in Americans* understanding of it,
206 Practice and possibilities
within marriage. I think this is much the easier part of the US American
pattern to explain. As the most intimate of institutionalized adult relation-
ships, and hence most likely to trigger infantile experience, monogamous
marriage is a predictable site, perhaps even a magnet, for the realization of
that experience (and herein lies the grain of truth in Derne's [1994] argu-
ment, which I disputed at the beginning of this chapter). Social practices
that allow or promote such emotional intimacy between husband and wife
can be expected to encourage "love matches" and the expectation of love
in marriage.
Social developments in US history would seem to have favored marital
love. Karen Lystra (1989:28) tells us that American middle-class youth
were selecting their own partners by at least 1800, with little interference
from parents, and that "'the heart' played an increasingly larger role in
mating as the century progressed, so that "by 1830, romantic love was fast
becoming the necessary condition for marriage in the American middle
class." Once married, spouses were also expected to continue loving each
other until death - divorce not being a legitimate alternative (Lystra,
1978:194,221). Lystra attributes this shift to love matches to the rise of the
ideal of companionate marriage, which challenged parental control over
mate choice; she argues (ibid.: 158) that "parents bowed out, not just
because the family became less of an economically productive unit in an
industrializing economy, but because acceptance of the ideas and values of
love and the self gave them no basis to act upon - except as advisers and
manipulators of the pool of eligibles." But I would give more weight to the
weakening of parental authority over children, because of the increasing
importance to these children's economic fortunes of education and other
intrinsic assets, over property and other inherited wealth. Though the rise
of the "romantic self* may have been responsible for the swiftness and
breadth of their acceptance among the American middle class in the
nineteenth century, love matches themselves were not new, and so cannot
be satisfactorily explained as the product of new ideas alone. John Gillis
(1985), for example, examining British marriage from 1600 to the mid-
1980s, has argued that at different periods throughout these centuries, love
matches existed among the propertyless, side by side with arranged mar-
riages among the propertied, and that the two exhibited quite distinct
patterns of fluctuation in popularity over time. As Gillis' observation
suggests, social practices that mitigate against spousal intimacy - say,
arranged marriages between strangers
for the transmission of property
or other economic or political ends, or markedly hierarchical gender
relations that distance husband and wife - make marriage a less likely site
for love. Then adult love must take some other route of expression, or
remain unexpressed.
Research on the psychodynamics of shared understandings 207
Whatever the experiences that lead us to want to refind the infant-
caretaker relationship in adulthood and realize this adult love in marriage,
and however these experiences may be the same or different historically or
cross-culturally, it is important to remember that given experiences must
have been shared in broad outline by substantial numbers of people in
order to have resulted in a cultural schema for married love such as that
Americans share. The framework developed in chapter 4 to think more
generally about how cultural schemas are learned has been helpful in
thinking about how this one might have come to be shared. This overly
brief consideration of how infantile love may relate to adult experience
should suggest some of the cross-cultural, subcultural and/or historical
differences in infantile and later experience on which future research needs
to focus. This discussion should also underscore the value, and indeed the
necessity, of a working relationship between cognitive and psychoanalytic
The languages of love
Finally, whatever history of love and marriage we favor, we should not
underestimate the historical continuity, with older ideas, of the complex of
understandings about these matters that I have described - understandings
that certainly predate Freud's Vienna. The tendency to regard our own
version of love and marriage as wholly contemporary is abetted by the
strikingly different idioms in which past understandings of them may have
been couched. Such idioms can disguise equally striking similarity, how-
ever, as the example of the nineteenth century suffices to illustrate.
Nineteenth-century American love was cast in terms of religious experi-
ence, true selfhood, and gender complementarity (Lystra 1989; Seidman
1991). Falling in love was a desexualized spiritual longing. True love, the
foundation of a lasting relationship, necessitated each person's knowledge
of the other's spiritual and moral qualities. In order to assure this exchange
of knowledge, courtship became a period of intense disclosure of the
spiritual and moral strengths and failures of each; it is then, rather than
after marriage, that the relationship was "tested" (Lystra ibid.:157-191)
for commitment and compatibility, in successive rounds of anxious self-
criticism and reassurance. Disclosure also promoted an intense experience
of union, expressed in metaphors of a holy union joined by God, of being
united, one being or one flesh, of being blended into one another, mingled,
or merged, and of being halves of a more perfect whole or, when apart,
"the absent half of our own very identity" (Seidman ibid.:45). Union was
underlain by powerfully felt spiritual attraction, and shared spiritual-relig-
ious aspirations - rather than shared social interests or activities, which
208 Practice and possibilities
would have been implausible in the context of separate men's and women's
"spheres" (Seidman ibid.:50). Again, the mutual benefit of marriage was
reciprocal emotional support in attaining these religious aspirations by
overcoming spiritual anxieties and filling spiritual lacks - rather than
reciprocal fulfillment of psychological needs. That mutual support re-
quired spiritual complementarity, which was much helped by the prevail-
ing theory of gender difference: each sex was understood to be governed by
its own masculine or feminine principle and drawn to the opposite sex
because each was incomplete and complementary to the other (Seidman
ibid.:53). Lastingness was inherent in the expectation that marriage was
for life, and was supported by the widespread idea of marriage as a
religious sacrament, with a transcendent moral purpose (Regan 1993:31);
by the doctrine of true love - which, unlike mere romantic love, was
enduring because based in knowledge of intimate motives and the true self;
by the bonding that came of mutual self-disclosure; and by the attraction
and meshing of opposites. On the surface quite different from contempor-
ary ideas and even odd by comparison, nineteenth-century American
understandings of love prove compellingly parallel. Underneath the un-
familiar metaphorical and other language that first draws our attention,
the Victorians expressed the same expectations that marriage be shared,
mutually beneficial, and lasting.
It would be easy to miss this similarity if we attended exclusively or
overly to the linguistic conventions in which love was wrapped in these two
different eras. Of all the extrapersonal renderings of cultural meaning,
language offers the fullest clues to this meaning. Like other such public
expressions of cultural meaning, however, it is indirect. As I tried to
demonstrate as well in the cases of metaphor and reasoning in the previous
chapter, discourse of all kinds must be carefully and systematically mined
to recover as much as possible of the interpersonal meanings the speaker
has in mind. These include not only its more obvious social meanings, but
its deeper psychodynamic meanings as well.
The theoretical framework developed in this book to explain the emerg-
ence of culturally shared beliefs and expectations accommodates widely
different empirical cases. This is because the processes by which sharing
arises are various. Sharing, we have said, requires only that different
people encounter essentially the same pattern of associations in the world.
In this chapter, I considered yet a different case - one in which the impact
of a constellation of early infantile experiences endures to recur in adult life
and powerfully motivate adult behavior. While the match between infant
Research on the psychodynamics of shared understandings 209
experience and adult understandings is persuasive, far from settled is what
learning takes place in between infancy and adulthood to augment earliest
experience in giving these adult understandings their particular shape. The
conditions of infant life and later experience under which this configur-
ation is learned must be widely shared to engender the culturally shared
understandings that result. To the degree that the early experience on
which this schema depends (as, in my hypothesis, unsparing nurturance
followed by early and extreme demands for the child's independence) is
provided by consecutive generations of caregivers, the schema will have
historical resilience as well. Once established, too, this understanding is
made even more indelible - even as it may be being reshaped - by public
forms, from popular song to psychotherapy, that may celebrate, elaborate,
and refigure the story.
The three different kinds of cultural sharing illustrated by my research
certainly do not exhaust all the ways in which such sharing arises. They do,
however, exemplify three significant ways. An important lesson to be
derived from this string of three analyses of Americans' shared under-
standing of marriage, is that the shared experiences that contribute to
cultural understandings may be of very different kinds. Hence these experi-
ences can have very different motivational resonances and durational
prospects. While conventional metaphors that have become habituated
and have the advantage of being widely understood may live for a long
time, they can also drop away unnoticed, as new ones can catch on readily.
Very little emotion or motivation attaches to our use of metaphor. By
contrast, expectations about love and marriage are freighted with emotion
and motivation. Therefore, the cultural story about love and marriage
appears to have enormous resilience in the individual life course. Core
elements of this story - those tied to a powerful set of motivations that has
been stable over time - also appear to exhibit considerable historical
durability. It is through the ethnography of the inner life, as advocated
here, that we can begin to identify and explain such differences in the
motivational force and the durability of cultural understandings.
8 Research on cultural discontinuities
Claudia Strauss
When I was an undergraduate, I wanted to understand where people's
beliefs come from, and I devised an interdisciplinary major with the
title, "World Views and Forces Which Shape Them," to study
this. As a graduate student in anthropology I was taught that beliefs are
culturally constructed. The phrase "culturally constructed," however,
glosses over a complex process in which beliefs of very different sorts are
learned from practices and discourses of very different sorts. Some of these
practices and discourses are mutually reinforcing in their messages, others
contradictory. Some leave us with fully formulated ideas, while others
leave us something much vaguer (Sperber 1985b). Some are explicitly
pedagogical, others inform only implicitly, and still others are deliberately
obfuscating (Kluckhohn 1941,1943; LeVine 1984). Some are emotionally
fraught, others inform but do not engage us (Spiro 1984, 1987a). Finally,
some present unrealizable ideals, others illustrate what really happens, still
others suggest bygone or emerging possibilities (Williams 1977). Meta-
phors of culture as "structure," "text," or "discourse" did not do justice to
this complexity. Eventually I became suspicious of theories that flattened
culture and reduced it to a single kind of thing and was attracted instead to
discussions of culture acquisition that took the perspective of learners
exposed to a variety of practices and discourses. Such challenges to notions
of culture as a tidy system were around before the current poststructuralist
and postmodern critique of the culture concept; among anthropologists,
for example, there was the insightful work of Barth (1975,1983), Sperber
(1975, 1985a, 1985b), and Wallace (1970). There were also sociologies of
knowledge that analyzed the contradictions between consciousness as a
product of a dominant ideology versus consciousness as a product of lived
experience (Bourdieu 1977; Bloch 1985; Comaroff and Comaroff 1991;
Gramsci 1971; Marx and Engels 1970).
This chapter illustrates three aspects of cultural complexity that affect
the learning of beliefs and that I have explored in my research: conflicting
messages, disparate forms and processes of internalization, and different
dynamics at the social and psychological levels. In the first section, I will
Research on cultural discontinuities 211
discuss the ways in which people internalize inconsistent messages and
relate this to my growing realization that any discussion of fragmentari-
ness and inconsistency has to be balanced by discussion of competing
forces for integration, which operate at both the social and psychological
levels, although in different ways at each. In the second section, I show the
dissimilar motivational effects of schemas that are learned and internalized
in different ways. Cultural representations vary not only in content, but
also in form, and these variations in form seem to have distinct conse-
quences for the way people act on their beliefs. In that section I also spell
out some methodological consequences of these disparate forms of inter-
nalization. The way understandings are acquired and cognitively represen-
ted has consequences for the way they come to be expressed - not only in
one-on-one talk of the sort I have analyzed, but also in mass media and
other sorts of public texts. In the last section, I spend more time on
differences between extrapersonal and intrapersonal culture. Public cul-
ture goes through a more complicated process of production than does
people's spontaneous talk. The result is the need to be very careful when
drawing conclusions about cultural meanings from public culture, as I
show in the case of a public policy change that is generally thought to have
widespread support in the United States but is in fact only a partial
realization of public sentiment.
These topics are illustrated with examples from my past and current
research on Americans' ideas about economic individualism. As a social
description, economic individualism posits a society of free individuals
(persons or corporations) interacting to maximize their economic gain.
Normatively, it is the idea that a person's economic standing should
depend on their own efforts. It is thus opposed (in theory) both to systems
in which economic standing is given by birth and those in which this is
determined by governmental intervention, although government interven-
tion to assure greater equality of opportunity is consistent with economic
individualism (McClosky and Zaller 1984:94-5). In the case of corpor-
ations, economic individualism is usually taken to be the same as laissez-
faire capitalism (the fewer the restrictions on the economic decisions of
corporations, the better), with the possible exception, again, of interven-
tion to break up monopolies in order to ensure opportunities for smaller
businesses. Economic individualism is probably the dominant economic
ideology in the United States and helps account for the odd fact (from a
cross-national perspective) that average US Americans tend not to band
together to fight economic and political policies that favor richer people at
their expense. (The relative quiescience of US American workers, com-
pared to those in other industrialized countries, was labeled "American
exceptionalism" at one time; see Mackenzie 1973. This tendency may not
212 Practice and possibilities
be so exceptional any more; see Halle 1984.) When I was planning my
dissertation research, I noticed that most discussions of US Americans'
ideas on this subject had been based on survey research, which is helpful
for impressions of large groups but unable to delve into attitudes in any
depth or with nuance. Even researchers who had conducted open-ended
interviews (e.g., Halle 1984; Hochschild 1981; Lane 1962; Sennett and
Cobb 1972, to name works that I read during that period) did not analyze
their interviews in the careful way I had admired in Quinn (1982).
So in my dissertation research (1984-5) I used the controversy surround-
ing a large chemical plant located on the edges of the neighborhood where
I then lived in Cranston, Rhode Island, as a way of exploring how
economic individualist ideologies were internalized by neighbors and em-
ployees of the plant. That factory, owned by the US subsidiary of the Swiss
chemical company, Ciba-Geigy Ltd, had been the subject of protests
because of its smelly and possibly health-threatening emissions into the air
and water. Just before I settled on my research topic, the company an-
nounced plans to close its Rhode Island plant and shift operations else-
where in the United States, forcing several hundred employees to look for
new jobs. Earlier in its history, I discovered, the plant had been the site of
serious labor-management conflict. After a year of background research
into the history of the Ciba-Geigy controversy, I conducted in-depth
interviews with its neighbors and employees to learn what they thought
about the "free enterprise system" in light of their experience with this
company. Over the course of in-depth interviews (six open-ended inter-
views per person, each approximately an hour-and-a-half long) with fifteen
men and women who were employees or neighbors of the chemical plant,
I asked about the conflicting rights of the neighbors, employees, govern-
ment (local, state, and federal), and the corporation owning the chemical
plant. Over the course of repeated visits we branched out to many other
topics, including interviewees' life histories and ideas about who does and
does not get ahead in this society. In 19901 conducted follow-up interviews
with each person, asking some of the same questions about getting ahead
to see if their ideas had changed. This research is the major source for my
discussion, in the first section below, of how inconsistency is managed
socially and psychologically, and in the second section below, of how
diverse social inputs give rise to differing forms of motivation.
In 19941 began a research project on attitudes about the welfare system.
This topic interested me both because it was the focus of major debates in
public policy and because it is linked to so many other sensitive topics in
this country, including ideas about individual achievement, federal gov-
ernment programs, class, race, gender, and immigration. The welfare study
has combined a variety of methods. In 1994 my students and I led several
Research on cultural discontinuities 213
focus-group discussions on the topic. Early in 1995 I conducted a phone
survey with randomly chosen samples in the greater Providence, Rhode
Island, area and the Research Triangle area of North Carolina to assess
how attitudes about welfare are influenced by other questions in the
interview context (an experiment I will describe further below). Since
mid-19951 have been interviewing or supervising interviews with some of
these survey respondents.
Each interviewee has participated in two
lengthy (approximately hour-and-a-half) sessions, discussing a variety of
topics related to welfare, other current events, and their own lives. All
along I have also been collecting examples of public discourse on welfare -
including on-line discussions, talk-radio commentary, newspaper articles,
and political party platforms. This research is the basis for my discussion
in the third section of an example of only partial congruence between
public policy and popular sentiments.
Analysis 1. How are conflicting discourses internalized?
There are several social and cognitive mechanisms for handling discrepant
ideas. At the social level, authorities can marginalize some ideas (as has
happened to pagan ideas in Judeo-Christian societies, for example) -
persecuting their holders, labeling them wacky, or simply ignoring them -
while validating and disseminating competing ideas. Another possibility is
that a compromise might take hold, a socially approved way of selectively
synthesizing the conflicting ideas. Naomi Quinn gives an example of this in
chapter 6: traditional views of marriage as a sacred vow, competing with
modern views of marriage as a voluntary contract, are resolved through a
typically US American emphasis on making an effort to overcome difficul-
ties. (To a certain extent, too, compromises have been worked out with
pagan practices and Christianity, as we all know from the examples of
Easter eggs and Christmas trees.) If either of these (marginalization or
socially approved synthesis) happens, the presence of conflicting idea
systems should not lead to psychological inconsistency because (in the first
case) the marginal ideas are either ruled out or can only be chosen while
rejecting the dominant view or (in the second case) a way of resolving the
potential conflict is readily available.
Marginalization and socially approved synthesis are not always present,
however, leaving people in every society to internalize some conflicting
ideas. (See, for a good example from another society, Abu-Lughod 1986
on the sentiments propounded in public, daily discourse versus those
expressed in the poetry shared with intimates among North African Bed-
ouins.) When that happens, there are several ways these conflicting dis-
courses might be internalized. First, a person could choose one and reject
214 Practice and possibilities
the rest. Or they could (unconsciously) select parts of the competing public
discourses and integrate them in a single schema. Another possibility is an
unconscious compromise. In this case the competing ideas are internalized
in separate but dynamically linked schemas so that acting on one creates
some anxiety or need to compensate by later acting on the other, but the
person is not explicitly aware of this psychic conflict most of the time. Still
another possibility is ambivalence, which is like unconscious compromise
except that no workable compromise has emerged and the person feels
A final possibility is compartmentalization (Singer 1972; Weiss
1990). In this case the competing ideas are internalized in separate, uncon-
nected schemas, so that expressions of one are unrelated to expressions of
the other. These schemas are activated in different contexts and the person
feels neither conscious nor unconscious conflict between them in the
ordinary course of events.
Under the heading of compartmentalization,
there are two further possibilities: horizontal compartmentalization, in
which the schemas are equally accessible to consciousness and verbal
expression, and vertical compartmentalization, in which one schema is
more accessible and easily verbalized than the other. (The metaphor of
verticality is not meant to imply that the less accessible schema is repressed
by the more accessible one; in both horizontal and vertical compartmental-
ization the two schemas are unconnected in the person's neural network.)
In a previous publication (Strauss 1990) I described and gave examples
of three of these possibilities: horizontal compartmentalization, vertical
compartmentalization, and integration. (I called the first two "horizontal
containment" and "vertical containment.") The point I was trying to make
was that just because people have internalized conflicting discourses, it
does not follow that their beliefs are a confused jumble; instead, they could
be highly organized (in one of those three ways, the only ones that had
occurred to me at the time). Since I illustrated each with the discourse of a
different interviewee, however, a reader could draw the conclusion that
some people have completely compartmentalized beliefs. Furthermore,
since I gave two examples of compartmentalization (one horizontal, one
vertical) and only one of integration, a reader might also draw the con-
clusion that integration is less common than compartmentalization. The
picture I presented there could thus be taken to support some poststruc-
turalists' theories that selves are shifting and plural (see Flax 1990 and
Kondo 1990 for summaries) as well as other theorists' claims that post-
modern selves are fragmented (e.g., Jameson 1984,1991a, b). (The differ-
ence - to use the distinction formulated by Jameson 1984:63 - is that the
poststructuralists are likely to agree that selves have never been unified
while the theorists of postmodernism see psychological fragmentation as a
new phenomenon.) I do not agree with either of these positions, however,
Research on cultural discontinuities 215
if by lack of a unified self or fragmentation of self we mean discordant
1 think that everyone's belief system is partly compartmentalized
and partly integrated and that this has probably always and everywhere
been the case. Furthermore, the forces responsible for social inconsistency
and integration are different from the forces responsible for psychological
inconsistency and integration, so we cannot infer the latter from the
First I will illustrate compartmentalization, then partial integration,
from my research. I will end this section with a discussion of the respects in
which I believe Jameson's discussion of postmodern psychological frag-
mentation is wrong and what might explain the sort of fragmentation and
partial integration I found. The following is adapted from the longer
discussions in Strauss (1997).
Since cognitive schemas are not readily observable, what do I take as
evidence for their compartmentalization or integration? I have paid atten-
tion to three clues: the content of people's talk, the voices they use to express
those contents, and temporal continuity or discontinuity in the expression of
those contents (Strauss 1992b). By "content" I mean simply the ideas
expressed. By looking at content we can discern that someone is expressing
inconsistent ideas, but this alone does not tell us whether the speaker has
integrated or compartmentalized the inconsistent ideas or is ambivalent
about them. By "voice" I mean characteristic mode of expression, including
repeated key words, metaphorical imagery, and emotional valence (cf.
Bakhtin 1981). Ideas that seem inconsistent to the observer but are ex-
pressed by the speaker in a single voice are probably integrated in the
speaker's cognitive networks. The reverse does not necessarily hold, how-
ever: Inconsistent ideas that are expressed in different voices are not
necessarily cognitively compartmentalized. For example sometimes people
copy their metaphorical imagery, key terms, and so on from distinct public
discourses while internalizing the ideas expressed in closely connected
schemas. To discern this I have paid particular attention to the third clue:
whether the ideas are expressed in connected or disconnected discourse
contexts. If someone typically expresses discrepant ideas both in distinct
voices and separate discourse contexts, I feel confident that their ideas are
compartmentalized. If different voices are intermingled in a single discourse
context, it suggests at least partial integration of the underlying schemas.
Evidence for compartmentalization
In my 1990 paper the person whose discourse I chose to illustrate horizon-
tal compartmentalization was Jim Lovett, a neighbor of the Ciba-Geigy
factory and former welder forced out of work by an occupational disabil-
216 Practice and possibilities
ity. He was in his late fifties when I interviewed him. Over the course of
many hours of conversation, Lovett expressed three schemas that seemed
inconsistent with one another.
These schemas I have labeled CAN' T
SYSTEM schema was shaped by populist social discourses and focuses on
the exploitation of working people like him by large corporations, big
government, the rich, and the poor. In this schema, rich people are bad.
His ACHIEVING ANYTHING YOU WANT schema derives from individ-
ualist social discourses; in it, anyone can become rich in the United States
and becoming a millionaire is a worthy goal. Finally, his FBELING RE-
SPONSIBLE FOR OTHERS schema is a version of communitarian dis-
courses, that value caring for other people, in one's own family or in a
larger community. Becoming rich, in this discourse, is a lower priority.
CAN' T FIGHT THE SYSTEM. When I asked Lovett to talk about Ciba-
Geigy, his former employers, and current events, he typically replied with
his version of a populist outlook that takes the perspective of the '"little
person" or "'average person" who is oppressed by the rich, big business,
and politicians and taken advantage of by the poor (Kazin 1994). (Note:
"left populists" criticize big business and the rich and "right populists"
criticize big government and the poor, Boyte and Riessman 1986, Reich
1987. Many US Americans are probably like Lovett in being critical of all
of these.) His version of this populist discourse can be called CAN' T
FIGHT THE SYSTEM because in these contexts he repeatedly expressed his
feelings of powerlessness to effect change. Here are some examples:
[fjt's just that the system, it's just so hard on the little guy, that's all. It's been the
workingman, it's been the workingman that has supported the country, the work-
ingman that has supported the world. The poor guy, he's on welfare; he doesn't pay
no taxes. And the rich man, right up until today, he's got some money he hides it;
he don't even pay taxes either. So who's it leave? The working guy, that is struggling
to support a family, to keep a home, to have enough groceries on the table, to send
his kids to school. He just don't have enough. He's struggling all the time. And he's
carrying the whole world. [85,1:17-18]
Corporations do not care about people. They - all they care about is satisfying their
stockholders, making money. They don't care about the little person, that's just
another pebble in the road, if it gets in your way, you kick it aside, [del. 1 line]
Somebody starts complaining about it, then they go on someplace else and they
start all over again. The same thing, [del. 4 lines] In this world they could go round
and round, fifty, a hundred years from now, they could come right back to Rhode
Island and start all over again. It's just an endless thing to - they don't care. [del. 4
lines] You can't fight that. You can't fight that. [85,1:11]
Politicians. They get away with murder. And they, you - everybody could be up in
arms about it, but until you can get a group, no one's listening. [85,1:9]
Research on cultural discontinuities 217
[speaking of high utility costs] Why should the little homeowner have to subsidize
the big corporations? It's not right, it's not fair. Again, what're you going to do
about it? I'm only one person. So what? They'll say, "Yes, sir, you're right. Yup.
You're right. We'll look into it, we're going to take care of it." Click, the phone off.
The numbers after the quotes refer to the year the interviews were conduc-
ted (1985 for each of the preceding quotes); which interview in the series of
six (the first, for each of these quotes); and the page number where the
section quoted appeared in my transcripts.
As the page numbers show, all
of these quotes came during the first interview, three of them close to-
Not only were these ideas typically connected in discourse, but they were
also expressed in the same voice. Notice that in all of them society is
imagined in terms of opposing classes or forces or interest groups (the
"little guy" or "workingman" or "little homeowner" versus "the poor
guy," the "rich man," "corporations," and "politicians"), with the agency
of people of the first sort constrained by a "system" that favors people of
the second sort. Lovett's temporal horizon and sense of change over time is
consistent with this focus on structural constraint: as he describes in the
second quote, in the long run the larger forces will prevail: "It's just an
endless thing." At another point he took an even longer perspective,
suggesting that over the course of millennia there is improvement until an
eventual "burnout point." Civilization starts over, "Until a burnout point
again. I, I think that it's endless" [85,3:17-18]. The emotional tone domi-
nating this talk is one of anger, resentment, and despair.
Much of Lovett's content and tone in such passages is typical of popu-
list language that I heard from other interviewees as well. But the voice in
which Lovett expressed these ideas differed a bit from that used by other
interviewees. For example, Lovett typically used images of trying to speak
and not being heard to express powerlessness. Note that in third and
fourth passages quoted earlier he said, "until you can get a group, no
one's listening"; and "They'll say, 'Yes, sir, you're right. Yup. You're
right. We'll look into it, we're going to take care of it.' Click, the phone
off." The lengthy discussion excerpted in the first passage closed with the
comment, "I think that it may relate to a lot of things that people are
thinking, but never have the opportunity to say"; and his discussion of
repeated civilizational burnouts began, "I don't think that we listen
enough to people of knowledge." By contrast, another working-class
interviewee, Carol Russo, typically used violent imagery ("crush," "step
on," "keep us back and down," "climbing all over everybody") to de-
scribe the way little people (whom she usually imagined as women and
children) are oppressed by people in power. That is why I stress that
21 $ Practice and possibilities
Lovett's CAN' T FIGHT THE SYSTEM schema was his version of popu-
ACHIEVING ANYTHING YOU WANT. Surprisingly, given Lovett's
CAN' T FIGHT THE SYSTEM discourse, when I asked him in the fourth
interview what he thought of the "free enterprise system," he responded
enthusiastically with his version of an economic individualist discourse in
which society is composed not of classes but of individuals, there are no
structural limits on economic mobility, and making big money is good:
I think it's terrific. I, myself, Irene [his wife] and I have been for a number of years
Amway distributors, [del. 3 lines] And we could build this business as much or to
any height that we'd want to, even to what they call direct distributorship, [del. 21
lines] Now this was free enterprise. We were given the opportunity to take some-
thing from the beginning and grow and go with it as far as we wanted to go.
[Continuing discussion of Amway] So it's quite a, they are quite a business today.
They started with a single product in a garage and [laughs] I've already lost track of
what they've grown to. You know. It's tremendous, [del. 6 lines] They're all over
the world now. They're all over the world. They're in Japan and China and
England and Germany, France, Canada. Just about every country. Australia, [del,
II lines] [gives the example of another Rhode Island Amway distributor who won a
new car as a bonus] This is available to anyone in the business. You know. There's
no, there's no restriction, there's no line about, well, I belong to the - you can't
reach this level. No, there's nothing like that. You can go from zero to the top. [del.
13 lines] Them men [the ones who started Amway] are, I don't know how much
they're worth, [del. 5 lines] So . . . it's just a, just proves that with an idea, that
there's no limit to any level that anyone in this country couldn't achieve.
There's a, we used to have a little saying that if your mind can conceive it, then you
can achieve it. It's like, you know, whatever your mind can conceive, you can
achieve. But you only get out of anything what you're willing to put into it. And if
you're going to lazy around, then you're going to lazy, that's what you're going to
earn, is lazying around. But if you're going to work, you're going to earn some-
thing. [85,4:21]
These ideas were repeated in the sixth interview, when I asked him several ques-
tions about "getting ahead":
CS: If people can't get ahead in the world, who is to blame for that?
JL: I don't know that you could blame anyone. You, you are the one to blame. It's
you. Because... you can achieve anything your mind can conceive, [del. 4 lines] So
we all have the potential of being a millionaire, if that is your goal. You've got to
have a goal. And if you're willing to work at it, hard enough, then you can achieve
it. [85,6:17]
CS: Is the system fair? Does everyone have an equal chance to get ahead?
JL: If you take it as a single person. Over everybody else. One person as - if he. Or
Research on cultural discontinuities 219
she. Wants that bad enough, then he and she will achieve it, because everything is,
everything is out there. [85,6:18]
Notice that in these passages his temporal horizon has narrowed to a
person's lifetime and within that time period, he sees progress. Society is
imagined as collections of atomistic individuals ("... a single person. Over
everybody else. One person . . .") and each person is free to "achieve"
whatever his or her "mind can conceive." The affective tone of these
passages is buoyant and optimistic, expressed in expansive spatial meta-
phors ("go with it as far as we wanted to go"; "no restriction"; "You can
go from zero to the top"; "everything is out there").
FEELING RESPONSIBLE FOR OTHERS. Appearing at various points
throughout the interviews was still a third, more communitarian (Bellah et
al. 1991)
voice. This voice posited a different version of "success," one in
which raising "responsible" children and caring for others were worthier
goals than becoming rich. As I will show, if we look at the contexts of their
expression as well as their contents and the voice in which they are
expressed, Lovett's CAN' T FIGHT THE SYSTEM and ACHIEVING ANY-
THING YOU WANT schemas appear to be compartmentalized, but his
FEELING RESPONSIBLE FOR OTHERS schema is partly integrated with each of
these. For the time being, however, let us just look at voice and contents: In
these respects, Lovett's FEELING RESPONSIBLE FOR OTHERS discourse
was quite distinct.
First, notice in the next passage that, unlike the ideas of structural
constraint and lack of progress over a long time span that he expressed in
his CAN' T FIGHT THE SYSTEM voice and the competing ideas of un-
bounded agency and linear progress in an individual's life that he ex-
pressed in his ACHIEVING ANYTHING YOU WANT voice, when Lovett
talked about raising children he described both continuity and potential
change over a period of three generations. For example, when I asked
Lovett to tell me a little about himself at the beginning of the first
interview, he told the story of how he incurred his disability, then went on
to the following:
I'm happily married. And I've got three terrific kids and wonderful grandchildren.
I feel so, that with my education, which was only to the 9th grade, I've been fairly
successful, um. . . raising my family.
And you can judge that by the way they
raise their children. My mother and father always said that they would always wait
to see how I handled my children, brought them up, to find out whether or not they
did a good job and I feel so that my brother and I never had any problems. His and
my children seem to be doing well. [85,1:2-3]
Over the course of subsequent interviews Lovett told a duTerent story
about his brother but consistently stressed intergenerational continuity,
220 Practice and possibilities
for example, in a discussion about the importance of teaching children
good habits:
And you can see it and I see it within my own children, compared to my brother's
children, where he. . . did things a little bit in reverse or opposite from me, so my
children, in my eyes, have developed in my way, which is correct, if you will, where
my brother may've taught his children, or may not've taught his children, the good
habits, so therefore they learned bad ones and they've continued with bad ones and
grown bad ones. Where I taught my kids, my children, good habits, and they've
improved on their good habits and I see it within their children. So, it is a direction
that you give them that they follow. And, and it's progressive. Everything is
progressive, I guess, if you will. You start something off in a good line and it's going
to grow in a good line, like a tree, if you've got a good seed, it's going to grow a
good tree, but if something contaminates that seed, then you're going to have a bad
[tree]. [85,3:17]
Notice also that unlike Lovett's self description as a member of the class of
working people (opposed to other classes and interest groups) in CAN' T
FIGHT THE SYSTEM and as a single individual in ACHIEVING ANY-
THING YOU WANT, his emphasis here is on his membership in a family.
Belonging to a group of people who should care about one another (and
certainly not hate one another) is stressed in the following passages as well,
when the topic is not his own private household, but how he would have
run Ciba-Geigy's Cranston plant if he had been in charge:
I would not've done what, what they did, because I care about people and I care
about myself. I care about the environment. I'm - I care about - if you want,
patriotic - 1 care about people. I like people. I don't - I've worked with men, that -
fact, my direct foreman, where I worked last, did not. He admitted, "I do not like
people.** And he did not associate with anyone within the building, [del. 5 lines]
[B]eingjust the opposite of that, I would, I am the one in the shop, if there had been
a fire, an accident, I was the first one on the spot whether I was the furthest away or
not, so I was that type of person. I would jump in and do something, even - 1 have
done several things, where it not - regardless of what it may harm, it may be
harmful to me. [85,2:2]
And it upsets me to no end to think that there are groups like the Ku Klux Klan
and the German organizations and things like that. [del. I line] I can't understand -
I believe that, I believe in free speech and all that but these people will get right up
and - in an audience - and say, I hate the Jews and I hate the Black and I hate. [del.
21 lines] [I]t just seems so sad that what we, the average person believes in this
country and then have these same people say, I'm an American, but I hate this or I
hate that one. Or we should kill all of these and all of them. That isn't what the
American way is, that isn't what our ancestors fought for. [85,4:25]
Finally, unlike the resentment and despair that permeated Lovett's CAN' T
FIGHT THE SYSTEM talk and the enthusiastic optimism of his
ACHIEVING ANYTHING YOU WANT talk, the dominant affective tone of
Research on cultural discontinuities
Table 8.1 Lovett's Multiple Voices
Opposing classes
resentment, despair
continuity & progress
2 or 3
his FEELING RESPONSIBLE FOR OTHERS talk was contentment, a word
he used when I asked him to describe himself:
I feel so that I'm content. And Fm satisfied. If I was to die tonight I'd have no
regrets. I really don't. I feel so that I have lived a good life. And fortunately for me
I've had [inaudible] an exceptionally good wife. [Irene is present.] And Fm very
pleased with my children. Fm very pleased with my grandchildren. I may've missed
out on something but because I don't know what Fve missed out on then I don't
really - 1 really haven't missed it. Until you have something and then lose it you
don't know so - Fm very content. And I don't know how my wife here - 1 can't
speak for her, but be - , my own personal feelings, I feel so that I have done perhaps
my thing that I was designated to do. [85,5:27]
It seems clear that Lovett has internalized conflicting social discourses,
which he expresses in distinct voices (see Table 8.1).
Other evidence for compartmentalization. A completely different sort of
evidence for compartmentalization comes from the survey on welfare
reform I conducted in January 1995. The survey was designed as an
experiment to see whether the priming provided by the initial survey
question would activate different schemas, producing divergent responses
to a set of three questions about possible directions for welfare reform.
(There is considerable research on such context effects in surveys, e.g.,
Clark and Schober 1992.) All respondents were told that this would be a
survey about welfare reform. Then some respondents were told that before
we got to welfare reform we wanted to know if they agreed or disagreed
222 Practice and possibilities
with the following statement. For some the statement was, "One of the
biggest problems in American today is that too many people avoid taking
responsibility for their lives." I anticipated that this statement, with its
focus on personal responsibility, would activate individualistic ways of
thinking, which put the blame for people's poverty on their own behaviors.
Others heard the statement, "One of the biggest problems in America
today is that we have forgotten that all of us, rich and poor, are in the same
boat. Giving help to some people now will eventually help other people
later." The reference to everyone being in the same boat was supposed to
activate communitarian ways of thinking, which focus on the mutual
responsibilities of members of a community. Still others heard the state-
ment, "One of the biggest problems in America today is that the average
person pays too much in taxes and doesn't get enough in return from the
government." I thought this statement would activate populist schemas,
which focus on the way the average person is disadvantaged in relation to
powerful persons and institutions. Finally, a quarter of the respondents
(randomly chosen in each case) heard none of these statements. Next, all
respondents were asked their opinions of three possible ways of changing
the welfare system: eliminating benefits for children born out of wedlock to
teenage mothers; providing welfare for two years only; or replacing welfare
with government-subsidized child care, health insurance and jobs for
everyone who needs them.
My hypothesis was that most Americans have
internalized all these ideologies (individualist, communitarian, and popu-
list) but they are compartmentalized. The way someone thinks about
various options for welfare reform at a given time will depend on which of
these schemas is most strongly activated.
I found that the priming provided by the initial question did seem to affect
the majority of the respondents* expressed attitudes about the three propo-
sals for welfare reform. Those who were asked to think about the statement
"too many people avoid taking responsibility for their lives" were more
likely than the rest to approve of eliminating benefits for teenage single
mothers. The populist-primed respondents were significantly more likely
than the communitarian-primed respondents to give their strongest support
to more extensive government benefits for all, while the communitarian-
primed respondents were significantly more likely than the populist-primed
respondents to favor a two-year limit on welfare benefits. Although I had
not predicted that pattern of results, they make sense. Asking people to be
more giving might provoke the resentful reaction: I'll give, but I don't want
to do it forever. Asking people whether the average person gets enough in
return from the government in relation to the taxes they pay, can easily
suggest the response that more government services would help balance the
scales. Not that the answer people gave is determined by the question, for the
Research on cultural discontinuities 223
populist-primed respondents could also have decided lower taxes would be
the better way to balance the ratio of taxes to government services. (Full
results and analysis are presented in Strauss 1995).
In sum, both my survey results and Lovett's talk suggest a large degree
of horizontal compartmentalization of discordant schemas. These
schemas are equally accessible; which is expressed depends on the context.
Evidence for partial integration
Both of these studies had results, however, that show exceptions to this
picture of compartmentalization. Thus, my experimental manipulation
did not work for all of my survey respondents. For each of the three
questions about reforming welfare, respondents had the option of agreeing
or disagreeing "strongly" or "not so strongly." Respondents who
answered "strongly" to all three questions were much less affected by the
wording of the initial statement than the majority (84 out of the total of
143 respondents) who answered "not so strongly" to at least one.
suggests that when people care about a topic, they are less likely to
I found this to be true as well for Jim Lovett and other interviewees
whose discourse I have analyzed closely. Each of them had one schema
that appeared across a variety of discourse contexts, partly integrating
their separate schemas, and this schema seemed to be related to emotion-
ally salient experiences earlier in their lives.
As I hinted earlier for Lovett, his partly integrating schema was
FEELING RESPONSIBLE FOR OTHBRS. Notice, in the following quote,
how the responsible, caring family man appears in the midst of what is
otherwise a typical CAN' T FIGHT THE SYSTEM passage:
[CS had asked Lovett to respond to the following statement: "People often talk about
there being different classes. Do you agree?" After clarifying that I meant economic
classes, Lovett agreed] Fm sure there's classes of people. Because - yes, Pm sure
that there are. So, there are, there are poor people. That are dependent upon
income from an outside source. There are the working class of people that are
working for someone else that are responsible people to see that their family is cared
for and maybe not a freedom of choice of- but they are responsible for other people.
Through taxes or welfare - it's all through taxes or deductions from their wages.
And then you have the upper class or the wealthy people that have people working
for them. That are in a class of their own. They, again, are looking for someone to
work for them to earn them money to make them wealthy or make them successful.
And not necessarily concerned with that person because there are so many of the
working class that if that one person is not doing his job, then he is replaceable,
expendable. Where he is in a, in a class of his own. And he only cares for his class.
The poor person does not - maybe, for some reason, he has no control over - is
224 Practice and possibilities
stagnant in his line. The workingman, it seems so that he is responsible to help the
poor man and he is also responsible to make the rich man richer. So he is a middle
class. He is not poor and he is not rich. And it seems so the middle class person is
carrying the whole country or the whole world. [85,6:24-5; boldfaced emphasis
This passage is almost identical to the very first ( CAN' T FIGHT THE
SYSTEM) passage I quoted from the interviews with Lovett. Note the
highlighted phrases, however: here the working man (who is "middle
class" because he is caught between the rich and the poor) is described as
"responsible people to see that their family is cared for" and "responsible
for other people," while the rich person is condemned because "he only
cares for his class."
A more subtle indication of partial integration is Lovett's use of "stag-
nant" to describe the poor person. I suspect he was thinking especially of
his poverty-stricken brother because earlier in the same interview Lovett
had said the following about his brother: "I was able to pick and grasp
onto things where he was, kind of got stagnated. And he stayed in this
jewelry shop just doing small nothing jobs for years and years and years
and years" [85,6:10]. Throughout the interviews mention of his brother
was highly correlated with Lovett's FEELING RESPONSIBLE FOR
OTHERS talk; recall that earlier I quoted Lovett as saying that his brother
"did things a little bit in the reverse or opposite from me," in bringing up
his children. Lovett also frequently criticized his brother for drinking
heavily, and agonized over whether to help him financially, given that
experience had shown that the money would probably be wasted. The guilt
he felt about his brother seemed to be reflected more generally in his
discussion of poor people. Thus, Lovett's statement, "The poor person
does not - maybe, for some reason, he has no control over - is stagnant in
his line," suggests to me that his FEELING RESPONSIBLE FOR OTHERS
schema is partially integrated with his CAN' T FIGHT THE SYSTEM
Similarly, Lovett's FEELING RESPONSIBLE FOR OTHERS voice also
occasionally intruded into his discussion of getting ahead in Amway:
[W]ith the business that we were in, as far as high as you wanted to go with this
then, even if you dropped out, there's always someone that's going to pick up so no
one is left stranded, [del. 3 lines] The Amway corporation code of ethic's been
copied by a lot of big corporations. [85,4:19]
In sum, if we look at the context of Lovett's talk as well as its content and
voice, then he appears to have some schemas ( CAN' T FIGHT THE SYSTEM
and ACHIEVING ANYTHING YOU WANT) that are compartmentalized
and another one (FEELING RESPONSIBLE FOR OTHERS) that is partly
Research on cultural discontinuities 225
integrated with each of the first two. Simply looking at the parts of my
transcripts in which these different schemas are expressed helps demon-
strate this. All of the CAN' T FIGHT THE SYSTEM passages I quoted
appeared in either the first or third interviews, when I was asking Lovett
questions about the work conditions that left him with chronic bronchitis,
the air pollutants from Ciba-Geigy that aggravated his breathing prob-
lems, the way company officials and state bureaucrats failed to deal with
these situations, and other topics that occurred to him in the midst of these
discussions. (Other, similar, passages appeared as well in the portions of the
second interview that returned to these topics, and in response to related
questions in the sixth interview.) Lovett's ACHIEVING ANYTHING YOU
WANT talk was limited to the fourth and sixth interviews, when I asked
questions about "the free enterprise system" and "getting ahead."
FEELING RESPONSIBLE FOR OTHERS talk, on the other hand, came from
all six interviews. Not only was it scattered throughout the course of the
interviews but, as I have just shown, these ideas were tightly linked to his
other schemas, so they were expressed jointly with these other ideas.
Partial integration can be further illustrated by contrast with a very
different form of temporal contiguity that appeared in the last interview. In
that interview, I had a standardized list of questions and asked inter-
viewers to give shorter answers than they had previously. The faster pace
of questions activated in quick succession his CAN' T FIGHT THE SYSTEM
and ACHIEVING ANYTHING YOU WANT schemas, forcing both into
awareness simultaneously and temporarily at least moving him from com-
partmentalization to ambivalent flip-flopping. When I asked Lovett, "Is
the system fair? Does everyone have an equal chance to get ahead?," first
he responded with clear ACHIEVING ANYTHING YOU WANT language,
"If you take it as a single person. Over everybody else. One person as - if he.
Or she. Wants that bad enough, then he and she will achieve it, because
everything is, everything is out there." But just minutes earlier, in response
to my question, "What things keep people from getting ahead in the
world?," he had replied, "in order to succeed, it seems so you have to have
some kind of financial backing and most people don't have it, therefore
they're afraid to risk. What they have. Because it took them, they had to
work so hard to accumulate it" [85,6:16]. With that CAN' T FIGHT THE
SYSTEM answer fresh in mind he continued his answer to "Is the system
fair?" Right after he said, "Everything is out there," he continued
You just have to be willing to - 1 realize that - . [sigh] All right. You have to, I'll
have to back that up [i.e., back up]. Because, again, there are certain obstacles that
are depending upon financial aid. And everyone does not have enough time or
energy to have enough money to achieve everything, so that - it would be a main
obstacle whether he -
226 Practice and possibilities
Again, Lovett struggles, and immediately switches back to his
[A]nd then I, you hear of cases where people start with zero pennies and, and do get
these college degrees. Because they are willing to work two jobs and study every
waking moment. So. It, it goes back to the personal person - his own drive I guess.
Unlike the tight integration of "There are the working class of people that
are working for someone else that are responsible people to see that their
family is cared for," where the narrative of class opposition converges with
the narrative of the responsible family man, here there is only a switching
back and forth between contained, contradictory narrative lines.
Other examples of integration. Lovett was not the only interviewee who
demonstrated examples of partial integration of cognitive schemas. Carol
Russo, a middle-aged school secretary whose violent metaphors I pres-
ented earlier, integrated populist, feminist, Pro-Life, and other discourses
in her BEING HURT BY PEOPLE ON TOP schema. Throughout the inter-
views she applied the same schema, dividing the world into the weak but
good (depending on the context, fetuses, most women, law-abiding citi-
zens, or the United States) versus the strong, violent, and bad (aborting
mothers, most men, criminals, and the Soviet Union - recall that the Cold
War was still on when these interviews were conducted). Michael Fields, a
fortyish labor relations consultant and former Vietnam veteran and anti-
war protester, partly integrated his current establishment and former
anti-establishment views with his KNOWI NG HOW THE GAME WORKS
schema. When he was younger, he said, he wanted to "blow the system
up"; now, as a successful professional, he works very well within the
system, but his discussion throughout was dominated by an assessment of
"life as some sort of a game" that he applied in cynical assessment both of
the Vietnam War and his work now. Anna Monteiro, a college graduate in
her late twenties working as a clerk in a social service agency, had a
different philosophy:
Life is a learning experience. You have to learn, from everything you do, you have
to learn something from it, whether it be good or bad. [95,2:3]
This self-confident LIFE IS A LEARNING EXPERIENCE voice partly integ-
rated the ideas she had acquired from her college social science courses,
feminism, individualist discourses, an immigrant perspective (she is a
second-generation Cape Verdean-American), and her consciousness of
color (she is dark-skinned).
Matthew Healey, a sales representative in his mid twenties, partly integ-
rated religious, animal rights, and Star Trek-inspired discourse in a schema
Research on cultural discontinuities 227
I described, using his own words, as LISTENING FROM THE MARGINS.
Here are some examples:
I do feel more comfortable dealing with people on this, I don't know, in the sense
of, I see their position and I will, sort of, be on a radius there around. You know, I
don't necessarily put myself in a power position and say, "Okay, this is what it's
going to be." I say, "All right, well, what do you think?" [95,2:8]
This way of thinking affected Healey's career planning. Even though he
did not have a girlfriend when I interviewed him, let alone a wife and
family, he was planning his career around the compromises he assumed he
would have to make to accommodate to these:
MH: Let's say if I have the option to take a job overseas and get a higher salary or
stay here and do something that perhaps wasn't as much something as what I
would like to do but I would have the benefit of having a family, I would definitely
choose the family.
CS: Huh. Why couldn't you bring your family with you, overseas?
MH: I could, I could, and that's certainly a possibility, [del. 1 line] I mean no one's
ever really put it to me that way, and if my family and my wife would be willing to
go with me, so -1 just assumed it would be a case where it would be a hard decision,
I wouldn't end up getting the easy decision, "Oh sure, let's go," but I'm prepared to
at least deal with a compromise, not saying, "Too bad, that's what we're going to
do." [95,2:14]
This sort of decentering could also be traced in Healey's thinking on other
topics. Just as he rarely places himself in a power position, in each of the
following examples, he shifts established structures of dominance. In the
first of the following two quotes he is discussing illegal immigration; in the
second, vegetarian dog food:
fTJhere was an episode [of Star Trek] where they talked about these computer
components that were made in Dakar, Senegal? I think it's Senegal. Dakar,
Senegal. And here's a country that doesn't necessarily have as much technology
right now but look it, okay, we made this world a unit, we are all together, invest
where everyone is given an opportunity. That just says to me, as an aside, look it,
okay, let's make Mexico or let's make Central America a technology center. Then
the people, perhaps the Americans will be going there to work. You know, it is
entirely possible. [95,1:14]
You're looking at an animal who is supposed to have meat, you're putting human
values on a dog. You know, you're imposing it, your ideas on it. [95,2:7]
In the first passage Healey envisions a time when the United States,
currently dominant in the Americas, is "on the radius" of economic
production; in the second, he suggests that human values should not
dictate how "lower" animals should be treated.
The integrating effect of emotionally significant experiences. For each
228 Practice and possibilities
interviewee, these partly integrating schemas can be traced to emotionally
significant experiences earlier in their lives. Lovett said that because his
father never spent much time or showed much affection with him, he had
made a point to be a caring, involved parent. During a second set of
interviews I conducted with Russo in 1990 it emerged that she had good
reason to expect to be hurt by people on top: her father had physically
abused her. Michael Fields's cynicism about "the game" seems to have
been heightened considerably by his experience as a Vietnam war draftee.
Anna Monteiro's self-confidence seems to stem from her close family, her
father's encouragement to do things differently from the crowd, and her
sense of being attractive to men. Matthew Healey's tendency to look at the
world from a lower or marginal position he attributed to having been very
heavy when he was younger:
CS: Can you tell me a little more about what kind of kid you were growing up and..
MH: Well, you know, I think one of the biggest things, in third grade, I had
pneumonia. For three months. And, well, when I was home I had a tutor here but I
ate and ate and ate and ate and ate. So I mean, I was an average weight kid before
and when I went back to school I was overweight. And from there on in, for the rest
of my life, that was it. It was a battle from there. And it's hard for people to
understand, but if you're a kid growing up heavy, there's - 1 mean, little kids are
cruel. Kids are cruel. And that was a very, sort of, I think it helped formed my
personality. I do feel more comfortable dealing with people on this, I don't know,
in the sense of, I see their position and I will, sort of, be on a radius there around.
[continues with passage presented earlier] [95,2:8]
It is also important to stress here that such emotionally salient experiences
and distinctive personal outlooks inflected the way each interviewee inter-
nalized and expressed even a single discourse, as we saw in the contrast I
provided earlier between Russo's violent populist imagery, which doubt-
less derived from the abuse she suffered as a child, and Lovett's images of
trying to speak and not being heard, which might be related to the
marginalization he had suffered by being illiterate until very late in life.
Emotionally significant experiences can shape the way people appropriate
a given social discourse and lead to some discourses becoming all-purpose
interpretive models for individuals, repeatedly activated by them to make
sense of the world.
Explaining compartmentalization and partial integration on the
extrapersonal and intrapersonal levels
Elsewhere (Strauss, 1997) I have used these examples (and others) to
critique the literary theorist Fredric Jameson's account of postmodern
psychological fragmentation. Jameson's argument for psychological frag-
mentation rests almost entirely on public culture (and mostly elite, high
Research on cultural discontinuities 229
culture at that). He points to the music of John Cage, "in which a cluster of
material sounds. . . is followed by a silence so intolerable that you cannot
imagine another sonorous chord coming into existence and cannot imag-
ine remembering the previous one well enough to make any connection
with it if it does" (Jameson 1991a:28). Another example he gives is a poem,
China, by Bob Perelman, a San Francisco poet of the "New Sentence"
school, in which there are only tenuous connections among the lines of
poem because, unbeknownst to the reader, each sentence is Perelman's
caption to a photograph in an album he found in a Chinatown market.
Postmodern art of this sort, Jameson notes, fits very well Lacan's
description of schizophrenia: "when the links of the signifying chain snap,
then we have schizophrenia in the form of a rubble of distinct and unre-
lated signifiers" (Jameson 1991a:26). While Jameson does not believe
postmodern artists like Cage or Perelman are schizophrenic in the clinical
sense, he does think that such art is symptomatic of a cognitive problem:
loss of a genuine sense of historicity and resulting psychic fragmentation:
If, indeed, the subject has lost its capacity actively to extend its pro-tensions and
re-tensions across the temporal manifold and to organize its past and future into
coherent experience, it becomes difficult enough to see how the cultural produc-
tions of such a subject could result in anything but "heaps of fragments" and in a
practice of the randomly heterogeneous and fragmentary (1991a:25)
Jameson's argument is that late capitalism - with its obliteration of all
traces of the premodern, its media saturation, its information overload,
and the undependability of work - has created a form of consciousness in
which we "channel switch" among different "compartments of reality"
(199lb:373, 372), unable to integrate these compartments into a larger
"map" of our world. Fragmented minds, in turn, create fragmented works
of art. Sometimes he focuses not so much on forms of consciousness as on
the world reflected in consciousness: if our world is fragmented, how can
our cognitive and cultural representations of that world avoid being
fragmented as well? In any case, he treats fragmented postmodern art as
good evidence for fragmented outlooks.
If I am right, however, that people's schemas are more integrated than
Perelman's poetry, Cage's music, and so on, then we cannot use these art
works to gauge forms of consciousness (not even the artists' forms of
consciousness). Nor can we say, as Jameson does at one point, "'We' thus
turn out to be whatever we are in, confront, inhabit, or habitually move
through" (Jameson 199lb:373); this oversimplifies the processes of social
and cultural construction we discussed in chapter 4. Once we recognize
that we cannot make this assumption, we can go on to consider some
interesting questions. How are diverse ideas managed at the social and
cognitive levels? Are there historical and cultural differences in the social
230 Practice and possibilities
management of diverse public ideas, and, if so, what are the cognitive
consequences of these?
At the beginning of this section I listed several ways that competing
discourses might be managed, socially and cognitively. Each of the social
possibilities I mentioned has a formally similar cognitive counterpart. Thus,
socially approved synthesis of public discourses is like cognitive integration
of them; social marginalization of one discourse in favor of another is similar
to the cognitive operation of vertical compartmentalization, where one
schema is easily expressed and accessible to consciousness while the other
remains more implicit (see Strauss 1990); and letting "a hundred flowers
bloom" on the social level is comparable to the cognitive operation of
horizontal compartmentalization, which I illustrated with Lovett.
Stressing these parallels, however, overlooks the fact that distinct fac-
tors come into play in the regulation of diverse ideas at the social and
psychological levels. First, consider the psychology of compartmentaliz-
ation. The connectionist learning model we outlined in chapter 3 suggests
that ideas will be cognitively linked only if they are associated in our
experience. If we hear populist ideas at the bar and communitarian ideas at
houses of worship, they will probably be stored in separate, unconnected
neural networks. For the most part we are only as consistent as we need to
be to get things done. Practical tasks do not always require complete
consistency, as we pointed out in chapter 5 with the example of the
contradiction between Paula's typically US American belief in self-reliance
and her desire to turn over the job of changing a flat tire or fixing a leaky
faucet to the nearest man.
On the other hand, if two contradictory schemas should repeatedly be
activated in the same context and if we need to decide between them in
order to act, we might go through the sort of ambivalent flipflopping I
demonstrated from my sixth interview with Lovett, but then come to a
resolution - either selecting one or reaching a compromise that enables us
to feel we have satisfied both. The need to do something sometimes
promotes integration. Furthermore, it seems that when a schema is linked
to strong emotions, it is frequently activated and used to interpret new
experiences - including disparate experiences - creating links among them.
For example, Healey's experience of being on the social margins when he
was younger seems to have given him a way of thinking that he now
extends to many other contexts, just as the abuse Russo suffered as a child
now leads her to expect to be hurt by anyone in power. It appears that
emotionally laden memories help create partial integration.
If we move to the social level, then we are asking why in some domains
diversity is tolerated or encouraged and in other instances one set of ideas
is dominant - a very different question. For example, under current
Research on cultural discontinuities 231
conditions of capitalist production and distribution, it works out well for
entertainment to be marketed to small market segments (e.g., multiple
cable stations) while at the same time news delivery is becoming increasing-
ly expensive and centralized - except on the Internet, which is highly
"compartmentalized" for technical reasons that are unrelated to the neur-
obiologies grounding of cognitive compartmentalization. Historical
changes in the regulation of diversity at the social level are not necessarily
mirrored at the psychological level.
Jameson's argument could be put as follows: false consciousness (verti-
cal compartmentalization) has been replaced by postmodern multiplicity
(horizontal compartmentalization), both in public discourse and in
people's consciousness. I think that it is premature to discount the extent
to which, hidden by multiple lifestyle possibilities, old-fashioned hegem-
ony remains, with some ideologies much more widely disseminated and
validated than others. Even if hegemonic control were a thing of the past,
however, it would not follow that people's outlooks are as divided as the
messages swirling around them because it is precisely when confusion
threatens that we are most likely to attempt to find unifying discourses to
restore an inner sense of order and predictability to help us know what to
think and do. (See Giddens 1991 for a related discussion.)
That is not to say that new media make no difference. On the one hand,
one effect of the new abundance of conflicting information could be
passivity and loss of interest (Ewen and Ewen 1992) that might promote
horizontal cognitive compartmentalization. On the other hand, new tech-
nologies that make it possible for citizens to gain access to official informa-
tion might have a countervailing, empowering effect, making the average
person feel more involved and interested, hence more likely to selectively
integrate information from diverse sources.
I should not overstate these divergences of the social from the psycho-
logical. As we discussed in chapter 5 and Naomi Quinn illustrated in
chapter 6, effective individual resolutions of widely shared and longstand-
ing conflicts are likely to be frequently repeated in people's everyday talk,
leading over time to widespread social diffusion. The organization of
people's schemas does not automatically replicate the organization of
social discourses, but they are part of the same larger conversation.
Analysis 2. The disparate motivational effects of different forms of
culture learning
As I have already suggested, disparate ideas are not just a matter of
analytic interest; they can raise practical issues for the belief holder. If you
232 Practice and possibilities
hold discrepant beliefs, on which ones do you act? Or, to put it in standard
psychological terms, which ideas are motivating? In chapter 4 we dis-
cussed, using our fictional Paula example, some of the factors that have to
be present for social messages to have motivational force for individuals.
Paula, as we described her, is self-reliant (in many realms) but does not set
aside much of her money in savings, even though she believes that she
should save more of her money just as she believes it is good to be self-
reliant. What accounts for the difference? We pinpointed seven conditions
that seem to be important. On the extrapersonal side, these are (1) positive
social discourses about the action; (2) teaching that attempts to link the
ideas being taught to strong emotions in the learner; (3) repeated presenta-
tions of the action; (4) social institutions, practices, and people that
facilitate or demand the behavior more than opposing behaviors (on this
point see Holland 1992a). Correspondingly, on the intrapersonal side, the
relevant conditions are (5) attention directed toward repeated presenta-
tions of the action, so a schema for how to do the action can develop; (6)
association of positive feelings with the action (or of negative feelings with
nonfulfillment of the action) that are stronger than the feelings associated
with opposing behaviors; and (7) a cognitive connection between the
schema for this action and a person's self image or identity. (Note, with
respect to the last point, that the first section of this chapter suggests that
someone could have several self images or identities - for example, Lovett
could be said to have one view of himself as a working man, another of
himself as a potentially rich person, and another of himself as a responsible
father and caring member of a community. I will discuss the situation of
multiple self identities shortly, stressing the importance of those identities
that are partly integrating, i.e., activated in a variety of contexts.)
We did not discuss very much in chapter 4, however, the fact that these
conditions for motivation refer to different forms of learning. If all of these
conditions are satisfied, they reinforce each other and no problem arises
about the differing effects of each. In the case of self-reliant behaviors, all
or most of these conditions are satisfied for most US Americans. That is,
most US Americans will frequently hear it said that you should be self-
reliant, be rewarded as children for being self-reliant, see other people
being self-reliant (at least in some domains), and be faced with situations
where they have to be self-reliant to a greater extent than in societies where
kin, community, or the state play a more supportive role. The result is that
most US Americans learn to feel good when they are self-reliant, may
come to include self-reliance in their image of themselves, and have
schemas for how to do many sorts of things for themselves. Of course,
there is intracultural variation in these conditions. For example, many
married middle-class US women are not self-reliant in financial matters.
Research on cultural discontinuities
There are also sharp limitations on the domains within which self-reliance
is expected (no one is expected to provide their food, clothing, and shelter
from raw materials).
But within the expected domains of self-reliance for
a given group in this society, the motivational conditions listed earlier
typically overlap and reinforce each other.
What, however, if all seven conditions are not satisfied? As Spiro (1984,
1987a) and Holland (1992a) have observed, sometimes people only pay lip
service to cultural values.
In the example we gave in chapter 4, this was
the case for fictional Paula's poor saving habits. Only the first condition
was satisfied - her parents often said that she should save for a rainy day.
Moreover, contradicting her parents' messages are many other implicit
social messages encouraging unnecessary spending, on which she acts in
the absence of any emotional investment in the less enjoyable goal of
putting money aside. (While the Paula example is fictional, we suspect
most US readers can think of someone to whom it applies.)
These examples suggest, then, that there are at least two kinds of
motivation. The first could be called lip-service motivation: we endorse the
value but do not act on it. The second is a motivation to enact routine
behaviors. These are the ones that follow from what Bourdieu (1977) calls
habitus or Geertz (1983a) calls common sense:
a kind of automatic,
unreflected-upon tendency to act as everyone else (like you) does. As we
explained in chapter 4, this sort of motivation is inertial: unless drastic
events force you to do something different, you continue to do what you
have done in the past or what nearly everyone like you (as far as you can
see) does. For many upper-middle-class adolescents, an example might be
going to college. In addition to these two kinds of motivation, a third sort
is the motivation to perform nonroutine actions. An example might be
going to college if most of your friends are not college bound, the school
guidance counselors do not expect you to, you do not know how to obtain
financial aid, and so on. These actions may not be physically harder than
the routine ones, but they require conquering the anxiety that comes from
doing something different, as well as taking the trouble to find out what to
do. A mundane example is driving over unfamiliar roads. Some people (I
am one of them) would rather take a longer, familiar route than a shorter,
unfamiliar route that would require paying more attention to where they
are going. A less mundane example is expressing unpopular opinions (e.g.,
religious belief among nonbelievers or atheism among believers).
Drawing on the list of motivating conditions presented earlier, let us say
that someone has only lip-service motivation if they have internalized
positive social discourses about the action but not connected these values
to a partly integrating self image, or to other strong feelings, or to schemas
for how to enact these social values. These values can be explicitly stated
234 Practice and possibilities
(otherwise you cannot pay them lip service). Someone is motivated to
enact routine behaviors, on the other hand, if they can observe or partici-
pate in repeated examples of the action, they pay attention to these
examples, and there are facilitating people and institutional structures.
Sometimes these routine behaviors are linked to positive social discourses
and a conscious self image, sometimes not. Finally, someone is motivated
to perform nonroutine behaviors in cases where there are few opportuni-
ties (normally, for someone like them) to build up a schema for the action
and little social support for the action but performing the action has
become associated with strong feelings and a partly integrating self image.
In the rest of this section, I will give examples of each.
Lip-service motivation
When he was talking about the opportunities to make money through
Amway sales, Jim Lovett enthusiastically endorsed the value of monetary
JL: I used to get so excited telling somebody about it [Amway] I could hardly talk,
you know? But it's the type of business where you can start with a dime and end up
a millionaire. If you want to. If you want, it's there. Really there. [85,4:21]
Later in the interviews Lovett generalized to the opportunities available to
anyone in the United States:
CS: If people can't get ahead in the world, who is to blame for that?
JL: I don't know that you could blame anyone. You, you are the one to blame. It's
you. Because... you can achieve anything your mind can conceive, [del. 4 lines] So
we all have the potential of being a millionaire, if that is your goal. [85,6:17]
Several of the Ciba-Geigy employees made similar statements, at least on
occasion. Anthony Gallucci talked about "trying to bring your kids up
properly" to show them "the advantages of having a good job - or
disadvantages of digging a ditch." When I asked Daniel Collins if he
admired successful businessmen, he said, "Anybody who succeeds, I ad-
mire him." (See Strauss 1990,1992b for more extended examples.)
Yet Lovett had decided not to move to Connecticut, where he said he
could have made more money as a welder, and he had never worked his
way up very high in the Amway sales ladder. Gallucci and Collins had
turned down opportunities to be promoted to a foreman's position. A
study of blue-collar workers at a New Jersey chemical factory found,
similarly, that out of fifty workers who had the seniority to become leaders
or chiefs, twenty-six declined (Halle 1984:154). This does not mean that
Research on cultural discontinuities 235
they do not care at all about getting ahead; the decision to work in a
chemical factory in the first place involves a knowing assumption of risk in
return for a higher salary than what is typically available to someone of
similar education working in a safer field. Lovett had learned the trade of
welding and thus had had better paying jobs than his brother, whom he
described as having "stagnated" as an unskilled laborer. Still, given that
Lovett knew about ways to make more money than he did and did not
pursue them, it is appropriate to say that economic success was a goal to
which he, Gallucci, and Collins paid lip service but which was not very
motivating for them. Why?
Part of the reason why Collins and Gallucci turned down opportunities
to go into management is that it had some negative associations for them.
At Ciba-Geigy, someone who joined management had to be prepared to
cut off social ties with his friends in the union. Friendships that crossed the
management-worker divide were discouraged from both sides: by man-
agers who thought friendships with the workers would undercut a fore-
man's authority and by workers who worried that their secret practices
would be communicated to management. Gallucci joked that he did not
want to see expletives linked to his name on bathroom walls; Collins said
he did not want to betray the workers who had elected him to a position in
the union's local. It is also important to note that since it is difficult to
become very rich if one starts out not so rich, economic success stories
probably circulate less among workingmen like these than they do among
the upper-middle classes, making it less likely they will have developed the
schemas that give a concrete set of actions they could take to get rich.
Still, it is not enough to say that Collins's and Gallucci's decisions were
the predictable result of a subcultural social message having greater force
than a message present in the larger cultural milieu, or of absence of group
support for the action. That does not explain why another man I inter-
viewed (Frank Hollingworth) began in the union at Ciba-Geigy but did
accept a promotion into management. Another interviewee (Al Cho-
quette) grew up in a working-class household, just like Collins, Gallucci,
and Lovett, but sought education and training so he could move into
white-collar jobs. Subcultural social messages do not necessarily motivate
people any more than dominant cultural social messages do, and people
are sometimes mightily motivated in the absence of group support. Other
ingredients are necessary.
The other crucial ingredient in this case was each man's self image. For
Collins and Gallucci, it seemed, it was important to provide a steady,
comfortable income for their families but it was not important to advance
as high as one could up the job ladders at work. The reliability and
adequacy of their incomes was sufficient for feelings of success in their
236 Practice and possibilities
breadwinner roles. Beyond that, self-esteem depended on other factors.
For Collins, advancement in the union was more important; for Gallucci,
having free time and being independent. (Unlike workers in some facto-
ries, the chemical operators at Ciba-Geigy had considerable free time and a
fair amount of freedom from supervision, depending on where in the
factory they worked.) Lovett said he did not move to another state, even
neighboring Connecticut, because he and his wife wanted to be near their
parents when they got older, a value that fits well with the importance he
places on taking care of others. Becoming well off was not part of a
frequently activated self image for any of them, and countering the value of
a higher salary were the negatives of lost friendships with other workers as
well as loss of personal time (management employees worked much longer
hours, typically) and higher stress. Simply put, other things were more
important for them than constantly trying to make more money. To put it
in the terms I presented in the last section, their getting-ahead schemas
were not connected to other, more emotionally salient and integrating,
schemas that determined the life choices they made.
Why do I speak of any motivation existing at all, in cases like these,
where someone accords value to a goal but feels only a weak urge to act on
it? Why not speak instead of a divide between cognition and motivation
(they know it, but are not motivated to do it) or label it more simply a case
of "weak motivation?" The difference between lip-service motivation and
other kinds of weak or absent motivation is that in lip-service motivation
there is social pressure to strive for the goal - pressure that the person
knows about and may have internalized. This is suggested by the defens-
iveness Lovett, Gallucci, and Collins showed in talking about their work
and incomes with me, given that they seemed to judge me to be better off
than they were. Lovett, for example, said,
So I was always able to provide a decent living for my family. We never had a lot of
- my own, my own family - a lot of elaborate things perhaps. We always were the
last one in the, in my group that I grew up with or went to school with, to have a
new car. But. We - our children never went without clothing. They never went
hungry. And Irene and I would probably not get to go to the movies or go out to
dinner as often as the others did, but our children never went without. So we were in
a sense good parents or providers or whatever. And we look back on it now with
the others that did afford, could afford themselves night clubs and, and restaurants
and new cars. That their children, because of it, have to've been left with a
babysitter a lot more than my children because we did stay with them. We were
involved with them and they weren't. So . . . their children have grown up and
gotten married and divorced where - and fallen away a little bit because maybe
they weren't attended to by their family, or, or afforded their parents as they grew
up as much as my children were. [85,5:18]
Research on cultural discontinuities 237
Collins recounted a conversation with the plant manager who had offered
him that promotion:
He said, "You stupe," he said, "you could have been a boss there." I said, "Well, I
chose to do what I'm doing." I said, "Maybe in later life," I said, "it might haunt
me." [laughs] [del.] I look back and I'm not mad at myself. Fm happy that I, you
know, didn't take that type of job instead of what I had to contribute to my fellow
workers and the union and international, [del.] I don't kick myself in the fanny as
some people would say for not taking it. [85,2:12]
When I asked Gallucci, "Is there anything about your life that you would
do over, if you could?," he replied,
[del.] I had a good time at Ciba-Geigy. I mean, I wouldn't, if I could find something
better, naturally, I would do something better, but if I - for a poor working man, I
had a good time there, over the years, for the most part. [85,5:20]
Interpreting statements like these is tricky. On the one hand, Lovett,
Collins, and Gallucci have acted on the alternative values they discuss here
(being an attentive parent for Lovett, dedicated to the union for Collins,
and having fun for Gallucci), so I know these statements were not being
made only for the sake of impression management. Still, it is clear that they
are aware that others might judge them to have made bad decisions from
an economic perspective and they accord enough value to that model of
success to explain themselves. Gallucci's attitude is hardest to pin down,
because he seemed to have a highly reflexive awareness of the impression
he conveyed, which he liked to play with. He said that at home, "I put on
my Archie Bunker act," meaning that he acts out the stereotype of an
uneducated "hardhat," even though he was very intelligent and well-read
(the magazines to which he subscribed included Harpers^ World Press
Review, and Wilsons Quarterly). Possibly he really thought of himself as a
"poor working man"; it is also possible that he anticipated that I would
look down on him and so talked about himself and his fellow workers from
what he imagined was my point of view, which he did not share. Either
way, he showed an awareness that the dominant social judgment is that
one should strive to be more than a "poor working man."
Motivation to enact routine behaviors
In an earlier publication (Strauss 1992b) I proposed that lip-service moti-
vation is always learned from explicit verbal messages and seen as an
optional value, while routine motivation is typically learned from unver-
balized actions and not usually seen as optional. I think now that the
contrast I drew between explicit and implicit learning was overstated: I
suspect I was too much under the influence of the stark opposition Bourdieu
238 Practice and possibilities
(1977) sets forth between the realm of explicitly stated, contestable values
(dogma) and unsaid routines that are not contestable because they are not
accessible to awareness (doxa). As we indicated in chapter 2, Bourdieu's
own examples do not bear out the hard-and-fast lines he draws between
these forms of learning, and Naomi Quinn has pointed out to me that
motivation to perform routine behaviors (like being a good mother [for
women] or being a reliable breadwinner [for men]) can certainly be acquired
from explicit social discourses, highly valued, linked to an image of self and
to emotionally salient experiences, and rejected by some people.
So my
earlier discussion needs to be modified.
In some cases the motivation to enact routine behaviors is learned in the
gradual, largely implicit way Bourdieu describes for the development of a
habitus, which we explained using connectionist models in chapter 3.
These are aspects of our behavior most of us take for granted, hence are
not linked to either validatory social messages or a conscious self image. In
these cases, no special push seems necessary to perform the behavior;
indeed, it would feel odd to do anything else. In the case of other routine
behaviors, however, our motivation to perform them rests not only on
repeated observation of routine practice but also on internalized social
discourses and emotionally laden experiences that shape a person's self
image. In these cases, people can probably imagine doing something else
and take pride, instead, in fulfilling their expected roles.
Still, what unites these different cases of routine motivation and distin-
guishes them from lip-service motivation is that routine motivation is
acquired through exposure to many examples of the action, that is, by
learning the (statistically) normal way to behave. Of course, mere exposure
is not sufficient: you also need to be paying attention. As we discussed in
chapter 3, social learning theorists have shown that there are many reasons
for emulating another's behavior (Maccoby and Martin 1983); therefore, it
can be difficult to predict whose behavior we will attend to and emulate, if
there is a choice. Once that largely unconscious choice is made, however,
then a lot of behaviors - posture, dress, speech patterns, food, and so on -
may be adopted without much thought. This is the whole realm of what
Malinowski called the "imponderabilia of actual life" (1961 [1922]: 18) and
is an aspect of "identity" that is overlooked by theorists who see identities
as only conscious self-representations and presentations.
As Bourdieu emphasized, people do not usually talk about these taken-
for-granted aspects of everyday life. Notice, for example, what Anna
Monteiro says and does not say about the way she dresses:
CS: What are some of things your father
told you or parents told you that you
have adhered to and have influenced you, do you think?
Research on cultural discontinuities 239
AM: [del 1 line] The main thing was, "The crowd goes one way, you go the other."
That nonconformist thing, you know.
CS: Can you give me a for-instance?
AM: Oh, fashion. Platform heels and hip huggers. [inaudible] Just because everyone
else is wearing them. Any kind of fashion, it's not necessarily that. Any fashion that
all of a sudden everybody's wearing, I'm not going to wear it just because every-
body's wearing it. I like to wear what's comfortable for me. [95,2:4]
Monteiro is proud to be a fashion nonconformist and indeed, she did not
dress in a trendy way. However, in addition to the fashion choices that she
makes deliberately as statements about her identity, there are the many
more respects in which she dresses the same way as other US women do
now, but differently than women did earlier or in other places. For
example, in India where I did a little research many years ago, respectable
women always covered their legs but often bared their midriff. In the
United States, it is the reverse. Nowadays, women unselfconsciously wear
pants to the office, but my mother can remember the day in the early 1970s
when the women in her office decided they would come in wearing pants
for the first time. As my mother's story suggests, even routine behaviors
can be noticed, discussed, and challenged, but this is unusual. Routine
behaviors of this taken-for-granted sort are typically left unmentioned;
they are what discourse analysts call, in the context of speech, presupposi-
tions of discourse (Brown and Yule 1983).
My male interviewees' actions of being the major breadwinners for their
families provide an example of routine motivation that is less taken for
granted. Like the routine aspects of dress that Anna Monteiro left unspeci-
fied, male breadwinning is the statistical norm in the United States.
the struggles that have made it easier for women to earn a family wage
make it more conceivable for men to give up the burden of being the sole or
even major breadwinner. At the same time, the scarcity of jobs that pay a
family wage for people at lower education levels and in some parts of the
country can make breadwinning very difficult and look less routine for
many men. Furthermore, men have always had the option of desertion, as
well as a great range of possible ways of interpreting and fulfilling the
breadwinner role. They have probably heard some talk about what is
expected in this role. My interviewees occasionally generated such talk
themselves. Tony D'Abrosca said, *Tm not rich, I'm not poor. I'm happy
with what I have. I've got a home and the kids have grown up pretty
good." Earlier I quoted Lovett's statement that he was usually the last in
his group to get a new car, "but our children never went without. So we
were in a sense good parents or providers or whatever."
In any case, beyond the pride they took in doing a good job at being a
breadwinner (however they defined this), there seemed to be the more
240 Practice and possibilities
taken-for-granted assumption on the part of the middle-aged married men
in the Ciba-Geigy study that of course they had to work. D'Abrosca, for
example, said about Ciba-Geigy,
Lot of people getting cancer there. So - over the years. And... it's a risk. We know
it. And - but the money's good. This state doesn't have too many good-paying
jobs. So. We're sort of trapped. So. I wouldn't want my kid going there. If he had a
choice. [85,1:5]
Of course, he would have a choice; nobody can force you to work in a
chemical factory. But by saying "if he had a choice," D'Abrosca implicitly
alludes to the situation he was in: If you want to be able to buy a house,
support a family, and pay all of your bills, you want to live in the same
state as your siblings and parents, you did not go to college (like most
people of your class at the time), and you assume your income alone (or
nearly alone) will be supporting the household, then you may leave your-
self no option (given the current state of occupational health and safety
regulation and enforcement, wage structures in the area, etc.) but to take a
dangerous job.
Nonroutine motivation
Still a third kind of learning comes into play when people acquire the
motivation to act in nonroutine ways. D'Abrosca, for example, began
competing in ultramarathons (races fifty miles or longer) when he was in
his fifties, making up for his asthmatic, unathletic childhood. He is also a
frequent contributor to the letters pages of the local newspapers, a photo-
grapher, and a collector of old books. His motivation to do these things
was not acquired by observing many examples of working men like him
doing those or similar things, nor was it the course of least resistance, given
current institutional structures and social practices. Instead, some of these
behaviors seem to have been driven by an image of himself as different
from other people ("TD: I'm different. CS: Yea. TD: I'm one of a kind
maybe. I don't know" [85,6:6]). This image was cognitively integrated with
many other schemas of his: it came up in many different contexts over the
course of the interviews, from discussions of politics to sports to his family
(Strauss 1992b). Also, it was linked to strong feelings associated with not
being able to perform athletically when he was a boy, which meant both
that he had come to value being "bookish" but perhaps also had always
wanted to distinguish himself athletically:
I had asthma as a boy. I still have it and I wanted to see what I could do. Wanted to
prove to myself I could do something because as a young boy I'd dream about
Boston [Marathon]. Impossible dream then, you know? 'Cause of the asthma. Now
I've done it three times. [85,2:12]
Research on cultural discontinue ties 241
Other interviewees provided remarkable examples of nonroutine behaviors.
For example, Linda Petty, a working-class woman who had two children
with severe disabilities, overcame a nervous collapse that hospitalized her
after the birth of the second child and went on to battle the state, winning
assistance for her children's home care and a voice in state policies on this
subject. Marianne Roberts wrote a workbook setting out a program of
Biblical economics and helped start a soup kitchen. The category of
nonroutine behaviors includes, as well, smaller-scale, less heroic actions (for
example, Gallucci's reading habits) as well as actions that are not necessarily
admirable (for example, Gallucci's pranks that could have hurt other people
at work). Deiinitionally, all that matters is that they are linked to a partly
integrating self image, strong feelings, and are not the course of least
resistance or readily observed given structures and practices in place.
To summarize this section, cultural messages are of different sorts, with
different psychological effects, and different consequences for people's
Methodological implications
Anthropologists have known for a long time that what people say cannot
always be taken at face value (see, for example, Malinowski 1961 [1922]).
More recently, there has been renewed attention to the way discourses are
constructed through interactions between researchers and interviewees
(e.g., Briggs 1986). Sometimes this awareness is taken to extremes and it is
implied that everything an interviewee says is deliberately fashioned for a
strategic end such as making a momentarily useful claim of group identifi-
cation. This approach to discourse analysis overlooks the extent to which
people's talk (as well as less spontaneous cultural texts) is shaped by a
variety of considerations: not only the momentary conscious ones but also
less conscious intuitions about what is interesting, funny, normal, and
right. As I suggested with the distinctions I have made between lip-service
motivation, routine motivation (taken-for-granted and not-taken-for-
granted), and nonroutine motivation, some goals or understandings are
assumed to be widely shared as ideals, others are assumed to be widely
shared as norms, others are in fact norms and are so deeply sedimented
that they are not thought about at all, and still others are recognized as
outside the norm or contrary to widely shared ideals. These differing sorts
of beliefs take on different guises in people's talk. (See also chapter 6 for
examples from Naomi Quinn's research of the way some aspects of speech,
such as metaphors and reasoning, are culturally revealing regardless of the
ways interviewees and interviewers interact, because they are automati-
cally produced rather than deliberately thought out.)
242 Practice and possibilities
To begin with, it is true that at least some of the form and content of
discourse in any context will be oriented (as Bakhtin 1981 pointed out) to
answering, anticipating, identifying oneself with or distancing oneself from
one's interlocutor. Two of the quotes used earlier in this chapter could be
expanded and examined more carefully from this perspective:
CS: What things keep people from getting ahead in the world?
D'Abrosca: [del. 2 lines] Like me, I'm not cut out for business. I couldn't charge a
guy ten bucks for something I paid five for. You know. And . . . There's a certain
makeup there you have to be, to be a businessman. I did the best I could with my
education and knowledge and skills. I'm not rich, I'm not poor. I'm happy with
what I have. I've got a home and the kids have grown up pretty good. [85,6:15]
Collins [recollecting a conversation with a plant manager at Ciba-Geigy] He said,
"You stupe," he said, "You could have been a boss there." I said, "Well, I chose to
do what I'm doing." I said, "Maybe in later life," I said, "It might haunt me."
[laughs] [del. 2 lines] I look back and I'm not mad at myself. I'm happy that I, you
know, didn't take that type of job instead of what I had to contribute to my fellow
workers and the union and international. [85,2:12]
As I noted earlier, in these passages Collins and D'Abrosca seem to be trying
to justify their life choices. Both sound defensive, seemingly aware that I and
perhaps others might be critical of their choices. (In Collins's case, clearly the
plant manager was.) It could well be that neither man is quite as satisfied with
the choices he made as he says he is, and each is constructing post facto
rationalizations. This may have been particularly true if they saw me not as a
student (which I was at the time) but as the professional I was aspiring to be
or as the wife of a university professor, which some interviewees knew me to
be despite my efforts to downplay this role. Another Ciba-Geigy worker
apologized at one point, after repeating criticisms of "the rich" that he had
made frequently throughout the course of the six interviews, "You probably
got big money, I shouldn't talk like that."
Yet, as I have already noted, even their deliberate justifications carry
cultural information: they indicate that D'Abrosca and Collins both know
that the dominant opinion is that they should have tried to get ahead more
than they did. In fact, I suspect they were more concerned about other
people's opinions than mine; both men had made other statements during
the course of the interviews similar to the ones I quoted here, suggesting
these thoughts had been expressed before in other contexts and had
become what I have called "verbal molecules" (oft-heard or oft-expressed
ideas that become internalized as relatively frozen verbal formulae, Strauss
More interesting still is the way they conducted their defense. Note that
D'Abrosca said, "I did the best I could with my education and knowledge
Research on cultural discontinuities 243
and skills" and Collins explained, "I chose to do what I'm doing." These
are individualistic explanations: each of us is different (has different knowl-
edge, education, and skills). All we can be expected to do is choose
thoughtfully, try hard to realize the goals we have chosen, and take
responsibility for our own decisions in the end. In another society they
might have said, "It's not my fate to be a boss" or "My family did not want
me to go into business." Those would not be appropriate ways to explain
one's actions in the United States, however; here individualistic explana-
tions are the taken-for-granted norm (even if people's actions in the United
States are often in fact motivated by concerns for others, which they seem
to have been, at least in part, in D'Abrosca and Collins's cases).
From the structure of Collins's and D'Abrosca's discourse, moreover, it
appears that making one's own decisions and doing one's best are less
contested, more taken-for-granted, values than getting ahead. As a general
rule, if someone defends action A on the authority of principle P, it follows
that they assume P is less controversial (or at least no more controversial)
than A. The same, then, can be concluded for the remainder of each man's
defense. In D'Abrosca's case that was, 'Tvegot a home and the kids have
grown up pretty good." For Collins it was, "I'm happy that I, you know,
didn't take that type of job instead of what I had to contribute to my fellow
workers and the union and international." This suggests that they assumed
that anyone (or, at least, I) would agree that success as a breadwinner and
a father or contribution to others and a larger cause are good values -
better ones than constantly trying to get ahead on the job more than is
necessary to support one's family.
As I noted earlier, what is left largely unsaid is also culturally revealing
(Hutchins 1980; Quinn and Holland 1987). Consider, for example,
D'Abrosca's statement, "I've got a home." Nowadays, for someone who is
very poor, that could mean, "I'm not homeless." In the context of the
conversation with D'Abrosca, however, it did not need to be said that he
meant, "I own a home instead of renting." At D'Abrosca's income level it
is assumed that one can afford some dwelling place; the question is whether
one can afford to own a house. And the value of owning a house is so well
understood that it needs no elaboration.
Thus analysis of discourse elicited in interviews or any other context can
tell us what values speakers believe to be widely shared, what they take for
granted, and the degree of sedimentation of cultural understandings, that
is, the degree to which they are taken for granted. Although Bourdieu
(1977:168) contrasts cultural understandings that are so taken-for-granted
as to be unrecognized (doxa) with those that are subjects of debate
(dogma), there is really a range here from the highly controversial through
the somewhat controversial (ideas about which it is recognized well-mean-
244 Practice and possibilities
ing people can disagree), widely shared (but still seen as values), and so
taken-for-granted as to be invisible.
Similar analyses could be conducted of the unstated assumptions behind
mass media and other public cultural productions. For example, when the
Disney children's film, Pocahontas, was released, it garnered praise for
being "the first Disney feature to crusade for the environment and for
racial harmony" (The Providence Journal-Bulletin 6/23/95:E3). Perhaps
reacting to criticisms of the stereotyped portrayal of Middle Easterners in
Aladdin, released just two-and-a-half years earlier,
Disney Studios very
deliberately made a movie in which the Native Americans are almost all
good and the English settlers almost all bad. (They also hired Native
American actors, Russell Means and Irene Bedard, to speak the parts of
Powhatan and Pocahontas.) At present, attitudes about cultural differ-
ences are highly contestable in the United States,
which is obvious if
Pocahontas is analyzed alongside contemporary commentary on it.
Undiscussed in any public commentary I have seen, however, was that
Disney had changed a story in which one adult has close friendships or
erotic relationships with two other people to a single (heterosexual) rela-
tionship. Commentators on Pocahontas pointed out that the Disney ver-
sion is not historically accurate because the real Indian princess would
have been only about twelve when she met John Smith, not a voluptuous
young woman. Some noted, too, that Pocahontas later married a different
Englishman, John Rolfe; there is no historical information about a love
affair between her and John Smith (who may, in fact, have invented the
story that Pocahontas placed her head upon his to save him from being
killed by her father, the chief, Powhatan). The facts of the matter are less
important than the legend, which at the time the film was made linked
Pocahontas to two English men: John Smith and John Rolfe. The Disney
film simplified this, making John Smith both the rescuee and the love
interest for Pocahontas.
The 1989 Disney film, The little mermaid, reworked the original Hans
Christian Andersen tale in a similar fashion. In Andersen's story the
mermaid kills herself, letting the prince marry another beautiful, sweet
young woman. In return the mermaid is given the chance to win an
immortal soul. In the Disney version the rival love interest is an evil witch
disguised as a beautiful maiden, and the mermaid gets her man in the end
(life in the hereafter not coming up as a possible alternative at all).
Commentators on the The little mermaid mentioned, approvingly, that
Disney was starting to give us movies with strong female characters
(probably a deliberate gesture toward feminism by the studio). I am not
aware of any discussion of the omission of the religious theme or the
simplifying assumption that each of us has only one true love, rendering all
Research on cultural discontinuities 245
other potentially erotic relationships suspect. This seems to be, at present,
a deeply sedimented understanding in the United States. In the previous
chapter Naomi Quinn discussed this assumption and suggested that it
derives its motivating force from an unconscious desire to recreate exclus-
ive mother-child bonds.
Insightful cultural analyses can be made of either public culture or of
texts collected in person-centered research. In both cases, there is some
pandering to the audience and deliberate attempts to make a point which
may not be the producer's deepest or only understanding. Still, with either
sort of production, it is of interest to note what controversial points the
producers make, what sedimented assumptions they express, and every-
thing in between.
Analysis 3. Disjunctives between shared understandings and
public culture
After having discussed some of the parallels between public culture and
culture in persons in the last section, it is necessary to return to the point
that in other respects there are disjunctures between these realms - not
only in technologically complex and socially differentiated societies like
the United States, but even in the technologically simpler societies anthro-
pologists often study.
Barth (1987), for example, studied variations in religious practices and
meanings among several neighboring Ok communities. These variations
were surprising, given great similarities in their technologies, economies
(based on raising taro and pigs), and languages. A key factor complicating
the production of "public" rituals in this region is that many rituals are not
open to everyone in the public: men's initiation rites at each level are open
only to males who have passed through the previous levels, Barth gives the
example of a ninth-step initiation rite, performed about once a decade in a
particular community. Among the Mountain Ok communities where
Barth did his fieldwork, there was no practice of memorizing a list of
procedures nor were the procedures obvious from other rituals. Instead,
the senior man in charge of the ritual has to recreate it on the basis of a
hazy memory of his own initiation, attendance at the rare performance of
it in his own and neighboring communities, and discussion with the
handful of other senior men at his level. The result is that the initiator gives
salience to imagery (e.g., gender and sexual imagery) that is particularly
meaningful from associations learned in his own life. Not that the initiator
has free rein: he is aware that his performance will be observed by other
senior men and needs to be approved by them. Others in his audience are
the novices, and Barth notes that the initiator is "sensitive to their moods
246 Practice and possibilities
and reactions" (1987:44). Within these constraints (including, above all,
his cultural intuitions about what is appropriate), he can introduce innova-
tions that are more reflective of his psyche than the "collective conscious-
This example demonstrates that a mirror metaphor (mental structures
reflect social practice, which reflects mental structures, which reflects prior
practice, and so on) oversimplifies the relation between public culture and
culture in persons. (Aside from being bi-directional it is no improvement
on the "fax" model of the relation between public culture and culture in
persons I have criticized in the past, Strauss 1992a, and that we discussed
in chapter 2.)
As I suggested at the end of the first section, a better
analogy is a conversation. As people express themselves they add to the
pool of public voices to which they and others respond and from which
they learn.
A US example of cultural production: "welfare reform"
Every public cultural production, whether a television show, home appli-
ance, or political speech, is the product of multiple subjectivities: it has to
be meaningful to the producers, acceptable to their backers, and judged by
the producers and backers as likely to be appealing to the potential
audience (whom they may know only superficially, or via the interpreta-
tion of another group of intermediaries, such as market researchers or
pollsters). "Popular" approval (which may, in fact, be the approval of only
a narrow segment of the potential public) then influences the formulae
culture producers use in future productions. Meanwhile, those cultural
productions available to a wide audience partially (but not completely)
shape the attentive public's understandings. For example, while Holly-
wood movies like the Disney films I discussed in the last section do reflect,
in part, widely shared US cultural understandings, their content is also
shaped by idiosyncrasies of their director's personality as well as by what
financial backers think will sell, based on what has sold in the past to the
largest movie-going segments of the domestic and financially important
overseas markets. Catering to these markets, for example, may inflate the
number of "action" films produced in the United States, since the largest
share of US moviegoers are young people (Klady 1995) and low-dialogue
action films do well overseas (Medved 1992, cited in Wasser 1995). (See
also Slotkin 1984, summmarized in chapter 5, n. 6, for a discussion of the
ways Hollywood movies do not just respond to audience tastes.)
Another example of this complicated process is the changes that were
made to state and federal welfare policies during the 1990s in the United
States, culminating in the Personal Responsibility and Work Opportunity
Research on cultural discontinuities 247
Reconciliation Act, signed into law in August 1996. This change in the
federal law ended the policy, in place since the Depression, that every
qualifying family with dependent children would receive financial assist-
ance (welfare) if they fell into poverty, no matter how many other families
had applied for help before them. In the year before the bill was signed,
welfare reform was one of the most talked about public policy concerns in
the United States, with Democratic and Republican governors alike rush-
ing to change their state welfare systems. What accounts for all of this
interest in reducing assistance that (in 1993) consumed only 1.1 per cent of
the federal budget and on the average only 2 per cent of state money (New
York Times, 6/19/94, Sec.4:4; Karger and Stoesz 1994:256-7)?
A simplistic analysis of these policy changes would be that the public
wanted them. Indeed, historians (Katz 1989) and pollsters (Smith 1987)
have found long-standing negative responses to "welfare" and "people on
welfare," and the welfare system had very few defenders among the twenty
people I have interviewed so far.
This is not surprising, given beliefs in
equality of opportunity, hard work, economic independence, small gov-
ernment, low taxes, and the injustice of targeted rather than universal
government programs, beliefs that are widely shared (although with differ-
ent emphases among different social groups, Strauss 1993) in the United
States. (For some US Americans, opposition to welfare is also fueled by
concerns about out-of-wedlock births, suspicion of independent women,
opposition to immigration, or racism - along with the incorrect belief that
most welfare recipients are immigrants and people of color.) Part of the
reason, then, politicians in the 1990s called for restructuring welfare is that
they knew the public would approve of this: it was a guaranteed way to
attract or keep votes.
This explanation is too simple, however. It misses the fact that while
welfare reform in the abstract is popular, the version of welfare reform
enacted in the federal Personal Responsibility and Work Opportunity Act,
which establishes time limits for receiving financial aid but does not
guarantee child care, medical insurance, or a job for recipients who are
trying to leave welfare, captures only a part of popular American senti-
ment on this issue. The survey I conducted in January 1995, discussed
earlier in this chapter, shows this. Among my North Carolina and Rhode
Island respondents, while imposing a two-year time limit was the most
popular proposal, there was almost as much support for replacing welfare
with "a system of government-subsidized child care, health insurance, and
jobs for all Americans who need them." Government-provided jobs,
health care, and child care met the approval of majorities of my samples in
both North Carolina and Rhode Island, as well as majorities of both men
and women, blacks and whites, respondents at every age level, and people
248 Practice and possibilities
at all income levels except those with a household income of over $75,000.1
should note, too, that my question about two-year time limits left open
whether a job would be provided at the end of that period; some of my
respondents indicated they assumed that it would.
The findings from my small, two-state survey match national survey
findings. In a December 1994 Yankelovich poll, for example, majorities of
the Americans sampled were opposed to cutting benefits for unmarried
teenage mothers and ending benefits after two years if recipients have no
other source of income (Time 12/19/94:32). In a May 1994 Yankelovich
poll 74 per cent of the respondents agreed with the proposal, "Replace
welfare with a system of guaranteed public jobs" (Time 6/20/94:26). Final-
ly, a 1987 review often surveys conducted from the late 1960s through the
mid-1980s revealed that changing the wording of a question from some-
thing like "Should the government to do more for people on welfare?" to
"Should the government do more for the poor?" altered the results con-
siderably: in every case but one, this change in wording raised approval of
increased government spending from a minority to a majority of those
polled (Smith 1987:76).
My analysis of these findings is that most Americans would like welfare
recipients to be working but do not want to see anyone left destitute and
recognize that there may not be enough private-sector jobs to go around
(especially jobs that pay enough to cover the cost of health insurance and
child care, in addition to other living expenses). A logical solution is
government provision of jobs and these other benefits, if they are not
available in the private sector, to someone who would otherwise be on
welfare. A majority of US Americans support this solution, despite their
tendency to favor limited government programs, generally. (Other survey
research shows that US Americans often like the sound of rhetoric about
small government, but in fact approve of most specific government pro-
grams, Free and Cantril 1967 and Gans 1988). It seems that the value of
self-reliance is much more strongly held by the public than concerns about
the size of government (as our analysis in chapter 4 of the psychological
roots of belief in self-reliance might lead one to expect).
Despite the popularity of expanded government jobs programs among
the public, however, they have not enjoyed much support by political
leaders in the United States in recent years. (In the summer of 1994
President Clinton's welfare reform task force called for expanded jobs
programs, but his public discussion of that provision remained muted.
If appealing to a majority of potential voters was all politicians cared
about, this would be hard to explain. However, appealing to a majority of
potential voters is not all that politicians care about.
First, politicians are understandably more concerned about appealing
Research on cultural discontinuities 249
to the sorts of potential voters who typically do vote than those who
typically do not. Low-income adults are less likely to vote than their
higher-income counterparts, so their interests are given much less weight.
As former New York governor Mario Cuomo pointed out in criticizing
politicians' cynical welfare scapegoating: "The poor people don't have
any power. That's why welfare's such a terrific issue. Who's going to march
against you...?'" (quoted in Rank 1994:233).
Furthermore, while middle-income voters favor many of the same pro-
grams as poor voters, politicians may not know this, being guided by
conventional wisdom or superficial studies of what the "silent majority"
(recently called the "angry white male") wants. For example, analyses of
the Republican victories in the 1994 midterm elections typically focused on
white males' preference for Republican candidates, ignoring the fact that
voters who felt their standard of living was getting worse preferred Repub-
lican candidates by a margin that was much larger than the difference
between men and women.
Also ignored was the fact that 94 per cent of
the members of the Progressive Caucus in Congress, the most left-leaning
legislators, were re-elected (Feldman 1995). These statistics suggest not a
widespread conservatism among the majority of voters (in that midterm
election, themselves only 39 per cent of the voting-age population
) but
rather displeasure at the state of the economy and with ideologically bland
Political leaders' disinterest in public jobs programs surely has other
sources as well. One factor that diminishes their support for these pro-
grams is that, given the present costs of running a political campaign, it
would be foolish to alienate wealthy potential donors. These wealthy
donors usually have no use for public jobs programs, which reduce the
pool of desperate job seekers, thereby increasing labor costs (Abramovitz
1996) and which also could require them to pay higher taxes.
Some of their disinterest may also reflect their own views, typically
shaped by less experience with the difficulties of making a living than their
constituents have. This is true of most opinion leaders (including people in
the media): even those who may have grown up under difficult circumstan-
ces come to earn, by virtue of their positions, an income that insulates them
from the struggles of the average American.
The difference in perspective
that results was illustrated tellingly on one Rush Limbaugh radio broad-
cast. The popular talk show host was responding to a caller who began
with the usual "megadittos" of praise, but went on to ask how Limbaugh
would respond to someone who questioned how former welfare recipients
are going to survive on $200 a week (then just a little over minimum wage).
In one of Limbaugh's rare moments of public discomposure, he did the
math on air: "Two hundred a week is eight hundred a month, times, uh,
250 Practice and possibilities
times, twelve, uh, is going to be," getting the math wrong, "eighty, what,
four hundred dollars a year." As if just realizing that this is, indeed, not
very much (even the correct figure of $9600 is not very much), he quickly
added, "Uh, and there are, I think, opportunities for a little bit more than
that. I don't think everybody is going to end up in those kinds of jobs," and
went on to suggest that former welfare recipients will be able to add to their
incomes by using the entrepreneurial skills they honed defrauding the
welfare system (Limbaugh radio show, 5/18/95). Clearly, he had never
given much thought to the total income earned at a minimum wage job.
To the many Americans working at low-paying jobs, however, such statis-
tics are a fact of life they cannot help knowing.
In the lengthy follow-up interviews I have conducted with some of the
people who participated in my phone survey about welfare, I have found a
similar difference between low-income and higher-income interviewees. So
far, higher-income interviewees, whether liberal or conservative, have
tended to focus on large-scale systemic, moral, or psychological issues,
while lower-income interviewees, both those who had been on welfare
themselves and those who had not, have tended to be more pragmatic,
focusing on the cost of day care, health insurance, transportation prob-
lems, and the like. The following passages illustrate this difference:
[director of nonprofit organization, a self-described liberal] I think that folks on
welfare get a bad rap a lot of times in terms of, you know, it being expressed by a lot
of people that, you know, that they have kind of a I-don't-want-to-work-hard,
I-don't-want-to-get-ofF-of-welfare attitude. You know, What's in it for me? But I
think that that's all of our attitudes. You could take that attitude and say teachers
have that attitude, [del. 2 lines] I think that that attitude's pervasive in society. I
think we all deal with it. I think I have the attitude, I think my wife has the attitude.
[former welfare recipient, responding to CS's question, "Why do you think most
people go on welfare?"] Money. [Spoken as if this is a dumb question.] You got no
job, you got no skills. I went because I had no job or money. My son's father was
gone. I didn't know where I was going to get a job and pay for day care. I didn't
have a car.
Focusing on large-scale versus more mundane, pragmatic considerations
can have important ramifications for the way one thinks about the useful-
ness of government programs in this area. The first approach makes it
seem as if change is difficult to achieve, and, indeed, that interviewee was
pessimistic that meaningful welfare reform was possible. The second inter-
viewee's focus on skills, jobs, day care, and transportation points to
specific forms of intervention that would be helpful to someone like her -
and, in fact, she was able to use the brief period of her welfare support to
finish her college education and now has a job.
None of this is terribly surprising, but it suggests that we need to be very
Research on cultural discontinuities 251
careful about what we conclude about cultural understandings from hasty
analyses of public culture alone.
I hope I do not need to say that I recognize that public discourses do shape
peopled understandings. In my current research, for example, almost
every one of my interviewees has repeated one of the myths about welfare
that are emphasized in television news and other media sources (e.g., that
welfare recipients keep having children to stay on welfare, when in fact
welfare recipients on average have the same number of children as all US
Americans do, New York Times 6/19/94, Sec.4:4).
The point remains,
however, that we should not expect public cultural texts to mirror people's
understandings. The need of public culture producers to please wealthy
backers is a constraint with different effects than the desire of interviewees
to present a positive self-image to an interviewer, for example. We need to
get away from ideas of culture as a single kind of thing, equally reflected in
social discourse and mental models, and move instead toward ideas of
culture as variegated public representations and psychological appropri-
ations of these, interacting in complex ways.
Beyond old oppositions
The model we have presented in this book makes it clear that the centripetal
cultural effects we have reviewed are a contingent product of interaction
between minds and a world shaped a certain way - not an inevitable
functional requirement of social systems or ecology or a product of timeless
mental structures or of the human need to find meaning through socially
given symbol systems (to put it in some of the theoretical terms of the 1950s,
1960s, and 1970s). It follows from our model that it is equally possible for
cultural inputs to result in understandings varying across individuals and
contexts, or for understandings to be learned without the emotional associ-
ations that give them motivational force. Furthermore, with changes in the
circumstances under which people grow up, understandings can undergo
historical change; and with effort, people can alter their habitual responses.
Thus, the centrifugal effects noticed by anthropologists in the 1980s can be
accommodated within a model that also accounts for the centripetal aspects
of culture that were the focus of earlier descriptions. In our formulation, we
do not have to choose between theories that acknowledge actors' intentions
and theories that acknowledge the role of durable, shared cultural schemas:
intentions depend on schemas. These may be widely shared or intracul-
turally variable; long-held or recently invented; thematically repeated or
juxtaposed in odd combinations. Regardless of which of these tendencies -
centripetal or centrifugal - a given researcher or a given research tradition or
a given disciplinary epoch chooses to emphasize, it is no less true that
meanings and the intentions that accompany them arise from schemas in the
minds of individuals. The cognitive theory we have outlined shows not only
why but how.
We have found connectionism a powerful framework for explaining all
of the abovenamed properties of culture and the centripetal and centrifu-
gal tendencies of ea'ch. We know of no other theory of cultural meaning
that approaches this one in its explanatory fullness. Nevertheless we have
been at pains to point out, the cognitive processes dealt with in connection-
ist models do not explain everything. The recognition that no one theory
does, makes one watchful for theoretical conjunctions and syntheses.
Beyond old oppositions 253
Throughout the book we have drawn freely and eclectically on theories
and findings from other subfields of psychology to supplement connection-
ism - from neuropsychology, cognitive psychology, social psychology, the
psychology of perception, psychoanalysis, and other branches of psycho-
logical anthropology. Nor were our cross-disciplinary appropriations
limited to psychological theory and research. We drew upon theory about
power and social movements, studies of public culture, historical and
religious studies, and political survey research, when these shed light on
our own material or extended our own analyses.
Admittedly, our presentations and uses of these various other theories
and findings have been more illustrative than fully worked out. One book
cannot do everything any more than one theory can, and we felt we had
enough to do in presenting our theory and its applications and in persuad-
ing others, especially anthropologists, of its merit. Our hope is that re-
searchers in several fields will be inspired by the points of potential
conjunction and synthesis we have identified between our theory and theirs
to pursue the implications of our approach for their future research. Even
more gratifying would be actual collaboration between cognitive and other
psychological anthropologists and the kinds of psychologists whose work
we have found valuable, and between psychological anthropologists from
our field and practitioners of other fields of anthropology.
Nor, we have said, does our theory encompass public culture or explain
the shape it takes. It is no more a complete theory of culture than of the
individual. We have not intended it to be, and no reader should close this
book still thinking that we did. Our theory of how cultural meaning is
internalized in individuals does not stand alone, but neither does a theory
of public culture. As we have argued, both intrapersonal and extrapersonal
processes are needed to make a whole theory of culture. Our account is a
necessary part of any such theory.
Why has the intrapersonal, internalized, meaning-making side of culture
been so overlooked in dominant anthropological discourse? A chief im-
pediment to recognizing its separate explanatory status has been the view
that meaning lies in signs or in the relations among them. Meanings, as we
have been saying, are bestowed by the users of signs. They are in people -
not in things or in some nebulous space between. In large part, the failure to
see this point falls naturally from the dualism opposing the collective to the
individual that is part of our intellectual inheritance (Bloch 1985), and it is
fixed in the modern disciplinary boundaries that define the turf that anthro-
pologists and sociologists defend from incursion by psychologists and biol-
ogists. In the recent past of our discipline the territorial boundary marking
the fault-line of this ancient dualism reemerged not only in Geertz's (1973e)
interpretivism but more broadly in the symbolic anthropology associated
254 Practice and possibilities
with him and other influential spokesmen like David Schneider (1968).
The position they developed still supplies the curious terms of present-day
antipsychologism in anthropology. Culture was defined as meaning, but
meaning, as we detailed in chapter 2, was deprived of its ordinary associ-
ation with what things mean to individuals, and restricted to those mean-
ings that can be "read" in public enactments and artifacts.
If this summary intellectual history and the fuller one we gave in chapter
2 make anthropology's antipsychologism seem merely wrongheaded, but
entirely innocent, that is an impression we should correct. What we called
"overlooking" in the last paragraph also took the form of active "eras-
ure." That this was so is exemplified as well as preserved in the following
passage Geertz wrote in "Thick description" in 1973:
The main source of theoretical muddlement in contemporary anthropology is a
view which . . . is right now very widely held - namely, that, to quote Ward
Goodenough, perhaps its leading proponent, "culture [is located] in the minds and
hearts of men."
Variously called ethnoscience, componential analysis, or cognitive anthropology
(a terminological wavering which reflects a deeper uncertainty), this school of
thought holds that culture is composed of psychological structures by means of
which individuals or groups of individuals guide their behavior. "A society's
culture," to quote Goodenough again, this time in a passage which has become the
locus classicus of a whole movement, "consists of whatever it is one has to know or
believe in order to operate in a manner acceptable to its members." And from this
view of what culture is follows a view, equally assured, of what describing it is - the
writing out of systematic rules, an ethnographic algorithm, which, if followed, would
make it possible so to operate, to pass (physical appearance aside) for a native. In
such a way, extreme subjectivism is married to extreme formalism, with the expected
result: an explosion of debates as to whether particular analyses (which come in the
form of taxonomies, paradigms, tables, trees, and other ingenuities) reflect what the
natives "really" think or are merely clever simulations, logically equivalent but
substantively different, of what they think. (Geertz 1973:11)
This published attack contributed to the banishment of not only the
cognitive anthropology of that time but all of psychologizing from main-
stream American anthropology for a long time. That this erasure of
psychological anthropology was successful can be seen in the struggle we
have nowadays getting colleagues in our own department to see the need to
include anything about it, or about the central place it has occupied in the
history of American anthropology, in core graduate and undergraduate
courses on anthropological theory and history.
We, of course, agree with Geertz's reaction to Goodenough's definition
of culture in one respect: as we have been saying throughout this book,
culture is not only what is in people's minds. But we find Geertz's argument
faulty in another respect. Many contemporary readers, lulled by the com-
Beyond old oppositions 255
bination of authoritative voice and heavy derision he used in this passage
and Geertz's inimitable, Latin-studded prose, may have missed his error,
but we think the subsequent history of cognitive anthropology has made it
plain: no particular view of how to describe the mental or intrapersonal
side of culture follows necessarily, as Geertz claimed it does, from the view
that culture is internalized. Rather, the methods of ethnoscience to which
Geertz alluded originated in a particular theory of meaning as being in the
contrasts among related words - a view derived, as we noted in chapter 1,
from the structural linguistics of de Saussure and as peculiar in its way as
Geertz's own that we critiqued in that chapter. This was a theory of word
meaning against which Charles Fillmore (1975), drawing on schema theor-
etical developments in cognitive science, was soon to provide a powerful
critique. It is a theory of meaning that has largely given way to schema
theory in several related cognitive sciences. In this book we have capital-
ized on this development in cognitive science that has influenced cognitive
anthropology so heavily (see Casson 1983; Quinh and Holland 1987), and
that has led us to quite a different, and we think more promising, theory of
cultural meaning. Incidentally, too, ours is a theory of cultural meaning
that prompts very different methods - some of them illustrated in chapters
6,7 and 8 of this book - from those to which Geertz, in the passage above,
would have had cognitive anthropology shackled.
Geertz's passage is a good illustration of the neighborhood-razing that
we used in the beginning of chapter 2 in our urban-renewal metaphor for
how each new group of anthropologists goes about eliminating the old.
Academic politics all too often proceeds in this way. Whatever status and
power such tactics gain the "winning" group, is paid for in missed oppor-
tunities for theoretical advance in the discipline as a whole. Theory ad-
vances most productively by learning and conserving what is most useful
about each approach and collaborating to synthesize these perspectives.
Instead, antipsychologism in anthropology led to the research strategy
of studying cultural meanings through the lens of popular symbols and
ignoring what particular individuals learn from those symbols, how the
meanings of different individuals overlap or diverge, and how changes in
context (both long- and short-term) affect individuals' interpretation of
events. We believe that looking at popular symbols can provide a great
deal of insight only if we do not ignore (or make unwarranted assumptions
about) the intrapersonal reactions they evoke.
A subsequent generation of anthropologists, that which we have assigned
to about the 1980s, critiqued symbolic anthropology soundly for its charac-
terization of culture, and especially "native" cultures, as unchanging and
monolithic. As we noted in chapter 2, the new preoccupation of anthropolo-
gists (whether in cultural studies, contemporary historical materialist
256 Practice and possibilities
approaches, or studies of situated cognition and discourse) is with the
shifting and inconsistent forms that cultural practices can take and with the
strategic considerations that lead individuals to adopt some practices or
identities and abandon others. The irony is that by not questioning sym-
bolic anthropologists' rejection of the intrapersonal, psychological side of
culture, the new generation is left without a convincing theory of either
stability or change. How can actors invent, negotiate, and contest their
cultural worlds if they lack learned, internalized motives and intentions to
do so? As we have pointed out in this book, under some circumstances
motives and intentions can be highly durable; certainly, they do not usually
change from moment to moment. Stated in its most absurd form, this latter
view makes culture change seem like changing clothes.
The implications for people pursuing cultural studies are the same as for
the older generation of interpretivists: do not presume that your readings of
public culture are the same as those of the people you are studying. And for
the postmodernists we need to add: stay alert to the possibility that some
understandings are more shared, stable, consistent, and motivating than
you might guess from the diversity of public discourses. The message for
historical materialists concerned with people's ability to resist coercive
institutions is that we need to remember that resistance, no less than the
practices that are resisted, rests on internalized cultural understandings. It
would be very useful if more researchers would study how people gain
understandings that motivate actions leading to both social reproduction
and social change. Finally, in response to colleagues studying situated
cognition and discourse, we agree that such studies are a valuable corrective
to less situated descriptions of cultural models, but the opposite stance, that
meanings and action are produced solely by the situation, does not follow,
nor does it make sense. Instead, we need to combine our approaches to show
how situation and internalized understandings interact to produce mean-
It is time for us to confront the contradiction in the definition of culture
we have inherited as meaningful, symbolic, signifying, conceptual, idea-
tional, but not to anyone in particular, that has encumbered the thinking
and limited the analyses of so many anthropologists over the last two
decades. It is time to question the idea of culture as invented, negotiated and
contested by human actors strangely lacking in any inner purpose and
motivation behind their creative and transformative efforts. It is time to
heed those who argue that culture is both public and private, both in the
world and in people's minds. By bringing the knowing subject back into
social process we can better account for what we have learned about culture
so far and we can begin to develop a deeper understanding of the problems
that still elude us.
1 Not that all anthropologists ignored these discrepancies. Malinowski (1961
[1922]) was one of many who urged attention to such contradictions, for
example, those between explicit beliefs and practice.
2 See, for example, anthropologists, Gomaroff and Comaroff (1991:17) and
Rosaldo (1989:102), as well as the sociologist, Derae (1995:2).
3 D'Andrade, drawing on a suggestion of George Mandler's, proposes that the
turn to studies of idea systems in British, French, and US anthropology in the
late 1940s and 1950s was influenced by "the tremendous expansion of the
importance of communication and information technology throughout indus-
trial societies by mid-century" (1995:12).
4 Focusing on the "frequently recurring and widely shared components" of
people's interpretations eliminates the possible objection that our definition of
"meaning" lets in too many idiosyncratic irrelevancies. "Dog," uttered at noon
on September 10, 1997, might evoke a different image and set of feelings for
Marida than it did for Martin at 4 p.m. on April 4, 1995, but the modal
components of Marida's, Martin's, Mary's, Marvin's (and so on, for all com-
petent English speakers) meanings over time is an interpretation that would
probably accord well with a dictionary definition and with the "central"
features of our knowledge of a concept that are most important in our mean-
ings (Langacker 1987:136, 159). Langacker posits that semantic centrality
"tends to correlate with the extent to which a specification is conventional,
generic, intrinsic, and characteristic" (ibid.: 159; emphasis in the original). We
talk about this again in chapter 5.
5 See Hannerz (1993) for a different defense of the culture concept and
D'Andrade (1995) for another definition of culture. D'Andrade's latest pub-
lished (1995:146,212) definition of culture is a Tylorian one: "the entire social
heritage of a group." More recently he has proposed that culture be defined as
"shared schemas that have an agreed upon external physical sign" (D'Andrade
1996). While D'Andrade recognizes that there are degrees of externalization,
we think that focusing on the correspondence between schemas and signs is too
much influenced by a linguistic model (taking as its prototype words and the
schemas that give them meaning) and is too restrictive. It leaves out schemas
acquired nonverbally and considered so mundane as not to require an agreed-
upon external sign, such as knowledge of correct posture. When this knowledge
258 Notes to pages 7-9
is enacted it is externalized as ways of walking, standing, sitting, and so on.
However, these actions are not agreed-upon signs (even indexical) of the
corresponding schemas; instead, they are the results or symptoms of the
schemas. Also left out are repressed understandings. In chapter 7 Naomi Quinn
discusses the common practice of US adults using baby talk with their spouses
and argues that this behavior arises from the desire to recreate in adulthood an
infantile state of dependency. These desires, however, are considered improper
for US adults and are defended against, leaving no agreed-upon external sign.
(Again, the baby talk itself is a symptom or result, not a sign of the schema.) In
both cases, we are discussing schemas that are historically and culturally
variable, with significant motivational and behavioral consequences. (Holding
one's body the right way is important for fitting into a group, for example, and
unfulfilled dependency desires in marriage can be a major source of dissatisfac-
tion.) Surely, these schemas are cultural. Yet, the lack of a single externalized
sign to represent these schemas makes them low on D'Andrade's externaliz-
ation scale, hence relatively acultural for him.
6 Peter Hervik and Holly Mathews (personal communication) each brought this
objection to our attention. See also Westen (n.d.).
7 See, for example, Kondo (1990:34-5) and Hutchins (1995:312,355), whose
arguments will be described in greater detail in chapter 2.
8 Freud (1965 [1923]) discussed "internalization" as a special type of learning
restricted to the formation of the child's superego in the oedipal phase. We use
"internalization" much more broadly, as do most psychologists today, to refer
to the psychological effects of experience.
9 See D'Andrade (1995) for a history of cognitive anthropology and Ingham
(1996) for an overview of some current work in psychological anthropology.
10 See also Goodenough (1981:51-54).
11 Each of these terms has a somewhat different connotation. Anthropologists
who do work on "the self' typically study folk psychologies, that is, not selves
so much as cultural models of the self (e.g., Geertz 1983a). Studies of "mean-
ing" are typically Geertzian as well (e.g., Geertz 1973c), focusing on the public
"forms [e.g., key words, rituals, works of art, etc.] through which people make
sense of their lives" (Rosaldo 1989:26). Work on "identity," on the other hand,
tends to focus on ethnic, race, class, gender, sexual orientation and other group
identifications. Typically, such work focuses on deliberate strategies people use
to claim or avoid such identities (e.g., Butler 1990). "Consciousness" is most
often used by Marxists and other critical theorists to refer to awareness (or
nonawareness) on the part of some subaltern group that they indeed are a
group with shared interests and, as a result of systematic social forces, have
suffered unjust treatment, received unequal resources, and the like. "Subjectiv-
ity" refers broadly to the actor's conscious experience (but see Johnson
1986/87:44 for a different definition); anthropologists of "experience" study
subjectivities humanistically and usually acknowledge the influence of Victor
Turner's work (e.g., Turner 1985). Some "reader response" approaches to
literary criticism argue that the meaning of a text depends not on the author's
intentions or content of the text but on the conventional interpretive frames
through which readers read it at a given time (Fish 1980). Others working in
Notes to pages 9-14 259
cultural studies talk about the need to study people's "readings" of texts
without implying that authorial intentions and text contents are irrelevant to
the meaning produced (e.g., Johnson (1986/87). Appadurai (1990:5) expands
on Anderson's (1983) term "imagined communities," using "the imagined" to
refer to collective representations, especially those related to identities and
created by modern media. (This is not to be confused with the "social imagin-
ary," typically a corruption of Lacan's usage in which "the imaginary" refers to
presymbolic fantasies.) Finally, "agency" takes its meaning from the opposi-
tion of "agency" to "structure." It is almost always used to refer to acts the
author sees as admirable resistance to hegemonic social structures and (as
Stanley Kurtz pointed out to us, personal communication) often carries the
unfortunate connotation that this resistance is an act of undetermined free will.
12 We use the term "explain" deliberately. Although we will not devote any
further space to the issue here, we deplore the rejection of explanation in favor
of interpretation in anthropology at present. (See Geertz 1973f:5 and 14 for
statements of the distinction between causal explanation and interpretive de-
scription, along with claims that anthropologists should pursue the latter
instead of the former.) Interpretation is important, but used alone, it leaves us
with tantalizing static descriptions without any understanding of process.
13 In this book we will use "implicit" to refer to psychological contents (such as
schemas or habits) that are normally outside of awareness but face no resis-
tance in coming to consciousness. These we distinguish from "unconscious"
ideas, feelings, and motives, which are repressed.
14 See, for example, Sanday (1990).
15 Sperber has become interested in the modularity of thought, while we draw on
neo-associationist models that are usually opposed to modular models (incor-
rectly, we will argue in chapter 3). Also, while Sperber's distinction between
mental and public representations is close to our distinction between the
intrapersonal and extrapersonal realms, we would not use the term "represen-
tations" for the latter, because not all publicly observable events and processes
have as their primary role the function of representing something for someone,
even if, incidentally, that is usually the case.
1 See, for example, Parsons (1961).
2 Where Geertz himself stood on this is not clear. His "turtles" anecdote
(1973i:29) is often taken to mean that texts are all there are, but he says other
things that explicitly contradict that (1983b:14 or even 1973i:13).
3 Geertz says "indeed" because the rest of this sentence is a close paraphrase of a
statement of Gilbert Ryle's that served as the epigraph for this essay (1973c:55).
4 Geertz went on, in the same essay, to say:
How precisely to accomplish this, how to analyze symbol use as social action and write
thereby an outdoor psychology is, of course, an exceedingly difficult business . . . But
what is clear, if anything is, is that to do so is to attempt to navigate the plural/unific,
product/process paradox by regarding the community as the shop in which thoughts are
constructed and deconstructed, history the terrain they seize and surrender, and to attend
therefore to such muscular matters as the representation of authority, the marking of
260 Notes to pages 14-22
boundaries, the rhetoric of persuasion, the expression of commitment, and the registerine
of dissent. (I983d:153)
The gendered imagery here is striking: "muscular" matters such as "seizing
terrain" for thought in public versus delicate experiences like "wishing or
regretting" for thought in private (instead of tending children for thought in
public and plotting revenge for thought in private). We could jokingly suggest
that this looks like a kind of protest masculinity. Perhaps these muscular
examples were necessary for real men to pay attention to meanings and feelings.
5 Many more examples could be given. See, for example, Geertz's discussion of
the "extrinsic theory of thought" in "Ideology as a cultural system" (originally
published in 1964).
6 See Ewing (1992) for a different response to Geertz on this point and Chafe
(1970) for a reply to similar claims.
7 See D'Andrade (1987a: 113) for a discussion of what it means for mental
structures to be intersubjectively shared.
8 See, for example, Holland and Reeves (1994); Holland and Skinner (1995);
Hutchins (1995); and Lave (1988).
9 In fact, this discussion gives the Beethoven Quartet example greater consistency
than Geertz did. This abbreviated quote omits "a coherent sequence of
modeled sound - in a word, music" (1973i: 11-12), which mixes the concrete
with the abstract. See Goodenough (1981) and Strauss (1992a:6) for a thorough
discussion of this passage.
10 Julie Tetel, personal communication. She was talking about sense for Frege or
this version of meaning for Geertz. See also Goodenough (1981:53-4).
11 Alexander (1987) also discusses inconsistencies in Geertz's henneneutics but of
a different sort: between collective determination and contingency. I do not see
that conflict as having influenced Geertz nearly as much as his struggle to
reconcile studies of meaning with the antimentalism that was dominant during
the critical period of his intellectual development in the late 1950s and early
12 Langacker (1987:5) makes the same claim for semantic theories generally. He
states, "I take it as self-evident that meaning is a cognitive phenomenon and
must eventually be analyzed as such," while noting that a cognitive theory of
meaning "conflicts with major traditions of semantic theory (much of which
can be read as an elaborate attempt to avoid this conclusion)."
13 See Sperber 1996:79-80 for a related argument.
14 See Barth (1993:332); Chafe (1970:74-76); D'Andrade (1984:101-105); and
Spiro (1987a: 162-3) for closely related arguments. Geertz comes close to
stating this himself at one point: "meaning is not intrinsic in the objects, acts,
processes, and so, which bear it, but , . . [is] imposed upon them; and the
explanation of its properties must therefore be sought in that which does the
imposing - men living in society" (1973f:405). The emphasis here, however, is
on the cross-cultural variability of meaning, not its psychological status.
15 Although she (Ortner 1989:60) traces her usage of the schema concept to
different origins than those to which we trace ours (see chapter 3). She points to
her own earlier discussion of "key scenarios" and to ideas coming from the
symbolic anthropology of the 1970s.
16 Roy D'Andrade has pointed out to us (personal communication) that these
Notes to pages 22-29 261
socialization patterns, in turn, might be a response to conflicts over property,
inheritance rights, and so on, as discussed by Barry, Child, and Bacon (1959).
We are not claiming that the explanatory variables Ortner identifies are irrel-
evant, only that they are insufficient.
17 See Ewing (1992) for the recurrence of the cave metaphor in anthropologists*
rejection of psychology.
18 See also the criticisms of Keesing 1987; LeVine 1982; and Wikan 1987, among
others. Keesing (1987:165, n. 4) does note, appropriately, "Geertz at his best
captures the evanescence and negotiatedness of cultural meanings and is able to
go beyond the assumption that those doing the negotiating need share a culture
or construct the same meanings - as witness the scenario of Cohen and the
sheep (Geertz 1973)."
19 A developmental psychologist has made a similar critique of her own field: "To
some extent, social learning theory seems to have a 'copy* theory of knowledge.
A sequence of events in the world is more or less copied as it is represented in
symbols and retained" (Miller 1989:260).
20 See, for example, Clifford (1988c); Clifford and Marcus (1986); Grossberg,
Nelson and Treichler (1992); Marcus and Fischer (1986); and Rosaldo (1989).
21 Postmodernists were not the first to raise these suspicions. See, for example,
Barth (1975); Roberts (1961); Wallace (1970).
22 See also Dirks (1992:75).
23 Those labeled "skeptical postmodernists" by Rosenau (1992).
24 See also Fox (1985) and Spivak (1987:205). Spivak recommends that subaltern
groups make a strategic appeal to a shared identity, even though there is no
shared group essence in fact.
25 See also: "For a long time ordinary individuality - the everyday individuality of
everybody - remained below the threshold of description" (Foucault 1977:191)
and "It appeared that I now had to undertake a third shift, in order to analyze
what is termed 'the subject.' It seemed appropriate to look for the forms and
modalities of the relation to self by which the individual constitutes and
recognizes himself qua subject" (Foucault 1985:6).
26 See also Foucault (1977:155).
27 This argument appears more strongly in The history of sexuality, Volume I than
in earlier works, such as Discipline and punish.
28 Butler draws on Mary Douglas's (1966) discussion of the cultural construction
of taboos about the body, especially those related to bodily boundaries, and
takes Douglas's ideas to suggest "that the naturalized notion of 'the' body is
itself a consequence of taboos that render that body discrete by virtue of its
stable boundaries" (Butler 1990:132-3). However, as Butler notes on the pre-
ceding page, that is not Douglas's suggestion (