Relational Models Theory: A Contemporary Overview 0805839151, 9780805839159, 9781410611413

Table of contents :
Contents......Page 6
List of Contributors......Page 8
Preface......Page 10
1 Relational Models Theory 2.0......Page 20
2 Research on the Relational Models: An Overview......Page 44
3 Four Modes of Constituting Relationships: Consubstantial Assimilation; Space, Magnitude, Time, and Force; Concrete Procedures; Abstract Symbolism......Page 78
4 Social Expertise: Theory of Mind or Theory of Relationships?......Page 164
5 The Domain of Work in Households: A Relational Models Approach......Page 184
6 Hidden Bias: The Impact of Relational Models on Perceptions of Fairness in Human Resource Systems......Page 214
7 Relational Models, "Deonance," and Moral Antipathy Toward the Powerfully Unjust......Page 238
8 Proscribed Forms of Social Cognition: Taboo Trade-offs, Blocked Exchanges, Forbidden Base Rates, and Heretical Counterfactuals......Page 264
9 Values and Emotions in the Relational Models......Page 280
10 The Four Faces of Trust: An Empirical Study of the Nature of Trust in Relational Forms......Page 304
11 Depressed Mood as an Interpersonal Strategy: The Importance of Relational Models......Page 326
12 A Relational Approach to the Personality Disorders......Page 352
B......Page 380
D......Page 381
G......Page 382
H......Page 383
L......Page 384
P......Page 385
S......Page 386
U......Page 387
Z......Page 388
F......Page 390
W......Page 391

Citation preview

RELATIONAL MODELS THEORY A Contemporary Overview

This page intentionally left blank

RELATIONAL MODELS THEORY A Contemporary Overview

Edited by

Nick Haslam University of Melbourne

2004

LAWRENCE ERLBAUM ASSOCIATES, PUBLISHERS Mahwah, New Jersey London

Copyright © 2004 by Lawrence Erlbaum Associates, Inc. All rights reserved. No part of the book may be reproduced in any form, by photostat, microform, retrieval system, or any other means, without the prior written permission of the publisher. Lawrence Erlbaum Associates, Inc., Publishers 10 Industrial Avenue Mahwah, New Jersey 07430

Cover design by Sean Trane Sciarrone

Library of Congress Cataloging-in-Publication Data Relational models theory : a contemporary overview / edited by Nick Haslam. p. cm. Includes bibliographical references and index. ISBN 0-8058-3915-1 (cloth : alk. paper) - ISBN 0-8058-5356-1 (pbk. : alk. paper) 1. Interpersonal relations. 2. Interpersonal communication. I. Haslam, Nick, 1963HM1106.R442 2004 302-dc22

2004046925 CIP

Books published by Lawrence Erlbaum Associates are printed on acid-free paper, and their bindings are chosen for strength and durability. Printed in the United States of America 10 9 8 7 6 5 4 3 2 1

Contents

List of Contributors

vii

Preface

ix

PARTI: FUNDAMENTALS

1

Relational Models Theory 2.0 Alan Page Fiske

2

Research on the Relational Models: An Overview Nick Haslam

3

27

PART II: COGNITION AND CULTURE

3

4

Four Modes of Constituting Relationships: Consubstantial Assimilation; Space, Magnitude, Time, and Force; Concrete Procedures; Abstract Symbolism Alan Page Fiske Social Expertise: Theory of Mind or Theory of Relationships? Nick Haslam and Alan Page Fiske

61

147 V

VJ

CONTENTS

PART III: JUSTICE AND FAIRNESS 5

6

7

The Domain of Work in Households: A Relational Models Approach Jacqueline J. Goodnow

167

Hidden Bias: The Impact of Relational Models on Perceptions of Fairness in Human Resource Systems Debra Louis Connelley and Robert Folger

197

Relational Models, "Deonance," and Moral Antipathy Toward the Powerfully Unjust Robert Folger and Rebecca Butz

221

PART IV: EMOTIONS, VALUES, AND MORALITIES 8

9

10

Proscribed Forms of Social Cognition: Taboo Trade-offs, Blocked Exchanges, Forbidden Base Rates, and Heretical Counterfactuals Philip E. Tetlock, A. Peter McGraw, and Orie V. Kristel Values and Emotions in the Relational Models Sonia Roccas and Clark McCauley The Four Faces of Trust: An Empirical Study of the Nature of Trust in Relational Forms Leah D. Houde, Dana M. Sherman, Tiffany B. White, and Blair H. Sheppard

247

263

287

PART V: RELATIONAL MODELS IN THE CLINIC 11

12

Depressed Mood as an Interpersonal Strategy: The Importance of Relational Models Nicholas B. Allen, Paul Gilbert, and Assaf Semedar A Relational Approach to the Personality Disorders Nick Haslam

309

335

Author Index

363

Subject Index

373

List of Contributors

Nicholas B. Allen

Department of Psychology, University of Melbourne, Australia

Rebecca Butz

Department of Psychology, Tulane University, New Orleans, LA

Debra Louis Connelley

College of Business Administration, Touro University International, Cypress, CA

Alan Page Fiske

Department of Anthropology, University of California, Los Angeles, CA

Robert Folger

Management Department, University of Central Florida, Orlando, FL

Paul Gilbert

Department of Psychology, University of Derby, United Kingdom

Jacqueline J. Goodnow

Emeritus, Department of Psychology, Macquarie University, Australia

Nick Haslam

Department of Psychology, University of Melbourne, Australia

Leah D. Houde

Fuqua School of Business, Duke University, Durham, NC

Orie V. Kristel

The Strategy Team, Ltd., Columbus, OH vii

Viii

LIST OF CONTRIBUTORS

Clark McCauley

Department of Psychology, Bryn Mawr College, Bryn Mawr, PA

A. Peter McGraw

Postdoctoral research associate, Princeton University, Princeton, NJ

Sonia Roccas

Department of Psychology and Education, Open University of Israel, Tel Aviv, Israel

Assaf Semedar

Department of Psychology, University of Melbourne, Australia

Blair H. Sheppard

Fuqua School of Business, Duke University, Durham, NC

Dana M. Sherman

Independent scholar, Scottsdale, AZ

Philip E. Tetlock

Haas School of Business, University of California, Berkeley, CA

Tiffany B. White

College of Business, University of Illinois at Urbana-Champaign, IL

Preface

The theory of relational models (RMs) was first proposed by Alan Fiske in the early 1990s. Its ambitions were high and wide. The theory aspired to be nothing less than an account of the fundamental forms of social relations that might be valuable to social and behavioral scientists of many persuasions. In the ensuing years, RM theory has received widespread attention, not only from psychologists, to whom it has most often been pitched, but also from management scholars, anthropologists, sociologists, and cognitive and political scientists. Many theorists and researchers have put the theory of RMs to work, generating a variety of conceptual links and empirical findings and pursuing a diverse range of methodologies. RMs theory has clearly had a measure of success within the sometimes harsh ecology of social theory. However it has also paid for this success with a kind of intellectual sprawl. Because of its wide-ranging relevance the theory has given rise to a dispersed set of applications, ranging from ethnographic studies in the rainforests of New Guinea and the factories of upstate New York to neuroimaging and priming investigations in the laboratory. Its ramifications have been charted in forbiddingly abstract cognitive science texts and in richly contextualized descriptions of social interaction, in the cool language of evolved cognitive modules and the warm language of emotions and motives. A consequence of this dispersion of RMs scholarship—disciplinary, methodological, and topical—is that it is very difficult for the newcomer or even the established theorist and researcher to grasp the theory's possibilities or to survey existing work. In the years since the theory first apix

X

PREFACE

peared there has been no attempt to draw this work together and to consider what it adds up to and where it might lead. The present volume is intended to remedy this state of affairs. It attempts to review the theory and the research that it has spawned comprehensively, and to lay out in detail some of the many directions in which theory and research are proceeding in the present. The chapter authors are a diverse group, drawn from a variety of disciplines. They have tried mightily not only to present their work, showing the contribution that RMs theory has made to it, but also to speculate on the theory's implications for their fields and the avenues along which it might be developed further. We hope that the chapters collectively demonstrate the theory's fruitfulness, document the state of RMs research as it enters its teens, and spur on researchers and theorists alike to take their work further still, finding new ways to test the theory's usefulness and validity.

A GUIDE TO THE VOLUME

The book is divided into five parts, containing 12 chapters in total. Part I, on "Fundamentals," begins the book with state-of-the-art summaries of RMs theory and research. It is designed to give readers a working grasp of the basic features of the theory as it is now formulated, and an overview of the RMs-related research conducted to date. Fiske's chapter 1 presents the indispensable concepts of RMs theory and describes the intellectual and historical context of the theory's development. Fiske describes how the theory explains social coordination by positing universal models embodied in culturally specific implementations and combinations, and strongly defends the need for a social psychology that addresses relationships rather than simply perceptions of other individuals. He goes on to discuss the issue of innateness and how the RMs might have evolved, before clarifying some existing confusions about the theory and proposing directions for its future development. Haslam's chapter 2 surveys the empirical research base of RMs theory in a way that complements Fiske's theoretical discussion. Research on the structure of the RMs, their influence on everyday social and cognitive processes at individual and collective levels, and their conceptual links to other theories is exhaustively reviewed. The author argues that RMs theory has accumulated extensive support for its structural claims, for its explanatory power, and for its capacity to generate new conceptual links and research programs. He notes also that research on the theory is rather thinly spread and in need of consolidation, and must address the development, crosscultural variation, and concrete implementation of the models in future.

PREFACE

XI

Although Part I lays out fundamentals of RMs theory and research, ideally it will not be bypassed by readers who already have some exposure to this work. Fiske's chapter does not simply reheat the theory as it was first presented, but offers a new version—"Relational models theory 2.0"—that responds to theoretical clarification, exchange, and the changing complexion of social and behavioral science over the past decade. Haslam's chapter provides a broad context for the more detailed presentations in later sections of the book. Part II, on "Cognition and Culture," addresses the contributions that RMs theory makes to the study of social cognition. Fiske's chapter 3 discusses the important and hitherto neglected issue of how relationships are constituted. Supported by an encyclopedic assortment of ethnographic and other illustrations, Fiske claims that each RM has a distinctive natural medium through which people perceive, represent, communicate, and learn about relationships. For example, people have a proclivity to constitute Authority Ranking relationships in terms of an iconic "social physics" of space, time, magnitude and force, and Communal Sharing relationships in terms of bodily connections (e.g., commensalism, blood rituals, body contact). In perhaps his most far-reaching theoretical extension to date, Fiske proposes that each RM has an integrated "conformation system" that governs how it is cognized, enacted, and culturally reproduced. Haslam and Fiske's chapter 4 continues the focus on social cognition in culture, issuing a challenge to recent claims that "theory of mind," the capacity to attribute mental states to others, constitutes the singular foundation of human social competence. The authors argue that such "mindreading" is only one component of our social capability, and that in many respects it is secondary to the sort of relational cognition—thinking about relations between people rather than the contents of other minds—that RMs theory captures. The mentalistic understanding of other individuals, they argue, is neither necessary for complex social interaction and organization nor universally prominent across cultures, and it may be best understood to operate in the service of a more basic form of relational understanding. The enthusiasm with which psychologists have received theory of mind as an encompassing view of social cognition may be symptomatic of its taken-for-granted individualism, which a relational perspective strives to correct. Part III, "Justice and Fairness," collects three chapters that address justice and fairness, moving away from the more abstract and intrapsychic concerns of the previous section into discussions of discord and negotiation in particular social contexts. Goodnow's chapter 5 opens a revealing window onto the meanings and tensions surrounding household work within families, complete with compellingly drawn examples. She argues

xii

PREFACE

that the RMs offer an integrative account of this domain, whereas many existing approaches to the field tacitly proceed within the terms of one model. Goodnow's chapter is notable for the serious attention it pays to the discord that arises when family members apply or wish to apply discrepant models to their work, a realistically warts-and-all portrait that shows how interpersonal life is not all smooth coordination but involves negotiation and conflict over how relationship are to be configured. Just as important is Goodnow's demonstration and insistence that the putting into practice of RMs depends crucially on settings and situations, and how necessary the concept of "implementation rules" is in this regard. A final virtue of this chapter is that it shows how much developmental considerations enter into the relational framing of household work, and how development remains largely neglected by RMs researchers. Connelley and Folger's chapter 6 presents another valuable demonstration of the interpersonal tensions that arise when people bring discrepant RMs to bear on their interactions, this time in an organizational context. The chapter reports on a study of a conservative workplace's struggles with internal diversity, and how different groups of workers interpret the organizational culture in starkly different ways based on distinct RMs. The traditional core of White executives understands the nature of their relations to one another and to the company in a communal fashion, which conflicts with the egalitarian expectations of African-American workers and the meritocratic or equity-based preferences of many White women. The result, well captured in quotes from protagonists, is a picture of disharmony and a lesson to managers. Connelley and Folger clearly show how relational discord can be grounded in intergroup differences, and raise the intriguing possibility that the RMs reflect, in part, discursive resources that can be deployed by people in ways that advance the interests of their social position. Folger and Butz's chapter 7 presents an innovative relational analysis of responses to injustice, focusing on the reactions of third parties to the abusive exercise of authority. They argue that these reactions are powerful emotional tendencies that must be understood in terms of Fiske's Authority Ranking model. Moral antipathy toward abusive authority has evolutionary roots, they suggest, and people are willing to sacrifice their self-interest to punish those who trigger it. Moral outrage such as this has an intrinsic hierarchical component, they speculate, and serves to ensure that power in organizational hierarchies is exercised in legitimate and circumscribed ways. This chapter reinforces the previous chapter's demonstration of how RMs theory can illuminate organizational phenomena, shows how it can be joined to evolutionary theorizing, and presents a good bridge to the next section of the book. Part IV, on "Emotions, Values, and Moralities," contains three chapters that show how RMs theory can illuminate affective and evaluative proc-

PREFACE

XIII

esses. Although it is often presented as a rather formalistic and abstractly cognitive approach, the theory was always intended to account for the warm-blooded dimensions of interpersonal life. Whereas Part HI focused attention on collective phenomena in organizations and families, the chapters in this section pay more heed to individuals' experiences of outrage, upset, and retaliation, and their distinctive motivational processes. Tetlock, McGraw, and Kristel's chapter 8 proceeds from an analysis of consistent violations of normative models of judgment. Rather than being intuitive scientists and economists, as normative models imagine, people often act as intuitive theologians, "trying to protect sacred values from secular encroachment." Our moral value commitments make certain kinds of "rational" judgment taboo, inviting moral outrage when breached. RMs theory helps to explain the sharp boundaries that acceptable forms of social judgment cannot cross, as it defines different relational grammars as incommensurable. To put monetary values on communal family bonds, to seek to buy body parts or to sell a gift from a loved one, or to judge individuals on the basis of attributes disproportionately associated with their social group when this conflicts with egalitarian efforts at reform, is to appear depraved and morally contaminated. Tetlock and his colleagues back their original theorizing with some path-breaking empirical work, which has real implications for the framing and presentation of public policy. Whereas Tetlock and McGraw discussed particular ways in which values linked with particular RMs conflict and elicit moral outrage, Roccas and McCauley's chapter 9 offers a broader mapping of links between RMs, values, and emotions. The authors argue for a variety of associations between the models and a well-established value taxonomy, so that people who "prefer" certain models tend to hold certain broad motivational goals. People choose to engage in relationships of particular sorts because they are compatible with their values, they suggest, and relational experiences also shape personal value orientations. Holding value priorities that are incompatible with an operating RM will tend to produce interpersonal discord and intrapersonal distress. Such mismatches, where behaviors or motives violate relational expectations, may have distinctive affective consequences depending on the RM involved. Roccas and McCauley propose that different moral codes may also be implicated in these emotionally fraught violations. Chapter 10, by Houde, Sherman, White, and Sheppard, extends its section's focus into the domain of interpersonal trust. Trust comes in different forms, they propose, according to the nature of the relationship involved, such as its degree of depth and interdependence, because different relationships carry different degrees and kinds of risks. They present some empirical evidence to show that people consider different risks to be relevant to different types of relationship, so that the RMs imply different varieties

XJV

PREFACE

of and bases for trust and perceptions of trustworthiness. This formulation indicates the problems of viewing trust as a unitary phenomenon, and encourages an understanding of it that is intrinsically relational rather than being grounded in personal attributes or broad contextual features, as in most contemporary research. Part V, "Relational Models in the Clinic," concludes the volume with two chapters that examine how the RMs help to make sense of psychological disturbance and its treatment. Allen, Gilbert, and Semadar's chapter 11 offers a compelling analysis of depression—the psychological "common cold" of modernity—and in the process demonstrate how RMs theory can contribute to the emerging field of evolutionary psychiatry. They argue that depression, and the causal influences that bear on it, can be understood in relational terms. Although the interpersonal basis of depression is increasingly acknowledged, they note that many interpersonal accounts are framed within a single model. They argue instead that several RMs help to illuminate and explain many features of depressed behavior and cognition (e.g., vigilance for signs of social exclusion and signaling to secure social support), and propose an ambitious "social risk" theory of depressed states. Their theory makes several predictions about how depressed individuals will behave in different relational contexts, which should stimulate an exciting program of research. Haslam's chapter 12 closes the volume with a relational analysis of personality disorders, enduring conditions that are poorly understood but cry out for some form of interpersonal understanding. He argues that many of these conditions can be conceptualized as reflecting inflexible biases in the implementation of particular RMs. These biases generate predictable forms of interpersonal disruption because they lead disordered individuals to construe relationships in ways that are discrepant with their interactants. Such discrepancies have been discussed in intrafamilial and intergroup contexts in earlier chapters by Goodnow and Connelley and Folger, but may also reflect individuals' interpersonal propensities. Empirical support for this relational account of personality disorders is presented, and the account is compared with some alternative interpersonal theories. Implications for discrepant relational construals in therapy—one of the more complicated forms of interpersonal encounter—are also discussed. These two chapters show how RMs theory enables new understandings of abnormal psychological phenomena, and can serve as an innovative alternative to the more well-trodden interpersonal models that are currently employed. They also suggest that although the theory has generally been perceived as an account of our shared social-cognitive endowment, it can also make sense of how people differ, whether in adaptive or maladaptive ways. Individual differences in relational tendencies represent a research area in which more work is clearly required.

PREFACE

XV

Together, I hope, the book's 12 chapters present the interested reader with a snapshot of a theory that is very much in motion, and that illuminates a wide variety of phenomena. On the book's evidence, it seems clear that the next decade will see a continuing extension and development of RMs research. —Nick Haslam

This page intentionally left blank

P A R T

I FUNDAMENTALS

This page intentionally left blank

C H A P T E R

1 Relational Models Theory 2.0 Alan Page Fiske

Relational models theory is simple: People relate to each other in just four ways.1 Interaction can be structured with respect to (1) what people have in common, (2) ordered differences, (3) additive imbalances, or (4) ratios. When people focus on what they have in common, they are using a model we call Communal Sharing. When people construct some aspect of an interaction in terms of ordered differences, the model is Authority Ranking. When people attend to additive imbalances, they are framing the interaction in terms of the Equality Matching model. When they coordinate their actions according to proportions or rates, the model is Market Pricing. Everyone uses this repertoire of relational capacities to plan and to generate their own action; to understand, remember, and anticipate others; to coordinate the joint production of collective action and institutions; and to evaluate their own and others' action. In different cultures, people use these four relational models in different ways, in different contexts, and in differing degrees. In short, four innate, open-ended relational structures, completed by congruent socially transmitted complements, structure most social action, thought, and motivation. That's the theory. These are the four fundamental, innate, human relational proclivities. To signify that they are cognitively modular but modifiable modes of interacting, I call them "mods." However, these open-ended generative potentials are insufficient in themselves to determine action or evaluation, or permit 'I thank Nick Haslam, Barbara Fiske, and Rick Wicks for helpful comments on this chapter.

3

4

FISKE

coordination. In order to use these mods to act or to interpret others' action, people need socially transmitted prototypes, precedents, and principles that complete the mods, specifying how and when and with respect to whom the mods apply. I use the term "preo" to signify the class of paradigms, parameters, precepts, prescriptions, propositions, and proscriptions that can be conjoined with mods. A mod must be conjoined with a preo that complements it to generate a specific cultural coordination device. This theory was first set out in primitive form in a 1990 article, and then developed fully in my 1991 book, Structures of Social Life, and an article that summarized the core ideas (Fiske, 1992). I expected then that the theory might have to be revised considerably once others had the opportunity to analyze it, and certainly once it was empirically tested. Surprisingly, discussions and extensive tests of the theory have not indicated any need for substantial revision (yet!). However, they have facilitated the development of a somewhat clearer, perhaps more cogent formulation of the theory, which this chapter presents.

THE FOUR ELEMENTARY RELATIONAL STRUCTURES The Communal Sharing (CS) mod bases sociality on the perception that a set of persons have something in common—something that makes them socially equivalent in some respect. (Typically, people are aware that the set is composed of individuals and subsets who are different in other respects.) The set may include the self, or may be a set of others who are treated as functionally equivalent for certain purposes. To organize any given aspect of a relationship in any given context, people must complete this mod with preos that define what who has in common with whom: "blood" kinship defined by descent through the mother, father, or both; blood consumed in an act of ritual brotherhood; a home and economic or parental responsibilities; religion; employment by a corporation; attendance at a school or college; membership on the same team or being fans of one; being followers of the same leader; nationality, birthplace, or culturally defined "ethnicity"; joint responsibility for a task or group decision; a shared resource such as a home, granary, commons, park, or road system; collective moral responsibility either for each other's acts or for each other's welfare; a type of misfortune or problem (breast cancer, alcoholism, an earthquake, epidemic, or terrorist strike); membership in a secret society; an age-set defined by initiation; or simply the sentiment of belonging together, with a shared past and future. The mod may structure the interaction of a dyad, a group, or an "imagined community" (Anderson, 1983) who never all assemble together and do not know the identities of all the others. People may be motivated

1. RELATIONAL MODELS THEORY 2.0

5

by what they see themselves as having in common with living humans, deceased persons, ancestors, spirits, gods, pets, domestic animals, or other social beings. Everyone has Communal Sharing relations with many primary or reference groups, as well as CS relationships that are less intense and more restricted in significance. Some of the equivalence sets to which a person belongs are wholly included subsets of broader CS groups; others merely intersect. The Authority Ranking (AR) mod bases sociality on asymmetrical difference, typically transitive and hence linearly ordered. This mod itself does not define how people are ordered with respect to specific social practices or values. Hence people using this mod must complete it with socially transmitted preos which define how to rank people: by age, gender, caste, seniority, promotion system, achievement on a task or test, contest or combat, passage through a ritual, possession of symbolic paraphernalia, bestowal of fiefdom, position determined by divination or revelation, charismatic performance, religious devotion, election, delegation or appointment by higher authority. In most cultures, each person participates in a variety of AR relations with all kinds of tangible and intangible beings. But there is a tendency to assimilate the preos used to rank any given set of persons, creating a single ordering across contexts. However, in some instances people use distinct preos, ranking the same set of persons differently in different contexts or with respect to different issues: for example, by age or seniority for some purposes, but by office for other purposes. The Equality Matching (EM) mod constructs relations according to additive interval differences, with even balance as the reference point. Examples include turn-taking; lottery or coin-flip; voting; eye-for-an-eye, tooth-fora-tooth vengeance; rotating credit associations; baby-sitting coops; balanced in-kind reciprocity such as exchange of favors or dinner invitations; matching contributions; distributions divided into equal shares; symmetrical playing fields and even numbers of players in sports; and equal starting points and resources in games and contests. The EM mod defines unit differences as additive and subtractive, but does not specify what constitutes a unit, or, for example, what delay should occur between events. Cultural coordination devices differ as a function of preos that indicate, for example, what constitutes a turn, a favor, a dinner, an episode of baby-sitting, the number of castles on a chess board or the number of minutes available to each player for making moves. Coordination requires socially transmitted preos determining the length of a half in a sporting event, the procedure for assigning teams to ends of the field, and who calls the coin flip. How long should a person wait before offering a return invitation? Is the offer of a bride sufficient restitution for the life taken in a homicide? If a kin group gives a bride to another group, who is entitled to the bride given in return? What sort of bride should they expect, and when? May they state their

6

FISKE

claim explicitly, or would that be rude? Who is eligible to vote, and what constitutes a valid vote? The EM mod is necessary for structuring each of these activities, but preos are necessary too. Neither is sufficient—they complement each other, fitting together to provide a definite model for generating, interpreting, coordinating, judging, and motivating coordinated action. The Market Pricing (MP) mod organizes interaction with reference to ratios or rates. Depending on the preos that complete the mod, it yields such cultural coordination devices as prices, wages, rents, interest, dividends, tithes and taxes, efficiency calculations, and cost-benefit analyses. Money is a common medium for MP, but relationships based on MP often do not involve money, and money is often a medium for relationships based on other mods. When people ask themselves, "Is what I'm getting out of this relationship proportional to what I'm putting into it?" they are using MP to judge relational equity. MP appears in utilitarian moral reasoning when people use a common metric for evaluating aggregate welfare in terms of the greatest good for the greatest number. It appears when the military makes decisions based on ratios of fighter planes lost or on infantry killratios. To implement the MP mod to coordinate an interaction, people require preos that indicate the factors that determine the specific ratios, as well as the measures for the numerator and denominator: How do you measure rice and labor? What qualities of rice determine its value? How many kilograms of rice does one owe in return for an hour of unskilled labor at harvest time? Does the baby sitter get paid per hour or per child per hour? People also need preos to guide them in determining who is eligible to participate in what kinds of MP entities: Should I pay for the soup my mother fed me? May I sell you one of my kidneys, or my gestational services?

IMPLEMENTATIONS AND COMBINATIONS OF THE FOUR BONDS

Preos indicate how to apply a mod, and they also specify which mod operates among which people with respect to which entities in which domains. Hence the relational mods represent alternative ways of doing anything social. However, the innumerable particular implementations varying across culture, history, and domain reflect just four generative structures. For example, there are four basic ways of making a collective decision: People can decide how to respond to a terrorist attack by consensus, by following a chain of command, by voting, or by cost-benefit analysis. Likewise, there are four basic ways of organizing work: People can clean the beaches by encouraging everyone to take responsibility as a community (perhaps volun-

1. RELATIONAL MODELS THEORY 2.0

7

tarily showing up for the annual cleanup), by assigning the task to subordinates, by dividing the coast into equal strips for each person to clean, or by soliciting bids from waste management companies. Similarly, there are four basic ways of using land to mediate social relations: A lot can be a commons held by all for all to use. It can be a fief granted by the king that constitutes the holder as lord over all vassals resident in it. It can be a homestead of fixed size granted to each citizen who farms it and establishes residence. Or it can be an investment to be rented or developed for profit. Again, there are four basic ways for a culture to organize the temporal aspect of a social interaction: People may take turns, alternating temporal intervals. Alternatively, time may mark precedence: In West African villages, you must offer a libation to the ancestors before drinking, and pass the calabash of beer along according to status, from senior to junior. Time can also be a denominator for remuneration: interest at 6% per annum, rent at $1,200 a month. In CS relationships, people may think of their relationship as timeless—they feel they have "always" been together and should be together forever. Obviously there are great cultural differences in the prevalence and ideological valuation of the four relational models, so what is natural in one community, historical epoch, or domain is abhorrent in another. There have been many societies where the sale of land was no more conceivable than the sale of the sun or the ocean, and there were societies in which it was routine to sell sex or slaves. The indeterminacy of the mods gives them a generative potential to combine with innumerable cultural preos to coordinate social activity in any domain of human life: labor, exchange, contribution, distribution, consumption, collective decision making, political ideology, moral judgment, religion, the interpretation of misfortune, aggression, sports and games, marriage, or the social meaning of time, place, and objects. This means that, for effective, mutually comprehensible social coordination of innumerable types of activities, humans need only the four social mods, along with mechanisms for discerning the cultural preos necessary for implementing the mods as specific cultural coordination devices. Because there are innumerable preos, people generate varied cultures. But because all humans have the same relational mods, children, immigrants, and social scientists are capable of apprehending and participating in any culture. The core of the theory is this idea that people use the same set of four implicit cognitive schemas to organize all of the diverse domains of sociality most of the time. For example, as the term Authority Ranking indicates, relational model theory posits that prestige/value/status hierarchies have the same cognitive basis as legitimate power/control/command. These are obviously distinct implementations of AR, since having greater prestige does not confer the authority to command another. In the one

8

FISKE

case, there is a linear ordering of social value, in the other a linear ordering of social control. But analytically, they have relational structures that are formally homologous, and people rely on the same implicit cognitive tools to construct them. There are four basic forces in the physical universe, structuring interaction among a small set of elementary particles. The complexity and variability of the physical universe result from diverse manifestations and combinations of these few fundamental forces and particles. If relational models theory is correct, the social universe may also be based on just four basic relational bonds. The diversity and complexity of human societies, institutions, and relationships results from diverse manifestations and combinations of the four mods. Indeed, most dyadic relationships, and all stable groups and institutions, are formed out of combinations of models, concatenated or nested. For example, survivors of an attack may seek vengeance in an EM mode, an eye for an eye and a tooth for a tooth. At the same time, the CS model structures both group responsibility for wreaking vengeance and the conflating of persons targeted as collectively responsible: The CS model operates when people feel that "we" have to get revenge against "them"—it does not particularly matter who among "us" carries out the counterattack, or which of "them" suffer. Everyone in the in-group is equivalent as victims and agents, so the group is the unit of suffering, justice, and retaliatory action; meanwhile "they" are a relational unit in which all persons are equivalent targets. MP may govern another phase of the event if leaders consider the costs and benefits of alternative means of wreaking vengeance. At the same time, AR may operate in leadership of the strategic decisions and among the military units who carry out the attacks—or in mediating a peaceful settlement. Likewise, consider a diplomatic or military dinner party. At formal functions where rank is marked, guests should arrive in inverse order of rank (subordinates arriving earlier and awaiting superiors) and leave in order of rank (no one leaving before anyone superior to them). Seating at the table also marks rank. However, the food was purchased and the staff paid in an MP framework. But guests share what they eat and drink—the food and drink is theirs collectively to freely consume. Indeed, eating and drinking together marks and constitutes a bond of solidarity among those who partake. At the same time, participants exchange toasts with one another, matching one-to-one, and may feel obligated to reciprocate the hospitality by later inviting their host to a corresponding dinner. So all four models may operate at the same time or in succession, organizing distinct aspects of the social event. Different implementations, combinations, and sequences of the four elementary models thus generate a complex world full of unique interactions.

1. RELATIONAL MODELS THEORY 2.0

9

"RELATIONSHIPS" As these examples illustrate, the term "relationship" is confusing in this theoretical context because, for example, it is not exact to state that people are "in" or "have" a CS relationship. This terminology reifies the models as if they were places that enclose persons, or kinds of objects that a set of people can possess, which are misleading metaphors. People may use a single model to coordinate the simplest, most transitory interactions. But generally the various aspects of interactions among a dyad or group are governed by more than one relational model; different aspects of an interaction may be simultaneously construed in terms of different models. Even if a set of people consistently use a single mod, they would have to implement it with somewhat different preos on different occasions. In other words, relational models are cognitive-affective-motivational models of and for aspects of interaction, not necessarily comprehensive or stable properties of sets of people. Nevertheless, if we keep this in mind, it may suffice and it simplifies expression to write or speak of "people in a CS relationship" rather than the more precise "people using the CS mod with certain preos to coordinate a particular aspect of their interaction in one domain at a certain point in time." Relational models theory posits that people are inherently sociable, by nature. Homo sapiens evolved these relational proclivities because of the benefits all parties tend to reap from participation in coordinated social action. Consequently, the relational models are not merely cognitive capacities: they are intrinsically motivating. People need and seek each of these kinds of relationship, although it seems that the four needs are not equally strong when they emerge in childhood. Furthermore, during development, cultural practices doubtless amplify or attenuate, redirect, and transform relational needs. Moreover, sociomoral emotions guide people to seek, create, sustain, sanction, transform, or terminate relationships (Fiske, 2002). Sociomoral emotions are proxies for the long-term expected adaptive value of relationships. That is, these emotions are motivational representations of the systemic state of relationships, directing people to act in ways likely to promote valuable relationships.2 For example, people who lack important or adequate relationships of a particular kind experience specific "appetites" for that type of relationship. Certain emotions reflect the positive value of good relationships, such as love in some CS relationships, awe/rev2 Most cultures lack clearly defined folk-concepts or lexical representations of many of these emotional experiences, just as they lack surface representations of many other basic psychological and social processes, such as the four mods, or syntactical features of language. We should not expect ethno-psychology to be psychological science, and if the lexicon were a good representation of basic psychological and social phenomena, social science would be just the compilation of dictionaries.

10

FISKE

erence in AR (looking up), comradeship in EM, and sense of satisfaction from good MP transactions ("got a good price on that!"). A person who has transgressed an important relationship experiences emotions motivating them to redress that relationship, and so on. Emotions are essential for sustaining social relationships because optimistic and misleading cognitive biases and heuristics, hyperbolic discounting, and the covert nature of many social sanctions would make it virtually impossible for people to reliably learn, or consistently make reflective decisions to act in accord with, their long-term relational interests (Fiske, 2002). Using the term, relationships in the colloquial sense, people often suppose that MP is not a relationship. But MP is, of course, a model—a structure or implicit schema, a syntax—for generating, coordinating, and evaluating interaction. However, it does seem to be true that MP relationships typically have less intrinsic motivational value for participants. And, compared to the other three models, when MP relationships do arouse strong emotions, it appears that, on average, much of the emotional intensity is linked to the extrinsic consequences of people's actions, rather than being intrinsic to the MP relationships for its own sake. People more often engage in MP relationships largely as means to extrinsic ends; this is much less common in EM and AR, and relatively rare in CS relationships. Economists, of course, generally assume that people interact solely in order to pursue individual self-interest. Nonetheless, it is notable that economists who focus on the advantages and problems of close coordination have discovered three basic types of institutions: markets and hierarchies (Simon, 1981; Williamson, 1975, 1985, 1996), and "clans" (Ouchi, 1980). The institutional approach argues that the form of organization used is the one which minimizes "transaction costs." Williamson assumes that markets are the default system, and people only adopt other structures when explicit contracts fail to control costly uncertainties. In contrast, Simon analyzes how markets and hierarchies correspond to different ways of using information to make decisions. (Note, however, that by "hierarchy" neither author means a simple linear ordering; they mean a dendritic pyramid kind of vertical integration that involves the inclusion of multiple lower level units under and within each higher level unit.) I was unaware of the transaction-cost approach to institutions when I developed relational models theory. Furthermore, the transaction-cost economists developed their ideas without being aware of Udy's (1959, 1970) discovery, based on analysis of a huge sample of ethnographies, of four basic forms of organizing labor in traditional societies. Indeed, my perception of the principal contribution of relational models theory is that it shows researchers in many fields that the systems of coordination they are studying are not limited to any particular domain—and therefore cannot be products of causes specific to any one domain.

1. RELATIONAL MODELS THEORY 2.0

II

The proximity of a set of persons, or their perception of one another, do not, ipso facto, constitute a "social relationship." Humans who are present in the same space or otherwise aware of another's physical existence are not necessarily coordinating their interaction with reference to relational models: Desperate to reach her train, a woman running along a sidewalk dodges parking meters, fire hydrants, signs, manikins left on the pavement, dogs, and humans—taking into account the expected trajectories of these obstacles but without necessarily coordinating her movement with reference to an implicitly shared model such as "keep to the right." This is a "null" relationship: The people are aware of each other's physical existence, but lack any joint, motivated schema to coordinate and evaluate each other's actions. Consider another hypothetical example: Two hungry men, strangers from different cultures, discover a troop of gorillas eating fruit from a tall tree on the other side of a river—and they all see each other. The men may establish a relationship, cooperating to cross the river, drive away the gorillas, and then harvest and consume the fruit. Or they may treat each other as mere obstacles to satisfying their own hunger, like the river and the gorillas. In the latter case, they have a null relationship. Relational models theory claims that there are only four universal, motivated relational models that people are disposed to use for virtually all aspects of every domain of social interaction. People use combinations of relational models to coordinate nearly all social activities. In that sense the relational models are fundamental and basic. However, they are not exhaustive, because people do invent specialized, culture-specific schemas for coordinating particular types of interaction. There are models for coordination in specific situations that do not seem to derive from any relational model: "Drive on the right," "discard high to signal your partner to lead the suit again," "wave a white flag to signal desire for a truce," or "only chiefs may build rectangular houses and place ostrich eggs at the roof corners."3 Special coordination devices such as these probably were invented once and diffused to certain other cultures but remain limited to one domain. In contrast, coordination devices that are derived from the four relational models are pervasive. Mods have certain characteristic features that distinguish them from special-purpose cultural coordination devices (adapted and shortened from Fiske, 2000): • They are extremely widespread across cultures and pervasive within them; that is, a distinctive, consistent structure occurs very widely, gen3 Such models can be linked to relational models, for example, if a person claims damages in an MP framework from a person who veered to the left across the median. Or a father playing bridge might treat his son's failure to discard high as an impertinence, violating their AR relationship.

12

FISKE

•

• •

•

•

• •

crating indefinitely many specific cultural coordination devices (e.g., many ways of taking turns or reciprocating in-kind). They are closely associated with highly structured learning mechanisms that result in reliable development of specific competences, despite highly variable and imperfect cultural input. Virtually all adults are capable of using them, despite vast differences in individual experience and general intelligence. They are each constituted, communicated, and culturally completed in a semiotically distinct medium corresponding to an homologous mode of cognitive representation (see Fiske, chap. 3, this volume). Children characteristically search actively for and take the initiative to engage in these forms of interaction. Children experiment and innovate in ways not readily induced from the input stimuli and make constructive mistakes that could not result from simple associative induction. People have focused, oriented motives for forming these types of relationship, along with emotions oriented to sustaining and sanctioning them. These types of relationship have significant adaptive advantages and result from plausible processes of natural selection. These relationships have neurological substrates characterized by a significant degree of functional modularity.

The four relational mods are the only ones with these features that structure social coordination in all domains of human activity. However, there seem to be other mods that generate ritual (Dulaney & Fiske, 1994; Fiske & Haslam, 1997), language, and taboos concerning sex and food. These mods cross-cut or intersect with the four relational mods and with each other: Language is an important medium for the conduct of relationships, while rituals are important in constituting and transforming relationships. There are also mods that shape how people construct, evaluate, motivate, and cognize combinations of relationships. These meta-relational mods constitute a distinct level of psychology and social relations, not yet adequately characterized by a general theory. Since the 1970s the principal focus of social psychology has been on how people perceive, categorize, and make inferences about other individuals. This research on person perception, stereotyping, and identity investigates how people process a particularly important kind of stimulus: humans (including the self). Relational models theory describes another level of social cognition: how people coordinate interaction. A third level of social cognition and evaluation involves "syntactical" models of how relationships combine with each other: Some sets of relationships imply or constitute other relationships, whereas other

1. RELATIONAL MODELS THEORY 2.0

13

sets of good relationships are mutually incompatible with each other. This is meta-relational cognition.

INNATENESS, AND HOW RELATIONAL MODS MIGHT HAVE EVOLVED

Are these four systems of coordination "in" the mind, or are they emergent arrangements that result from logical or systemic constraints on the possibilities for organizing any social interaction? Would any intelligent agent with sophisticated communicative abilities quickly discover and adopt these four forms of sociality? The universality and pervasiveness of these four forms of relationship does not imply that they are innate. Suppose these models were not prepared cognitive proclivities built in to our neural architecture as it emerges during development. We might still observe these structures in every domain of sociality in every culture because of their functional advantages or because they represent some other kind of stable equilibrium such as an evolutionary stable strategy (Maynard Smith & Price, 1973). These four relational structures do seem to have special functional advantages, so people who invent or imitate these forms of organization may benefit from doing so. Under some conditions, insight into these advantages, or cultural group selection, may result in the cultural reproduction and diffusion of forms of organization beneficial to the group (Boyd & Richerson, 2002). Consider, however, how natural selection probably shaped human cognition and motivation to fit advantageous forms of sociality (Fiske, 2000). A system of social coordination that is typically advantageous to all participants is an adaptive niche, an opportunity for enhancing fitness. Natural selection will favor individuals who have heritable tendencies to learn to set up such systems, engage others in them, and participate advantageously. (This kind of evolution is often called "Baldwinian selection" for the assimilation of learned behaviors; for an overview, see Fiske, 2000.) Every time humans invent or stumble into such an arrangement in any domain where it confers benefits, individuals who readily learn and effectively engage in the coordination system will have greater fitness than those who are less capable or less disposed to do so. That is, it will be adaptive to structure and orient general imitative skills, selectively tuning them so people become adept at reproducing adaptive coordination systems observed among others—and inventing new adaptive systems. Moreover, once humans launch a jointly beneficial coordination system in one specific domain, it will be tremendously advantageous for individuals to learn to generalize their acquired capacities to other domains where they will also function advantageously. We would thus expect natural selection

14

FISKE

to yield cognitive-emotional proclivities structured enough to facilitate their easy, reliable, sophisticated use, yet flexible enough to permit them to be used in an indefinite variety of domains. People should evolve capacities and motives that are effectively organized but open-ended so they can be readily used to solve novel or locally variable adaptive problems. Indeed, this kind of structured cognitive flexibility permits humans to take advantage of the maximum range of social opportunities and thereby exploit many novel ecological niches. In short, because the same four relational systems work well to coordinate most aspects of social action in innumerable domains of joint activity, it is extremely advantageous to be able to apply them adeptly and flexibly. Thus asking whether the relational mods are emergent or innate is somewhat like asking whether color is "in" the environment or "in" the perceptual system. Relational mods are mental adaptations to fundamental affordances of social coordination—adaptations to relational niches. Baldwinian theory and the approach of Tooby and Cosmides (1992) explain how and why specialized domain-specific cognitive capacities would evolve. But neurons are metabolically expensive and the birth canal limits the size of the human brain as well, so there are strong costs and constraints to evolving innumerable specialized modules rigidly dedicated to narrow tasks. Rigid, domain-specific modules would impede construction of adaptively flexible social systems and adaptive functioning in them, and would block rapid adaptation to new conditions. Whenever possible, natural selection should thus produce a small number of structured yet flexible, open-ended, generative proclivities that facilitate adaptive social coordination across diverse domains in diverse environments. However, we know little about the phylogeny of the relational models. (For useful reviews, see Haslam, 1997 on primates and Dugatkin, 1997 on "cooperation" among animals.) Yet many social birds and mammals and some reptiles form dominance hierarchies: Is the human proclivity for AR a direct evolutionary homolog of these social dispositions and cognitions? That is, did phylogenetically deep psychological proclivities to create dominance hierarchies evolve directly into human AR? Or is AR a novel, distinct, analogical adaptation that evolved independently out of other cognitive capacities? Nonhuman dominance hierarchies are constituted by threat-deference asymmetries related to priority of access to desirable resources such as scarce food, nest sites, or (for males) sexually receptive females. Primate dominance consists largely of the subordinate animal conceding defeat when attacked by a dominant individual or coalition, and either ceding a resource when threatened by a superior, or offering an appeasement-deference signal. Human hierarchies sometimes involve this sort of aggressive confrontation resolved by threat, submission, and appeasement.

1. RELATIONAL MODELS THEORY 2.0

15

Also, human AR often does regulate access to resources and mating opportunities. So AR could be a generalization of dominance hierarchies. Eventually, fMRI studies of the functional neuroanatomy of relational models in humans may provide clues about the phylogeny of AR (lacoboni et al., in press). As of this writing, however, virtually nothing is known about the neuroanatomy of primate dominance hierarchies, so we have nothing to compare the human data to yet. It seems likely that human CS evolved as a flexible generalization of mammalian parental care, just as parental care seems to have evolved several times into pair bonding in certain birds, a few mammals, and a couple of primates (including Titi monkeys; Manson, 1997). Only a very few mammals actively share food with other adults or share the feeding and care of offspring, notably some species of canids and, more extensively, mole-rats (Bennett & Faulkes, 2000). Many of the callitrichidae (marmosets and tamarins, New World primates) are polyandrous, with multiple males actively cooperating to carry, feed and care for the large twins that are normally born. No other living primate is known to do this, so human CS probably is an independently evolved analog of food-sharing and cooperative proclivities in other taxa. Alternatively, CS may be an evolutionary outgrowth of collective group defense and attack against nongroup members, evident in a number of primates and other mammals (on chimpanzees, especially, see Manson & Wrangham, 1991). Many people have suggested that Market Pricing, in particular, is not innate—or that if it is an innate, socially specialized capacity, it has little or no inherent motivational force unless this is culturally fostered during development or early adulthood. My hypothesis is that MP is currently in the process of being assimilated into cognitive and motivational proclivities: It is becoming a mod. If so, it provides an opportunity to study an intermediate stage in the progression from invention and transmission based on general cognitive and imitative skills, along the way to assimilation into a structured generative mental proclivity. However, because most learning in most organisms requires comparison of temporal and other rates (Gallistel & Gibbon, 2000), MP may possibly represent a unique case in which a general cognitive mechanism is developing an offshoot specialized for social coordination. That is, the capacity to use rates and other proportions to coordinate social interaction may have evolved from general-purpose, nonsocial capacities to calculate reinforcement ratios. If so, it is not clear why only humans evolved this social capacity, unless MP depends on symbolic representation—as indeed it does (Fiske, chap. 3, this volume). The question that remains is what permitted humans to evolve these flexibly generative relational capacities out of the much more rigid, limited, domain-specific precursors of their ancestors. The most important factors

16

FISKE

may have included the synergistic effects of combining language with social relationships, along with the emergence of increasingly abstract cognition, including analogical reasoning.

A PERSONAL HISTORY OF THE DEVELOPMENT OF THE THEORY

Relational models theory has roots in many other theories and research traditions; one of my principal aims in developing the theory was to integrate the classical theories of human sociality. From the beginning, I was committed to the idea that an adequate theory of human sociality must integrate evolutionary, psychological, developmental, social, and cultural processes. This commitment was fostered by my interdisciplinary training: My undergraduate major at Harvard College was Social Relations. When I graduated from Harvard I joined the Peace Corps, working in public health for 4 years in Malawi and the Congo. This experience awakened me to the importance of culture in shaping human life, and nearly steered me into a career in public policy and international development. When I opted to pursue my intellectual interests, I knew I wanted to explore human social psychology from a perspective that incorporated culture. Exploring my options, I entered the Social Science Divisional MA program at the University of Chicago, a flexible program where I specialized in something called "Cross-Cultural Problems." That winter I took a course from Robert LeVine that provided a framework for what I wanted to do: psychological anthropology. I entered the University of Chicago PhD program in the Committee on Human Development. In Human Development I studied broadly across psychology, anthropology, and evolutionary studies. When Bob LeVine left for Harvard I started working with Richard Shweder; so I am an intellectual grandchild of their Harvard mentors, John and Beatrice Whiting (although unfortunately I took no classes with them as a Harvard undergraduate). My notion of science as a mode of understanding and explanation was (and still is) that it is a search for patterns, an attempt to construct simpler descriptions of regularities in a complex universe. My first efforts along these lines were a series of literature review papers I wrote in several courses, bringing together empirical research in developmental and social psychology and animal behavior. These papers analyzed aspects of what I called "The Group Bond" in humans and other mammals. I looked for diverse evidence for a preadolescent sensitive period when group identity is most salient and labile, out of which adult group identity is likely to crystallize. This work later influenced my conception of Communal Sharing relationships.

1. RELATIONAL MODELS THEORY 2.0

17

The interests that I developed in graduate school became defined by the intersection of two goals: first, understanding how the psyche that all humans share at birth develops into culture-specific commitments and practices; second, understanding how natural selection, cognition-motivation, development, socialization, and other processes result in sociable, generally moral adults. I was skeptical of the socialization theories I encountered. The descriptions of moral thinking in Jean Piaget's The Moral Judgment of the Child captured my interest, although his rationalist-individualist theory did not. Years earlier, I had struggled with Max Weber in an undergraduate political history class, so I was somewhat resistant when Donald Levine (a sociologist on my PhD committee) pointed me toward Weber's massive unfinished Economy and Society, he told me no one could be a social scientist without reading it. Don knew it in the original, but I bought the excellent English translation and spent months working through it, coming to Don occasionally to get help and discuss it. At some point I perceived a congruence between three of Piaget's stages of moral development and Weber's typology of forms of legitimation of authority. I was very intrigued by this apparent correspondence between forms of ideology and stages of moral justification. To simplify somewhat, Piaget's immanent justice, heteronomy, and mutual respect and cooperation seemed to correspond respectively to Weber's traditional, charismatic, and rational-legal legitimation (although eventually I came to see that traditional legitimation conflated what I later called CS and AR). A little later, someone (I think a fellow student) recommended to me a book by Paul Ricoeur, The Symbolism of Evil, which dissects the history of Christian theodicy. Ricoeur describes early conceptions of evil as defilement (dread of the impure), supplanted later by sin (rebellion against God), and finally a rationalized ethicization. I was astonished to discover what I perceived as a third congruent typology in this new domain. (Ricoeur was at Chicago, but I failed to avail myself of the opportunity to meet or work with him. The nature of the congruence among these three theoretical taxonomies is discussed more fully in Fiske, 1991, pp. 25-30.) Bob LeVine told me that I couldn't be an anthropologist without doing fieldwork, and cheerfully rejected my pleas that I didn't need to do fieldwork because I had just returned from 4 years in Africa as a Peace Corps Volunteer. So I decided my dissertation fieldwork should explore the ethnographic validity and utility of the tripartite theoretical framework that resulted from synthesizing Piaget, Weber, and Ricoeur. I wrote a dissertation proposal and totally rewrote it a score of times. I wanted to study foragers, partly as a change from the farmers I had worked with in the Peace Corps, and partly because my interests in evolution suggested that understanding foragers was far more important than understanding farmers or

18

FISKE

pastoralists. But my reading and inquiries indicated that there were no African foragers with intact cultures living independently of farmers and pastoralists (I didn't know of the Hadza). So I decided to study the Samburu, cattle people of northern Kenya. The Social Science Research Council gave me a dissertation grant to do this, but before I left I was unexpectedly offered a chance to be a Peace Corps country director. Excited by the prospect of international development policy formulation and programming, I accepted the offer, chose Upper Volta as the most challenging and rewarding of the openings, abandoned the Samburu, postponed my fieldwork, and went to Ouagadougou. My plan was to study the largest culture in Upper Volta, the Moose ("Mossi") after my term as country director. So in Ouagadougou I studied the Moore language and, in 1979, went with my wife and infant daughter to live and conduct my fieldwork in a village for 2 years. Once I picked a field site, I lived and worked without an interpreter or assistant, trying to immerse myself in everyday village life; it was the most confusing, overwhelming, exciting, and rewarding intellectual experience of my life. Working among the Moose, I found not just the three relational structures I was looking for, but important social practices that just did not match any of those three. When I returned to write my dissertation, I formulated the concept of Equality Matching to encompass these observations. To my surprise, I was hired by the University of Pennsylvania Department of Psychology, so I quickly finished my dissertation and then began to think about a book. Despite my fieldwork, I lacked confidence in my quadriform theory. I could readily see how to incorporate Tonics' Gemeinschaft and Gesellschaft and Durkheim's mechanical and organic solidarity, but I still wondered whether I was just projecting into all these theories, and into Moose life, what I wanted to see. So, to test my inferences, I started reading about other domains of social life, deductively testing my emerging theory on the literature: If indeed people had four fundamental ways of organizing social relations, the four structures should be evident in any well-studied domain. Soon after, I came to Karl Polanyi's analysis of the basic forms of exchange: householding (clearly CS), redistribution (by chiefs, AR), reciprocity (EM), and markets (MP). Then I found Udy's two comprehensive surveys of how people recruit labor and organize production: familial (CS), custodial (AR), reciprocal (EM), contractual (MP), and voluntary (null). I began to be less skeptical about my synthetic theory: I had the eerie feeling that these congruencies could only be the result of four fundamental forms of human social cognition. Influenced by Geertz's extension of Ryles' theory of culture as "models of and for" action, and by Roy D'Andrade's emerging theory of cognitive models and schemas with directive force, I named these entities "relational models." I was so excited about my assimilative synthesis that I wrote much of the book before it even occurred to me that this theory had to explain the differ-

1. RELATIONAL MODELS THEORY 2.0

19

ences in the manifestations of the four elementary models across different domains: How could the same models be used to organize such disparate practices? This led to my formulation of the concept of implementation rules or, as the idea developed, preos. I am still thinking about this problem, focusing on three issues: First, how do children (and adult outsiders) discover how to implement their mods in ways that permit them to coordinate in locally appropriate and meaningful ways? Second, what is the cognitive form in which each relational model is represented in the human mind? Third, what are the relative advantages of specialized domain-specific modules versus flexible generative capacities that are effective in diverse, varying, and novel domains? My most recent conceptualization of the first two issues is in the third chapter in this book; the third issue is explored in Fiske, 2000. I wrote The Elementary Structures of Social Life (Fiske, 1991) for my own intellectual pleasure, and of course for a general social science audience. But I also had in mind my colleagues in the University of Pennsylvania Department of Psychology—a group of smart people that I expected to be cogently skeptical of such broad claims. The book was also implicitly an argument with the relativistic cultural constructionists who were dominant in the Department of Anthropology when I was a graduate student in Human Development. I had many challenging University of Chicago-type arguments with Anthropology faculty and graduate students there, most of whom regarded psychological and biological explanations as anathema. I hoped (futilely, no doubt) to convince them that it was possible for universal evolved mechanisms to generate cultural diversity and uniqueness—and to show why. A theory is intellectual poetry, and a good theory has a beautiful esthetic purity. I didn't want to pollute my theory by bringing it into contact with dirty data. Yet I knew psychologists would never take this synthetic-inductive theory seriously without deductive hypothesis-testing. I didn't really want to test the theory—there are so many good reasons why data might fail to support a valid theory! But I knew I had to. So, intrigued by the experience of my children occasionally calling me "Mom," I launched a series of studies of naturally occurring social errors, working with Kathryn Mason (the "Mom" whom the kids sometimes called "Dad"), and hiring an unassumingly brilliant student, Nick Haslam. Nick was very skeptical, but he took the job and went to work on the study. We were all somewhat surprised when the data came in clearly supporting the theory—in study after study. Nick then thought up the idea (entirely on his own, while I was spending a quarter at UCSD) of testing the book's even more implausible claim that the four models were discrete cognitive categories, rather than being cognized in a continuous dimensional space. This was such a fundamental tenet of the theory that if I had known what Nick was up to 2,800 miles away, I wouldn't have let him get away with seriously jeopardizing the

20

FISKE

already-supported theory. But he did it, ingeniously adopting Meehl's taxonomic methods to this new domain, with clear results that surprised us all again in their support of the theory. When I gave job talk and then other colloquia on the theory, I did so with trepidation. Surely someone would raise a hand at the end and suggest an obvious fifth relational model; perhaps the audience would come up with dozens! They never did. The core of the talk was the claim that the same four structures were evident in many different social domains, so I always wondered if someone would show that the structures in different domains were, in fact, basically different. So far, no one has.

ISSUES AND CONFUSIONS My conceptions of the relational models have developed somewhat since the early 1990s. My approach in the 1991 book was very assimilative, as I attempted to find connections across diverse theories and research. However, as the book neared completion I realized that I had been somewhat indiscriminate, or at least theoretically vague, in my enthusiasm to find homologies among the many theories I was attempting to integrate. I realized that I should try to discern the defining features of each model, distinguishing these universal mods (as I have come to call them more recently) from the infinitely diverse cultural coordination devices that grow out of them. Hence I suggested that the core of MP is the use of ratios to coordinate interaction. Other authors define similar constructs in terms of selfishness, individualism, maximization, competitiveness, contractualism, and other attributes. I argued that these features, though perhaps correlated with using ratios to coordinate interaction, were nonessential implementations that vary across domains and cultures. Similar questions arise with regard to CS. The most parsimonious definition is that it is an equivalence relationship—a relationship organized with respect to something that people perceive they have in common that makes them collectively distinctive. This means that positive feeling and mutual altruism, although typically strongly associated with CS, are not essential features of the underlying mod. Similarly, the orientation of action and affect toward "need" is not a defining feature. If people treat a resource as a jointly held commons, they do not necessarily utilize it according to need: One person's use of the resource is, in principle, equivalent to any other person's use of it. Reviewing the original statement of relational models theory, Turner (1992) and Whitehead (1993) cogently characterized its principal deficiency, the limited account of conflict, aggression, coercion, exploitation, and dysfunction. The theory was developed partly as an antidote to the asocial,

1. RELATIONAL MODELS THEORY 2.0

21

selfish economic/behaviorist or Machiavellian views of human sociality. So, despite some discussion of how the models organize aggression and conflict (see the master table in Fiske, 1991 and 1992), it did indeed present an overly rosy account of harmonious sociality. Whitehead's lengthy review essay provided some of what was missing, pointing out especially that people may impose the use of a relational model on others. To this I would add that people are disposed to get angry and punish those who violate the models that they themselves are using, but the targets of such sanctions often do not acknowledge that that particular model applies, or that their acts were transgressions, so they perceive the intended sanctions as illegitimate aggression. This may launch rounds of mutual recrimination or violence. Furthermore, from a theory of the basic forms of social relationships it is possible to make inferences about potential sources of social dysfunction; Nick Haslam and I have, in fact, developed several implications of the theory for psychopathology. Social problems are likely to arise when participants in an interaction use different models, implement the models with different preos, or apply the models too strictly and inflexibly. If a person uses a model excessively, underuses a model, or consistently uses a model in ways that are discrepant with the cultural preos that others around them are using, that person will experience persistent relational difficulties that are likely to be diagnosed by a clinician as personality disorders (Haslam, chap. 12, this volume).

THE NEXT STEPS Relational models are structures for constructing and construing social action. Beyond relational models, there is another major level of social cognition, evaluation, and motivation: People use meta-relational models to combine models or keep them separate. That is, people have strongly motivated models to understand, coordinate, and evaluate combinations of relationships. These meta-relational models are powerful constraints on when, where, how, and with whom people use the relational models. At the level of relationships, developmental studies and cultural comparison will eventually identify the universal features that constitute the mods and distinguish these from the preos that inform variable cultural implementations. Theoretical questions remain, however, about whether there is a hard, clear line between mods and preos. Clearly some cultural implementations of a given mod are much, much more common than others. This suggests that different mods lend themselves most readily to different implementations and are prone to be used in particular domains, perhaps because mods are selected for their advantages in those typical implementa-

22

FISKE

tions. That is, although each relational model can be used in indefinitely many ways to structure any domain, there may be psychological propensities, functional advantages, or systemic factors that make some implementations more prevalent than others. There are some obvious patterns: CS among close co-resident kin (the "household") is virtually universal, while more distant kinship involves lesser degrees of CS. Furthermore, the consumption of cooked food, drink, and other comestibles tends to be communally organized everyday, and commensalism may be more inclusive at rituals, although other models come into play to lesser extents. These patterns may reflect the links between CS and inclusive fitness of individuals sharing a high proportion of genes. AR is pervasively organized according to age, gender, and descent categories such as "race" and caste. In traditional societies, at least, men tend to be more actively involved than women in AR activities outside the household. EM typically involves people of roughly corresponding age and social role. Also, it seems that violence is most likely to be structured in terms of EM as vengeful retaliation. MP becomes more prevalent the greater the number of parties and kinds of entities involved, the more transient the transactions, and the less the certainty of sustained interactions across any particular dyad. In its current form, relational models theory is a descriptive theory, characterizing how people think about sociality. But the theory would be much more powerful (and interesting) if we could deduce why and when people implement particular models. Ideally, these deductions should be based on the characteristics of the relational models themselves and the differences between them. In the immediate encounter, clearly people are obliged to use the models pretty much as their predecessors and associates use them, by virtue of the social transmission that results in local cultures. Otherwise, people could not coordinate. In novel situations, people probably use analogies from other important domains. But what affects the selection of models in the long term? Presumably there are functional constraints and potentials that affect the prevalence of the models in different social domains—each model has weaknesses and strengths for organizing a given domain, if we can define appropriate performance criteria. These "capacities" of the models also should be evident in the consequences (e.g., the kinds of problems) that are typical of implementations in particular domains. Certain combinations of the models may be more functional and stable than others because those sets of models perform complementary functions. If we understood where and how different relational models "work" and the nature of the problems they produce, the theory could be applied prescriptively. However, if we aim

1. RELATIONAL MODELS THEORY 2.0

23

to explain when and why people use particular models, determining the consequences of using each model is not sufficient. For an explanatory theory, we need to analyze when and how functional effects of the models become causes of processes that select among them. The relative functional advantages of each relational model are probably related to the differential costs of collecting and storing the information necessary to use each model, the differential costs of making the calculations necessary to use each model, and the differing degrees of specificity of the preos that have to be negotiated in order to achieve smooth coordination with different models. For example, to utilize a resource in the CS mode, people merely need a preo that defines who has access to that resource. The University of California gives faculty, staff, and students unlimited access to athletic facilities; others are excluded. They just have to make sure each person seeking access is in the user group. However, if people allocate a resource according to AR, they must have a preo that specifies the ranking of each person, they must keep track of the ranks of users who seek access to the resource, and they must agree about whether the precedence accorded does correspond to personal rank. For example, if faculty had priority of access to the gym according to seniority, and faculty had precedence before students, then whenever a person sought access to, say, a tennis court, someone would have to determine if a court was free, and if not, whether any current user had a lower rank than the person seeking the court. This requires collection of more information, more processing of that information, and possibly more negotiation. Access structured according to EM would require preos determining what counted as a unit of use, and how long people could delay taking a turn before losing their priority. There would have to be a mechanism to record and compare everyone's utilization and make that information available so people would know when they were entitled to use the facility. And people would have to agree that these preos were properly implemented and the mechanisms correctly applied. The use of MP would require definition of an auction mechanism or price scheme, along with bookkeeping and billing. Then, with MP, people would have to agree about what had been used and what payments were owed and had been made. With each successive model, things get more complicated: more information to collect and store, more calculations to make, more details to be negotiated. As the terms of interaction become more quantitatively specific, the negotiations are likely to become more arduous. Balanced against these information and transaction costs are the advantages—and disadvantages—of exactly structured coordination. At one end, these benefits may be greatest toward the MP end of the series, where the coordination is most precisely structured. But there are countervailing advantages of the wider latitude or indeterminacy toward the CS end. CS per-

24

FISKE

mits more latitude but less precise coordination than AR, which is less structured than EM. MP provides the most precisely specified coordination, but offers less flexibility. So, if we are building a school in an MP framework, we can specify precisely how much each person has to do, and what they get in return. That has considerable advantages in terms of predictability and precision of coordination, but it imposes strict constraints. If someone is ill and can't work, they don't get paid through the MP mechanism in this domain—though their illness may make their need very desperate. If someone has free time and is motivated to volunteer to help build the school, then paying the volunteer increases the costs of school building compared to a CS framework in which everyone would simply pitch in until they all got the job done together. In short, from CS to AR to EM to MP, use of the models imposes an increasing burden of information collection, storage, and processing, along with greater specificity of the terms that must be agreed upon in order to coordinate. Likewise, the determinacy of the interaction increases step by step, while flexibility decreases. Over the long run, the functional trade-offs may select for models that, in given domains, offer the greatest difference between coordinative opportunities and information-processing burdens. In general, more tightly structured coordination is likely to provide the greatest advantages to participants. However, in some cases, looser, less restricted forms of coordination may permit beneficial flexibility. Crosscultural and developmental research will permit us to test this approach and better understand the apparent patterns in the implementation of the models.

REFERENCES Anderson, B. R. O'G. (1983). Imagined communities: Reflections on the origin and spread of nationalism. London: Verso. Bennett, N. C., & Faulkes, C. G. (2000). African mole-rats: Ecology and eusociality. Cambridge, England: Cambridge University Press. Boyd, R., & Richerson, P. J. (2002). Group beneficial norms spread rapidly in a structured population. Journal of Theoretical Biology, 215, 287-296. Dugatkin, L. A. (1997). Cooperation among animals: An evolutionary perspective. New York: Oxford University Press. Dulaney, S., & Fiske, A. P. (1994). Cultural rituals and obsessive-compulsive disorder: Is there a common psychological mechanism? Ethos, 22, 243-283. Fiske, A. P. (1990). Relativity within Moose ("Mossi") culture: Four incommensurable models for social relationships. Ethos, 18, 180-204. Fiske, A. P. (1991). Structures of social life: The four elementary forms of human relations. New York: Free Press. Fiske, A. P. (1992). The four elementary forms of sociality: Framework for a unified theory of social relations. Psychological Review, 99, 689-723.

1. RELATIONAL MODELS THEORY 2.0

25

Fiske, A. P. (2000). Complementarity theory: Why human social capacities evolved to require cultural complements. Personality and Social Psychology Review, 4, 76-94. Fiske, A. P. (2002). Moral emotions provide the self-control needed to sustain social relationships. Self and Identity, 1, 169-175. Fiske, A. P., & Haslam, N. (1997). Is obsessive-compulsive disorder a pathology of the human disposition to perform socially meaningful rituals? Evidence of similar content. Journal of Nervous and Mental Disease, 185, 211-222. Gallistel, C. R., & Gibbon, J. (2000). Time, rate, and conditioning. Psychological Review, 107, 289-344. Haslam, N. (1997). Four grammars for primate social relations. In J. Simpson & D. Kenrick (Eds.), Evolutionary social psychology (pp. 297-316). Mahwah, NJ: Lawrence Erlbaum Associates. lacoboni, M., Lieberman, M. D., Knowlton, B. J., Molnar-Szakacs, I., Moritz, M., Throop, M. J., & Fiske, A. P. (in press). Watching social interactions produces dorsomedial prefrontal and medial parietal BOLD fMRI signal increases compared to a resting baseline. Neurolmage. Manson, J. H. (1997). Primate consortships: A critical review. Current Anthropology, 38, 353-374. Manson, J. H., & Wrangham, R. W. (1991). Intergroup aggression in chimpanzees and humans. Current Anthropology, 32, 369-390. Maynard Smith, J., & Price, G. R. (1973). The logic of animal conflict. Nature, 246, 15-18. Ouchi, W. G. (1980). Markets, bureaucracies, and clans. Administrative Science Quarterly, 25, 120-142. Simon, H. A. (1981). The sciences of the artificial (2nd ed.). Cambridge, MA: MIT Press. Tooby, J., & Cosmides, L. (1992). Psychological foundations of culture. In J. Barkow, L. Cosmides, & J. Tooby (Eds.), The adapted mind: Evolutionary psychology and the generation of culture. New York: Oxford University Press. Turner, J. H. (1992). Structures of social life [book review]. Contemporary Sociology, 21,126-128. Udy, S. H. (1959). Organization of work: A comparative analysis of production among nonindustrial peoples. New Haven, CT: Human Relations Area Files Press. Udy, S. H. (1970). Work in traditional and modern society. Englewood Cliffs, NJ: Prentice-Hall. Whitehead, H. (1993). Morals, models, and motives in a different light: A rumination on Alan Fiske's Structures of Social Life. Ethos, 21, 319-356. Williamson, 0. E. (1975). Markets and hierarchies: Analysis and antitrust implications. New York: Free Press. Williamson, O. E. (1985). The economic institutions of capitalism. New York: Free Press. Williamson, 0. E. (1996). The mechanisms of governance. New York: Oxford University Press.

This page intentionally left blank

C H A P T E R

2 Research on the Relational Models: An Overview Nick Haslam

Philosophers of science disagree about what constitutes a good theory, but most researchers seem to share a rough understanding. A good theory should do several things well. The explanatory concepts that it employsstructures, processes, mechanisms, and the like—should map onto real, demonstrable states of affairs. These concepts ought to make powerful sense of a wide range of phenomena within the theory's field of relevance. Ideally they should do so economically, elegantly, and better than alternative theories. The theory should offer original understandings of existing findings, and yield bold and testable predictions about new ones. It should apply universally, not only in particular times or places. It should generate active research programs and inspire new and ramifying lines of theoretical inquiry. As philosophy of science goes this list of desiderata is a mishmash, but it surely captures much of what theories aspire to do. The theory of relational models (RMs), reviewed by Alan Fiske in the previous chapter, has given rise to a large and diverse body of research in the last decade, and the time is ripe to take stock. How well has the theory stood up to the many tests that theories face? What are we to make of its past, present, and future, not as a set of abstract tenets but as a living body of ideas and a growing tradition of work? In this chapter I survey the research literature that has drawn on the theory of RMs, and come to some conclusions about how well it has performed, where it is likely to go, and where, perhaps, it needs to go further.

27

28

HASLAM

Deciding where to mark the perimeter of this research literature is no easy task, for three reasons. First, RMs theory has been widely cited in a variety of fields in the decade since its inception. Second, "research" can be defined more or less narrowly, depending on methodological taste. Third, it is frequently difficult to decide when a citation of the theory represents a substantive inspiration for a piece of research, or an original piece of conceptual research in its own right, rather than being merely an acknowledging nod to the theory's relevance in a particular empirical context. I have attempted to resolve these ambiguities by being comprehensive in my coverage of the existing work, by defining "research" in a broad and inclusive way to accommodate everything from experiments to speculative exercises in conceptual linkage, and by referring only to studies whose use of RMs theory was more than incidental. The review is divided into four main sections. In the first section, entitled "Structures," I discuss work on the structural basis of the theory, its adequacy as a specification of the cognitive structures that govern social relations. The second section, entitled "Influences," surveys research on the role of the RMs in accounting for psychological and social processes and practices. It is divided into subsections that address various individuallevel and collective phenomena, as well as individual differences. A third section on "Connections," reviews a diverse collection of theoretical linkages that have been proposed between the models and other concepts within the social and behavioral sciences. A final "Conclusions" section offers some thoughts in closing on the status of RMs theory in light of the existing research and on directions for future work.

SECTION I: STRUCTURES

A significant amount of empirical research has addressed the structural component of RMs theory. Fundamentally the theory makes three basic structural claims: that there are four basic models governing social relationships, that these models are best understood as discrete and incommensurable categories rather than as dimensional continua, and that their core features match Fiske's characterization. None of these claims is selfevident. Theorists have proposed classifications containing anywhere from two to very many categories, and have often failed to specify whether these categories represented groupings of actual relationships or of the psychological structures that generate them. Often theorists have eschewed categories altogether, preferring to formulate differences among relationships in terms of dimensions of warmth, formality, inequality, and so on. Needless to say, no theory has formulated its basic relational categories in quite the way that Fiske did, so it is entirely conceivable that he was mistaken in his

2. RELATIONAL MODELS: AN OVERVIEW

29

specification of at least some of their features. The three fundamental structural claims of the theory of RMs therefore require thorough empirical validation. In the first attempt to assess whether the RMs as described by Fiske (1991a) conformed in a coherent manner to four underlying psychological structures, Haslam (1995a) conducted an exploratory factor analysis of undergraduates' ratings of their personal relationships. Each participant rated a sample of their relationships on items measuring features of the four RMs and four resource classes—Love, Information, Status, and Services—drawn from Foa and Foa's (1974) resource exchange theory, an alternative relational taxonomy. Two bipolar factors emerged from this analysis, one opposing Communal Sharing (CS) items to Market Pricing (MP) items and the other opposing Equality Matching (EM) to Authority Ranking (AR). There was some tendency for EM items to fall closer to the CS region of the factor space than to the MP region, and for AR items to show the opposite pattern, suggesting that the two apparent dimensions may not be truly independent. Items within each model generally cohered adequately. The resource class items, in contrast, showed relatively poor coverage of the two-dimensional space and failed to adhere to either the underlying dimensions or a spatial ordering proposed by Foa and Foa (1974). Although the two-factor solution obtained by Haslam (1995a) provided some evidence for the coherence of the RMs, in other respects it was contrary to theoretical expectation. Two factors emerged with particular RMs at each pole whereas four factors, each corresponding to one model, would have been conjectured. Nevertheless, there are several grounds for treating this analysis as less than final. First, the factor structure of the RMs may have been distorted by the inclusion of the resource class items. Second, the analysis may have suffered from the methodological limitations of exploratory multivariate research, so that the factor solution may reflect unsystematic influences peculiar to the study and its dimensionality may be questionable. Finally, the analysis constrained its factors to be independent, contrary both to theoretical expectation and to the apparent structure of the RMs items as revealed in the analysis, implying that it may have misrepresented the latent structure of the models. Two studies have since attempted to overcome these methodological limitations by performing confirmatory factor analyses on improved item sets restricted to the RMs. In the first, Haslam and Fiske (1999) used items from the now-standard questionnaire measure of the RMs—the Modes of Relationships Questionnaire (MORQ)—that assessed the operation of each model in multiple domains (e.g., social exchange, morality, social identity). A group of non-student adults rated a representative sample of their personal relationships on this instrument. The fit of three alternative factor models to this data set was compared: two- and four-factor solutions, with

30

HASLAM

the latter either allowing for or not allowing for correlations among the factors. The two-factor solution reflected a reductive understanding of the RMs as poles on bipolar dimensions of authority or inequality (AR vs. EM) and warmth (CS vs. MP), comparable to the solution obtained by Haslam (1995a), and the four-factor models represented the predicted distinctness of each model as a single unipolar factor. Permitting factor intercorrelations (obliqueness) represented the predicted systematic covariation of some models as a function of contingent social and cultural norms, contrasted with the normal assumption that relationship factors are statistically independent. Haslam and Fiske's (1999) findings demonstrated the clearly superior fit of the theoretically favored factor structure containing four oblique dimensions. All MORQ items loaded satisfactorily on their correct RM, in support of the internal coherence of Fiske's characterization of the models. The CS factor was strongly associated with the EM factor, and MP was less strongly associated with both EM and AR. In a second confirmatory factor analysis, Vodosek (2000a) developed a new questionnaire concerning beliefs about how work teams should function, with items representing distinct goals linked to the four RMs. Graduate business students rated their goals for consulting teams in which they worked, rather than a sample of personal relationships. Nevertheless, Vodosek's findings concur impressively with the earlier analysis. A structure composed of four oblique factors demonstrated significantly better fit than one in which the factors were constrained to be orthogonal. All items loaded substantially on their predicted factor. The CS factor was highly correlated with the EM factor, and AR was significantly associated with MP. In addition, the fit statistics in Vodosek's analysis were decidedly better than in Haslam and Fiske's (1999), and indicated that the oblique factor solution was a very close approximation to the obtained data. Discreteness

The factor-analytic work reviewed earlier furnishes strong evidence for the structural adequacy of RMs theory. Collectively, these studies indicate that four correlated, unipolar factors capture the structure of social relationships, and that these factors are well described by coherent sets of relational features proposed by the theory. These factors fail to correspond to more established taxonomies, which have generally invoked a smaller number of independent and bipolar factors. People's intuitions about the structure of their personal relationships also seem to map onto the four models, and they have little trouble assigning particular relationships to a RM. Although it has substantially confirmed the structural basis of RMs theory, this factor-analytic research fails to address a bold, central postulate of

2. RELATIONAL MODELS: AN OVERVIEW

31

the theory, namely its insistence that the RMs are categories that apply to particular relationships in an all-or-nothing fashion. In contrast to the standard assumption that relationships differ by degrees along a set of continuous axes (e.g., Wish, Deutsch, & Kaplan, 1976)—an assumption that is built into factor analysis—the theory maintains that each model provides a discrete and incommensurable way of organizing relationships. Two studies have tested between these dimensional and categorical alternatives, using taxometric procedures (e.g., Meehl, 1992). In the first taxometric study, Haslam (1994a) had undergraduate participants rate a representative sample of their personal relationships on items that assessed aspects of the four models, as well as items assessing four of Foa and Foa's (1974) six resource classes. Item sets for each proposed relational form were analyzed using Meehl's MAXCOV procedure, which yielded unambiguous evidence for the discreteness of all four of the RMs but only two of the resource classes (Love and Status), which other evidence suggested were essentially redundant with one another and with one of the RMs (CS). Besides the main finding that the hypothesized features of the RMs cohered in ways that imply discontinuities between different relationships, four further findings of the study are noteworthy. First, very few of the sampled relationships did not belong to one of the RMs. Second, a majority of relationships were governed by more than one model, consistent with Fiske's (1991a) claim that different aspects or domains of particular relationships may be organized according to different models. This is consistent with the argument that the RMs do not represent a taxonomy of relationships but a set of discrete principles for organizing social cognition and interaction. Third, taxometric evidence indicated that Americans may tend to disavow AR and MP relationships, because the estimated proportions of these relationships in the sample was substantially smaller than the proportions participants give when they freely assign their relationships to the models. Finally, the RMs seemed to be more powerfully informative about relational features than colloquial role terms such as boss or friend, supporting their primacy as determinants of interpersonal activity. In a second study (Haslam, 1999), another sample of relationships was gathered and rated on a different set of items assessing the RMs. These items were again employed in taxometric analyses of the models, using two different procedures (MAMBAC and admixture analysis) in addition to MAXCOV. Once again, the results of these analyses were highly consistent with RMs theory's conjecture that the structure of social relationships is discontinuous. Although all of the studies reviewed to this point have used self-report data, none of them have directly asked whether the RMs map cleanly onto

32

HASLAM

people's intuitive understandings of their relationships. Structures matching the models might be detectable in people's ratings of their relationships without these statistically inferred structures corresponding to their folk taxonomies, or to the ways in which they classify relationships in an explicit and self-conscious fashion. A lack of correspondence would be no great surprise, and would not harm the theory of RMs, which are conceptualized as implicit social grammars rather than explicit folk understandings. A scientific theory of grammar would be on shaky ground, after all, if it had to mirror laypeople's intuitive formulations of grammatical rules, although their self-reported intuitive judgments of grammaticality are indispensable. Nevertheless, if correspondences were found between latent and manifest structures in this way it would extend the theory's claim to capture fundamental distinctions among social relationships. With this in mind, Haslam and Fiske (1992) conducted a study in which intuitive understandings of the structure of relationships were assessed in two distinct ways. Participants freely sorted a sample of their personal relationships into groups and rated the similarity of pairs of these relationships, from which relationship groupings were derived cluster-analytically. In addition, participants classified each relationship according to five relationship taxonomies: the RMs, Foa and Foa's (1974) resource classes classification, Mills and Clark's (1984) communal versus exchange dichotomy, Parsons and Shils' (1951) role expectations, and MacCrimmon and Messick's (1976) social orientations. The extent to which participants' intuitive groupings of their relationships mapped onto each of these taxonomies was then examined. The RMs classification performed well in the company of its four contenders. It was associated with the intuitive groupings just as strongly as the role expectations and resource classes, compared to which it was employed with somewhat greater reliability and ease, and it was more strongly associated than the social orientations and communal versus exchange classifications. The RMs also showed a relatively complete coverage of a multidimensional relational attribute space derived from the five theories, uniquely capturing the crucial authority and equality-related aspects of relationships and failing only to recognize degrees of formality or rule-governedness. Within RMs theory such variations might be understood as contingent implementation parameters rather than basic relational forms. The results also indicated that intuitive relationship groupings were best modeled by about four categories, consistent with RMs theory. Simpler classifications, such as the communal versus exchange dichotomy, appear to be insufficiently differentiated for that task. In sum, the theory of RMs emerged from the study as a superior reflection of participants' reflective understandings of their relationships. Its relative success appears to rest on a

2. RELATIONAL MODELS: AN OVERVIEW

33

taxonomic structure of sufficient complexity, economy, and comprehensiveness for that purpose. A second study assessed intuitive understandings of relationships in a quite different manner. Rather than having participants classify personal relationships in intuitive as well as theory-guided ways as in Haslam and Fiske (1992), Haslam (1994b) had them rate the prototypicality of a sample of hypothetical relationships, or choose which of several forms of behavior of one interactant—described in terms of interpersonal circle octants (e.g., Kiesler, 1983)—would be most appropriately paired with a specific form of behavior of a second interactant. The fit of three alternative structures—dimensional, lawful, and categorical—to these rating and choice data was then compared. The dimensional structures modeled judgments in terms of their degrees of warmth and dominance, whereas the lawful structures corresponded to complementarity rules proposed by interpersonal circle theorists (e.g., Wiggins, 1980). The categorical structures accounted for relationship prototypicality judgments in terms of discrete relational categories corresponding to the CS and AR models. Regression analyses indicated that these relationship categories modeled the rating and choice data more strongly and economically than the alternatives. Although confined to only two of the RMs, this study supports their capacity to make sense of people's intuitive understandings of relationships. The research reviewed in this section makes a solid case for the soundness of the structural component of RMs theory, and gives affirmative answers to its three fundamental claims. The confirmatory factor-analytic and taxometric studies support the existence of four models irreducible to a smaller set of more basic structures, and Haslam and Fiske (1992) indicate that four relational structures may be close to sufficient to account for intuitive understandings of relationship types. It is impossible to rule out the existence of additional models—although their theoretical grounding in the four established scales of measurement (Fiske, 1991a) would argue against this—but the finding that the models appear to govern all but a small minority of relationships, most of which are probably asocial or Null relationships (Fiske, 1991a), suggests that the four models are not lacking an indispensable fifth. The taxometric studies and Haslam (1994b) support the discreteness of the RMs, and directly challenge any understanding of them as dimensional constructs, making this the first relational taxonomy to have established such support. The factor analytic and taxometric studies confirm the coherence of the RMs, indicating that several distinct sets of items intended to tap diverse aspects of the models all covary systematically. In addition to noting this chorus of support for the structural basis of the theory, it is also worth acknowledging that when the RMs taxonomy has been compared to its primary alternatives it has tended to fare well (Haslam, 1994a, 1994b, 1995a; Haslam & Fiske, 1992).

34

HASLAM

SECTION 2: INFLUENCES Establishing the empirical credentials of the structural component of RMs theory is an important step toward validating it as a whole. Any research that employs the taxonomy is only as sound as the taxonomy itself, and if the latter draws spurious distinctions or conflates distinct structures by failing to draw correct ones, such research will have limited success. All the same, validating the theory's structural component is just a preliminary step, as the theory's true test is whether it can account for social phenomena. A number of studies attest to the power of the RMs in accounting for individual-level social cognition and behavior and collective phenomena. Individual Cognition and Behavior

The first paper to demonstrate the usefulness of the RMs theory in accounting for individual cognitive and behavioral phenomena reported a series of seven studies (Fiske, Haslam, & Fiske, 1991). These studies examined social errors, inadvertent slips in which one person is mistakenly substituted for another. Errors of three types were examined: "misnamings," in which an addressee was referred to by the wrong name, "person memory errors," in which the wrong person was recalled as having been present in a situation or event, and "misactions," in which an interpersonal behavior was addressed to an unintended person (e.g., calling the wrong person on the telephone). In each study, a sample of participants collected personal examples of one or more kinds of these errors and recorded them using a standard diary methodology. In addition to describing each error, participants described the people who figured in it on multiple dimensions, including the RM they employed with each of these people. In all seven studies participants demonstrated a significant tendency to substitute people with whom they related in the same manner (i.e., according to a particular RM). This finding held equally for the RMs taxonomy as a whole and for each model individually. The effects of RM on confusability were more robust than the effects of such variables as acquaintance age, social role, race, and name similarity, of which they were statistically independent. They were also stronger than those of two alternative relational taxonomies, Foa and Foa's (1974) resource classes and Mills and Clark's (1984) communal versus exchange distinction. Only gender rivaled RM as a determinant of slips. Importantly, the RMs appeared to underlie and largely account for the effect of social role. Participants tended to substitute acquaintances designated by the same culturally available role term in large part because they shared with these acquaintances the same kind of relationship, indicating that the "deeper" relational structures were more important and basic for social cognition than the colloquial role categories.

2. RELATIONAL MODELS: AN OVERVIEW

35

In a second paper on social errors, Fiske (1993) replicated the earlier studies with participants from four diverse cultures (Bengali, Chinese, Korean, and Vai [from Sierra Leone and Liberia]). All participants were living in the United States but were relatively unacculturated, and the studies were conducted in their native tongues. In every sample, social errors systematically preserved the RM governing the participants' relationships with the substituted persons. This effect of the RMs was, if anything, stronger among less acculturated participants. Social role, as well as individual attributes such as gender and age, also predicted errors, but these effects were frequently accounted for by participants' tendencies to associate with, and hence substitute, people of a particular demographic profile. Consequently, at least some of these effects do not reflect cognitively appraised social equivalence. This study therefore demonstrates an impressive level of cross-cultural validity for the RMs and their influence on social cognition. In a follow-up to the error studies, Fiske and Haslam (1997b) examined intentional substitutions, in which an initially intended interactant was replaced with another when the former was unavailable or the latter changed plans. Using the same diary methods for collecting naturally occurring substitutions, Fiske and Haslam again found that participants substituted interactants with whom they had the same kind of relationship. Although substitutability was also predicted by interactants' personal attributes—gender, ethnicity, and age, but generally not personality—their effects again appeared to reflect the participants' demographically differentiated patterns of affiliation rather than cognitive equivalence. The findings of the study therefore offer further support for the role of the RMs in determining social equivalence. Seeking further evidence for the wide-ranging role in social cognition of the RMs, Fiske (1995) had two samples of participants freely list their acquaintances by name and subsequently classify them according to relational properties (models, resources, communal vs. exchange, role term, situations in which they interacted) and personal attributes (gender, age, race). The extent to which the order of recall of the acquaintances was clustered into "runs" by each of these characteristics was then compared. Recall order was most powerfully determined by the relational characteristics, all of which yielded stronger clustering values than all of the personal attributes, with social situation preeminent. The RMs yielded the strongest effect of the relational taxonomies, once again demonstrating their influence on social cognition in a new methodology and cognitive task. The social error, substitution, and recall studies indicate that social cognition—whether it involves the naming or representation of persons, recall of social episodes or individuals, or formulation of interpersonal intentions—is powerfully influenced and guided by the nature of people's relationships with their interactants, rather than simply by the personal attri-

36

HASLAM

butes that inhere in them. Moreover, these consequential relationships are well captured by the RMs. Much social cognition truly is "thinking about relationships" (Fiske & Haslam, 1996), and approaches to social cognition that implicitly conceptualize the social field as a collection of attribute bundles, rather than as a web of interrelations, risk a serious misrepresentation. Such a misrepresentation, which treats social cognition as similar in kind to object cognition, can be understood as a reflection of the taken-for-granted methodological individualism of much American social psychology. It also reflects a tradition of laboratory research in which the social element is represented either by strangers or hypothetical persons rather than real acquaintances. The studies conducted by Fiske and colleagues on the influence of RMs on social cognition have run counter to this tradition, implicating participants' actual relationships. However, a few studies have taken different approaches to the role of the RMs in social cognition. Baldwin (2000) used a priming methodology to demonstrate that the RMs can be activated by particular words in ways that allow inferences about the implicit structure, correlates, and chronic activation of these models via reaction times to the identification of other target words. Baldwin obtained intriguing evidence for gender differences in the salience or activation of the RMs and in their associations with parental figures. Female participants tended to identify a CS target faster and MP and AR targets slower when primed by "Dad" than by "Mom," suggesting that these models are selectively associated with the two parents. The male participants demonstrated precisely the opposite pattern of reaction times, and also tended to identify a MP target quicker than women, implying greater chronic activation of this model. Baldwin's work suggests that the RMs are amenable to rigorous laboratory methods, which may enable studies of automatic or implicit cognitive processing about relationships. lacoboni et al. (2002) present a particularly exciting demonstration of how laboratory methods can illuminate RM theory, employing cutting-edge neuroscientific procedures. Using functional magnetic resonance imaging (fMRI), they investigated the brain regions involved in CS and AR relationships, activating these with a number of carefully controlled movie clips depicting relationships of each type. CS and AR clips activated largely overlapping bilateral brain regions relative to a resting control, and these regions differed substantially from those activated under standard cognitive tasks. AR clips activated certain temporal and frontal regions more than CS clips, and both clip types generated somewhat distinct activation patterns when two interactants were presented rather than a lone individual. The findings suggest that social cognition employs somewhat different brain areas than nonsocial cognition, and that somewhat different neural mechanisms underpin the interpretation of social interactions governed by

2. RELATIONAL MODELS: AN OVERVIEW

37

different RMs. They are consistent with the view that we understand social relationships by interpreting the actions of others as embedded in a meaningful context, and by simulating them as if they occurred to us. This research on the functional anatomy of the RMs should help to deepen our understanding of the models and place them in an expanded theoretical and empirical context. It should, for instance, help to understand what mental functions the RMs draw on, and how they emerge and develop. It might help to establish homologous brain structures in other species and thereby enable speculations about the evolution of social competence. It might also afford new perspectives on individual differences in the use of the RMs, such as those that might underpin certain psychopathologies (see chaps. 11 and 12, this volume). At the other end of the cognitive science spectrum from this neuroimaging work, Jackendoff (1992, 1994, 1995, 1999) has presented formal analyses of the conceptual structures underpinning social relations that support the existence of a distinct faculty of social cognition of which the RMs might be a part. Arguing on multiple grounds for the plausibility of a central module dedicated to social cognition, Jackendoff (1994) proposed that cultural learning must involve the fleshing out of skeletal abstract frames or notions like the RMs. Any simple inductive account of cultural learning founders, he maintains, on the existence of irreducibly social conceptual primitives that are not linked perceptually to the world and hence must be in some way grasped intuitively. Cultural knowledge, that is, must have a "fairly substantial innate endowment" (p. 204). Besides its theoretical consistency with Fiske's account of cultural learning (Whitehead, 1993) as an externalization of universal frameworks and the linkage of them to local implementation parameters, Jackendoff's work supports RMs theory by suggesting that the models are apt to be some of the fundamental frames in any social module. Collective Phenomena The empirical studies of the structure of the RMs and of their influence on social cognition and behavior bear on important facets of the theory, but focus almost exclusively on the mental activity of individuals in relation to their personal social milieus. The theory of RMs was also intended to make sense of social or collective phenomena involving shared activities, understandings, and values; that is to say, culture and social organization. A variety of research projects have trained their attention on phenomena of this sort. Perhaps the richest evidence for the fruitfulness and explanatory reach of the RMs comes from ethnographic studies. Two ethnographic studies have thus far been conducted: Fiske's (1990, 1991a) among the Moose of Burkina Faso and Whitehead's (2000a, 2000b) among the Seltaman of New

38

HASLAM

Guinea. Fiske's ethnography represents the founding study of RMs theory, in which its abstract conceptual synthesis touches dusty ground and establishes that it is workable. The ethnographic material is rich with examples that reveal the operation of the four models not only in family contexts of social exchange, distribution, work, kinship, and so on, but in less straightforwardly interpersonal spheres of Moose culture, such as moral systems, relations with divinities, and orientations to land, time, and artifacts. Especially as presented in Fiske (1991 a, Section IV), this work serves as an unmatched source of examples that convey the distinct flavors of the RMs. It also provides empirical support for the theoretical claim that humans are inherently sociable (Fiske, 1991b), demonstrating how calculative individualism is hardly the universal social logic it is sometimes assumed to be within strands of Western social science. Whitehead (2000a, 2000b) employed RMs theory among the Seltaman of Papua New Guinea. She showed that food sharing and food taboos could be illuminated by the RMs. Like Fiske's ethnographic work, Whitehead's brings into sharp relief the extent to which basic relational assumptions in one culture may diverge in recognizable ways from those in another. The American tendency to disavow and avoid the hierarchical dimension of social life, and to esteem the "higher" position within the AR relationships that they acknowledge, contrasts with the taken-for-grantedness of AR in Seltaman society, and their eagerness to take subordinate roles. Several researchers have used RMs theory to approach questions associated with social justice. Folger, Sheppard, and Buttram (1995) present an overview of distinct justice principles often identified as equity, equality and need, and show how these principles correspond to, and are elucidated by, the MP, EM, and CS models, respectively. They show further how these principles apply at multiple levels—individual, interpersonal relations, and institutions—and extend them to refer to procedural as well as distributive justice. AR is seen to have a place in justice research as a model for ensuring procedural aspects of fairness, although in a study by Sondak (1998) MBA students rated this model least procedurally just as a way to allocate positions in elective classes, preferring MP. Folger et al.'s (1995) work helpfully discusses the ways in which the multiple "lenses" for making sense of justice can lead to ambiguity and lack of consensus within organizations— especially when provoked by rapid organizational change and intercultural contact—and proposes ways of handling this multiplicity. Goodnow (1998) has examined justice perceptions outside of the organizational context, focusing instead on divisions of work within households (see chap. 5, this volume). She finds that RMs theory offers an illuminating account of the distribution of responsibilities among household members. Decisions regarding distributions depend crucially on the models members believe should apply to the relevant relationships, and these models

2. RELATIONAL MODELS: AN OVERVIEW

39

are implemented in varied ways as a function of age and family role. Moreover, the RMs clarify relationship errors of particular kinds, such that different family positions are associated with sensitivity to confusions that these positions invite, mothers being alert to confusions with maids (MP) and adolescent children to receiving "orders" (AR). Goodnow's analysis presents a valuable explication of implementation parameters and reveals the many dimensions of relational tension and negotiation that arise within households. Mikula (1999) has also investigated perceptions of justice in relation to the distribution of household tasks, studying both adolescents within families and college students living in shared flats. In three studies, he inductively developed a classification of distributive arrangements and rules, and examined how perceptions of these varied according to the nature of the task and of the social system (family vs. cohabiting students). There was a general tendency for egalitarian and balanced (EM) arrangements to be perceived as more just, although less regulated communal (CS) arrangements were considered somewhat more acceptable within families than within the student households, and strict apportionment of tasks to the person who created the need for them less acceptable. Consistent with this finding, participants in a scenario study regarded unreciprocated assignment of a task to a particular person less unjust among close friends than among acquaintances. The perceived justice of particular distributive arrangements varied considerably by task as well, according to whether they required special skills, benefitted or were necessitated by identifiable individuals, and were time consuming, suggesting several determinants of the conditions under which particular RMs may be implemented. In addition, there were large discrepancies between the arrangements judged to be most just and those actually implemented, and possible self-interest biases in such judgments, with those who contributed the least household work rating arrangements that do not enforce balance or prevent free riding more just. These findings imply a large amount of conflict and disagreement surrounding distributive arrangements within the home, and suggest that a RMs-based analysis may help to understand the forms that this dissent will take. Three studies have investigated how the RMs taxonomy might illuminate group psychology. Sondak (1998) presents a re-analysis of a study of string quartets, investigating how they handle problems of democracy, leadership, rivalry, and conflict management in terms of the models. He makes a number of informal observations about how the more successful quartets navigated these issues—cultivating a strong CS ethos but also allowing the first violin to adopt an AR position within the group, for example—and how the implementation of every model needs to be studied for an adequate picture of the group dynamics to emerge.

40

HASLAM

In a detailed study of science research groups, Vodosek (2003) demonstrated the negative group outcomes and atmospheres that occur when members lack a shared definition of intra-group relationships. Dissimilarity in the relational models that group members applied to their relationships with one another was associated with lower satisfaction with and commitment to the group, more negative affective response to it, and greater intention to quit. Interestingly, this dissimilarity was greater among more culturally diverse groups, as predicted. Vodosek's work therefore points to an important source of negative outcomes in groups, and also offers a useful framework for understanding how relational models dissimilarity affects group outcomes. Lickel, Hamilton, and Sherman (2001) used the RMs taxonomy to investigate people's intuitive understandings or "theories" of different kinds of groups. Their work starts from a robust finding that people intuitively recognize four qualitatively distinct group types: intimacy groups such as families, task groups such as work teams, social categories such as women, and loose associations such as people who like classical music. These group types are perceived to differ on many dimensions, including size, duration, permeability, internal similarity, subjective importance to members, and so on. Lickel et al. (2001) show that people also perceive that different types of relationship occur among members of the different group types, and draw strong inferences from group types to likely relational norms. Intimacy groups are associated with CS and to a lesser extent EM relations, task groups with AR and to a lesser extent MP and EM relations, and loose associations with MP relations. Social categories were not perceived to have a consistent relational style. Study participants also made reliable and symmetrical inferences from prevailing relational norms to probable group types and other group features, such as size and permeability, indicating the extent to which intragroup relationships figure in lay theories of groups. For instance, groups organized in a CS fashion were found to be judged the most unified, coherent, and entity-like, and to draw stronger judgments of collective responsibility when one member committed a wrongdoing. These studies demonstrate the potential role of RMs theory within group psychology, and within the broader domain of lay theories of the social domain. A significant quantity of research conducted under the banner of RMs theory has focused on conflict, at intrapersonal, interpersonal, and intergroup levels. The theory lends itself to this focus because it proposes discrete and incommensurable frameworks for social relations. If interacting individuals or groups bring different prescriptive frameworks to their engagement, the result is likely to be friction of some sort. Not only is there likely to be mutual incomprehension and a moralized sense of violation, but there will also be no ready way to mediate the disagreement because the

2. RELATIONAL MODELS: AN OVERVIEW

41

models are in principle irreconcilable (Fiske, 1990). The implications of RMs theory for the depth and quality of conflict differ from those derivable from standard dimensional accounts of social relations. By these accounts, divergences between individuals' or groups' prescriptive standards for interaction simply amount to differing parameters along continua of warmth, inequality, and so on, and should therefore yield to a gradual recalibration between the parties. There is no obvious reason why differences in relational norms should be difficult to reconcile, as there are no in-principle discontinuities between them. Fiske and Tetlock (1997, 2000) have spelled out a relational theory of conflict that relies heavily on the issue of incommensurability (see chap. 8, this volume). They observe that people commonly resist making tradeoffs that involve apparently incompatible values—such as placing a dollar value on a human life—when such "taboo tradeoffs" are inescapable for many decision makers and should, in principle, be readily performed by laypeople if they operated according to the utility calculus of Homo economicus. Fiske and Tetlock (1997) argue that tradeoffs are taboo whenever they "require assessing the value of something governed by the socially meaningful relations and operations of one RM in the terms of a disparate RM" (pp. 256-257). When confronted with such an intermodal comparison, people react with confusion, indignation, attribute immorality or irrationality to the person making the tradeoff, and want to punish and ostracize the person. They commonly refuse to make such tradeoffs themselves. Fiske and Tetlock present a series of hypotheses about the conditions under which tradeoffs are taboo. For example, they propose that political ideologies correspond to preferences for particular RMs for social organization and decision making, and that intrusions of MP valuations into the domain of CS relations will be judged particularly offensive. In sum, Fiske and Tetlock's taboo tradeoff account illuminates conflict at multiple levels: within the person contemplating a tradeoff, between that person and others who abhor it, between ideological positions that take different stances toward it, and between cultures if they draw the boundaries between distinct relational domains such that an ordinary tradeoff for one is a boundary-crossing atrocity for the other. Several empirical studies have been conducted on taboo tradeoffs. In preliminary work, Tetlock, Peterson, and Lerner (1996) confirmed that taboo tradeoffs in which MP rules were applied in inappropriate domains elicited punitive and moralized trait attributions, anger and disgust, and social ostracism. Tetlock, Kristel, Elson, Green, and Lerner (2000) demonstrated in two studies that taboo tradeoffs generated greater outrage toward those who perform or even seriously entertain them than routine within-model tradeoffs, and that this effect is moderated by participants' political ideology. Interestingly, they also showed that people appear to find the mere

42

HASLAM

contemplation of taboo tradeoffs morally contaminating, prompting attempts at "moral cleansing" to restore their sense of virtue. Tetlock, McGraw, and Kristel (2000) further showed that a taboo tradeofflike phenomenon affects the valuation of objects. In three studies, participants proposed ridiculously inflated valuations (selling prices) for objects acquired outside the "normative orbit" of the MP model, or commonly refused to sell these objects at any price. Consistent with a prediction based on Fiske and Tetlock (1997), distress and outrage at the prospect of selling the object, and monetary "overvaluation" of the selling prices, tended to be greatest for objects acquired in CS relational contexts and steadily diminished as the context more nearly resembled MP (i.e., CS > AR > EM). These findings represent striking violations of the basic economic principle of source independence—the independence of an object's value from its manner of acquisition—and strongly challenges the reducibility of all values to a single utility metric in everyday cognition. Indeed, Tetlock, Kristel, Elson, Green, and Lerner (2000) suggested that findings such as these call for a new metaphor of people as intuitive "moralist-theologians" who strive to "protect sacred values from secular encroachments" (p. 857) at least as much as they strive to achieve accurate prediction and control or to maximize utility, as the prevailing intuitive scientist or economist metaphors would have it. In a further extension of Tetlock and Fiske's work on taboo tradeoffs, Poulson (2000) investigated everyday interpersonal conflicts. In two studies, participants made judgments about their imagined reactions to conflict scenarios involving contribution, distribution, decision making, or social influence. In each conflict scenario, a protagonist held interpersonal expectations based on one RM which were violated by another person's behavior. These violations either preserved the expected RM but contradicted its desired implementation or, more radically, stemmed from the use of a different model. Contrary to hypotheses based on model incommensurability, between-model conflicts did not produce more negative reactions than within-model conflicts. Interesting incidental findings of the study included tendencies for within-model conflicts involving CS to be judged very negatively, and for those in which EM was enacted to be judged relatively inoffensive. The latter findings support a weak default preference among the participants for egalitarian relationships and the former points to the special emotional fraughtness of communal relationships. Although all of the research on conflict presented earlier has employed experimental methods and artificial conflict scenarios, a field study by Connelley (1997) shows how the RMs can make sense of real and consequential conflicts (see chap. 6, this volume). Connelley conducted focus groups within a Fortune 100 manufacturing firm that had recently attempted to expand the hiring of women and minority group members. This

2. RELATIONAL MODELS: AN OVERVIEW

43

initiative had produced misunderstanding and distrust both from the targeted employees, who complained about hiring, performance evaluation, and promotion decisions and whose retention rates were poor, and from the predominantly White male workforce, which nurtured an angry backlash. Connelley established that three groups of employees—the White male "establishment," and the White female and African-American targeted minorities—thought and spoke about workplace relationships and human resource systems in the terms of different RMs. The White males conducted their work-lives according to a CS model in which ingroup solidarity and consensus-driven decision making tended to define other employees as outsiders and deprive them of consultation and the career benefits of informal workplace networking. The White females preferred a MP definition of the work environment, with competence and efficiency governing the allocation of rewards along meritocratic lines. African-Americans, in turn, favored the strict parity of an EM model, supporting firm hiring and promotion targets. Connelley's study is valuable not only in its qualitative detail and its clear delineation of the relational fault lines of a deep organizational conflict, but also in its implication that the "choice" of a RM may be at least partially strategic and positional, motivated as a response to a prevailing ethos, in this case CS, that does not serve one's relational goals. It figures as the first investigation to use RMs theory to understand conflict in organizations, although Vodosek (2000b) has laid out a useful integration of organizational research on work group conflicts that strongly supports this line of work. Vodosek demonstrates the limits of existing theories of intragroup conflict—based variously on demographic and value differences and on social identity or equity theory—which presuppose a particular form of relational cognition (CS for social identity and demographic accounts, MP for equity theory). RMs theory promises a more differentiated understanding of intragroup conflict, and Vodosek presents a set of predictions about conflict as a function of divergences among group members' operating models. All of the research reviewed to date has addressed relationships between individuals or between groups within organizations. An innovative analysis by Sheppard and Tuchinsky (1996) suggests that relationships between business firms can also be understood in terms of the four relational "grammars." Focusing on relationships between "customer" firms and their suppliers, they argue that American management thinkers, dominated by economic, network, and behavioral negotiation theories, have relied too much on hierarchy and the market (AR and MP) as organizing principles. These principles imply a level of distrust and fear of opportunism in customer-supplier relations that may be counterproductive. Sheppard and Tuchinsky discuss ways in which these relations might be reconfigured along CS and EM lines which, they argue, are likely to be more effective

44

HASLAM

when the firms' products are complex, must be swiftly introduced to the market, and employ parts that are highly integrated and depend on rapidly changing technology. Integrated relationships in which suppliers closely "tailor" (EM) their products or services to the customer's specifications or "align" (CS) with them directly to develop products collaboratively embody such models. Besides demonstrating that the RMs can be applied at this macro level of organizational analysis, Sheppard and Tuchinsky's work illustrates just how intellectually and practically limiting a field of study's core assumptions about the nature of social reality can be. RMs theory offers a set of distinct social grammars that can, they imply, call any univocal approach into question, and force investigators to be more self-conscious about the assumptions of their own preferred set of explanatory concepts. Individual Differences

Relatively little research has pursued the implications of RMs theory for the study of individual differences. The models have been conceptualized primarily as components of the universal psychological endowment of the human mind, rather than as sources of differentiation among people. Nevertheless, ways in which the models might account for differences between people in the domains of psychopathology and personality have been developed in several pieces of theoretical research (Fiske, 199la; Fiske & Haslam, 1997a; Haslam, 1995b, 1997b). Running through this work is the supposition that people may differ in a systematic, trait-like manner in their tendencies to employ the models in making sense of their interpersonal worlds. More specifically, some people might completely lack the capacity to construe social relations in terms of one of the models, others might consistently over- or underimplement a model with respect to the normative expectations of their culture, either from a cognitive bias or a motivational abnormality, and still others might fail to implement models in accordance with socially appropriate rules and parameters. In all of these cases, affected people should show a distinctive pattern of interpersonal aberration in which their expectations and forms of conduct conflict with those prevailing within their society. Such an account of interpersonal aberration has several theoretical merits, discussed at greater length in chapter 12 of this volume. It presents an innovatively relational analysis of interpersonal disturbance, both in its reference to aberrant forms of relational cognition and to the interpersonal context in relation to which disturbed individuals are aberrant. Alternative interpersonal accounts tend to be resolutely individualist in their focus, referring to the individual's deviant forms of behavior within social interactions, rather than to his or her forms of relational construal, and making no reference to the normative expectations of the surrounding social milieu.

2. RELATIONAL MODELS: AN OVERVIEW

45

This mindfulness of social context enables the RMs-based account of interpersonal disturbance to account for cultural variations in the forms of interpersonal disturbance, such as personality disorder. In addition, the account offers an analysis of the cognitive processes and structures that underpin interpersonal traits rather than leaving these traits as unanalyzed entities. In contrast, some interpersonal theories of personality disorder employ such dispositions—dependency, vindictiveness, and so on—as explanatory entities when they might be better understood as purely descriptive. Is paranoia "explained" by high levels of dispositional vindictiveness or simply characterized by it, and might not a more satisfactory explanation refer to tendencies to construe relationships in terms of calculative rationality and individualized competition in interpersonal settings where such construals are culturally inappropriate? The first specific analysis of individual differences to make use of RMs theory was developed by Allen and Gilbert (1997), who proposed an innovative evolutionary account of depression (see chap. 11, this volume). According to their account, depression originated phylogenetically as an adaptive mechanism for ensuring a risk-averse strategy in social interaction when people perceived their social resources to be at critically low levels. This strategy dictates vigilance to social threats, general behavioral risk aversion, and agonistic signaling to others that the person poses no threat, and serves to protect and defend the person against a further, calamitous loss of resources. Allen and Gilbert propose that this depressive strategy should manifest itself in a changing profile of relationships comprehensible in terms of the RMs. They argue that it might entail an increased dependency on resources grounded in CS relationships, perhaps coupled with an increased attentiveness to the risk of exclusion from the communal unit, both intelligible as increases in the activation or implementation of the CS model. They also postulate a strategic withdrawal from exchange relationships, specifically EM and MP, given the greater risks of nonreciprocity and being cheated. Finally, they propose that in AR relationships the depressive strategy would promote an increase in submissiveness, a reduction in uprank challenging, and perhaps a generalized increase in sensitivity to rank and status. In effect, this account of the cognitive and behavioral underpinnings of depression represents it as an intrinsically relational phenomenon. The strategy represents a motivated alteration in ways of thinking and behaving within relationships of different kinds. Only one study to date has assessed the claims of Allen and Gilbert's "social risk hypothesis." Semedar (2000) examined differences between depressed and nondepressed students—as well as between students with and without cyclothymia—in the implementation of the RMs in role-based relationship types. Consistent with hypothesis, depressed students were more likely than their nondepressed peers to understand their relationships with

46

HASLAM

family members and employers in CS terms, and also to construe family and close friend relationships in AR terms. Contrary to expectation, they showed an overimplementation of EM with family and with colleagues. Although these findings are preliminary and not entirely consistent with Allen and Gilbert's theory, they do appear to support an increased dependency on communal relationships and increased sensitivity to rank and perhaps inclusion-exclusion among the depressed. In addition to this work on depression, one study has investigated links between the RMs and personality disorders. As adumbrated in a series of theoretical treatments (e.g., Fiske, 1991a; Haslam, 1997b), some personality disorders—maladaptive and inflexible dispositions predominantly expressed in interpersonal relationships—may represent aberrations in the use or implementation of particular RMs (see chap. 12, this volume). Such a claim can take stronger and weaker forms, ranging from a conceptualization of a disorder as specifically caused by the categorical absence of a particular model to a formulation in which the disorder is simply taken to be associated with an aberrant degree of implementation of a particular model. It is possible to develop predictions about the form of aberration specific to a particular personality disorder. For example, the schizoid personality's aloofness and detachment might be understood as the expression of unusually weak motivation for CS relationships, and the narcissist's grandiosity, need for admiration, and sense of entitlement as expressions of an under-implementation of EM and overimplementation of AR. Haslam, Reichert, and Fiske (2002) conducted a study of aberrant use of the RMs in a sample of people with self-identified interpersonal problems. These participants completed self-report measures of their personality disorder symptomatology, their interpersonal problems, and their implementation of, motivation for, and difficulties with relationships governed by each of the four RMs. Although the sample was nonclinical, it showed elevated levels of personality disturbance. Numerous predicted associations between specific personality disorders and relational aberrations received support, despite the study's modest statistical power. Interestingly, the RMs captured distinctive interpersonal components of disorders that have no interpersonal profile according to the interpersonal circumplex (e.g., Wiggins, 1980), the most prominent interpersonal account of personality disorder, and distinguished between pairs of disorders to which the circumplex ascribes identical profiles. For example, the circumplex implausibly identifies both schizoid and avoidant personality disorders with a pattern of cold submissiveness, but the study found them to be marked by unusually high and unusually low motivation for CS relationships, respectively. The personality disorder study was followed up by Caralis and Haslam (2002), who examined relational correlates of normal personality dimensions. In a sample of psychiatric outpatients, measured on implementation

2. RELATIONAL MODELS: AN OVERVIEW

47

of and motives for the RMs as well as the five-factor model of personality, numerous predicted associations were obtained. People who strongly implemented or desired CS relationships tended to be low in Neuroticism, and high in Agreeableness and Conscientiousness. Tendencies to implement EM were associated with low Neuroticism, and high Extraversion and Openness. Preferential use of AR was linked with Neuroticism and low Agreeableness and Openness. These findings point to a rich variety of links between normal personality and relational tendencies, and suggest that contrary to claims that Extraversion and Agreeableness are the interpersonal trait dimensions, all such dimensions have relational aspects or correlates. Research on individual differences in the use of the RMs is plainly in its infancy, but the evidence to date suggests that the RMs approach offers a promising new way of making sense of interpersonal tendencies associated with chronic personality abnormalities, episodic mood disorders, and normal personality.

SECTION 3: CONNECTIONS

Much of what can properly be called research on the RMs consists in demonstrating links between the models and other concepts and theories in the social and behavioral sciences. These conceptual connections have the potential to open new vistas for empirical research and to offer fruitful reconsiderations of existing theoretical accounts. RMs theory appears to propagate an unusually rich variety of connections. Evolution One area in which RMs theory has been put to work is in relation to the evolution of sociality. In a review of primate social relationships, Haslam (1997a) argued that three of the RMs evident in humans have precedents and parallels in the social behavior and cognition of other species. CS is linked to kinship phenomena, such as kin recognition and the unconditional grooming observable within some primate kin groups, evidence that categorical inclusion-exclusion distinctions are drawn and differentiations within the ingroup largely ignored. AR is linked to dominance hierarchies, and the transitive inferences about status within them that some primate species appear to make. Social reckoning akin to human EM can be seen in the formation of alliances and coalitions and related phenomena understandable abstractly in terms of reciprocal altruism. Only MP seems to lack parallels among nonhuman primates, its reliance on ratio-based figuring appearing to exceed their cognitive capacities. Haslam (1997a) argues that MP therefore represents one aspect of human social-cognitive distinctiveness in the midst of several im-

48

HASLAM

portant continuities with primate social cognition. Consequently it is a mistake to imagine that essentially discontinuous capabilities such as language and "theory of mind," the attribution of mental states, exhaust the human social-cognitive apparatus (see chap. 4, this volume). Whereas Haslam (1997a) examined the phylogenetic precursors of the RMs, Fiske (2000) has recently discussed the evolutionary status of the models in broader terms. He offers the models as examples of "cultural coordination devices" for which humans have evolved proclivities. These proclivities have been selected to support and channel the learning, construction, and enactment of local cultural paradigms and for discerning these paradigms in the process of social learning. Culturally specific forms of social coordination are based, by this account, on a form of coevolution of congruent proclivities and cultural paradigms. In this "complementarity theory," the relation between the universal and culturally specific aspects of the RMs is clarified, and their development is understood as the child's active externalization of proclivities in guided search of their local realizations, rather than as internalization. Emotion The evolutionary implications of RMs theory have been further exploited by TenHouten (1996, 1999a, 1999b, 2000) in a series of papers presenting a "socioevolutionary" approach to the emotions. TenHouten strives to give sociological content to Plutchik's (1980) evolutionary account of the emotions, which represents eight primary emotions as adaptive reactions to the negative and positive polarities of four basic existential problems of life: territoriality, hierarchy, temporality, and identity. TenHouten (1996, 1999b) proposes an isomorphism between these survival-related themes and the RMs, such that the models figure as "sociological generalizations" of the themes that were evolved to enable them to be addressed collectively. Territorial issues of ownership and control of resources are linked to MP, hierarchy linked straightforwardly to AR, temporal issues of reproduction and loss linked to CS, and identity concerns related to acceptance as a equal member of a group linked to EM. Implications are drawn for likely emotional responses to social situations in which people are successfully or problematically engaged in social relations of a particular kind. For example, CS relationships are proposed to have privileged connections to joy and sadness, and AR to be associated with anger and fear. In addition to these wide-ranging speculations on their affective linkages, TenHouten (1999a, 1999b) argues that the RMs are associated with four distinct forms of time consciousness which are especially evident in crosscultural perspective. MP, for instance, is aligned with a form of linear time consciousness suitable for quantitative measurement and the figuring of ef-

2. RELATIONAL MODELS: AN OVERVIEW

49

ficiencies. CS, in contrast, carries with it a "patterned-cyclical" sense of time marked by strong links of the present with the past which are re-enacted in repetitive fashion by ritual and ceremony. Temporal modes such as this are preferentially associated with distinct modes of cognition, ranging from more holistic "Gestalt-synthetic" processing for CS to the more "logicoanalytic" reasoning characteristic of MP. TenHouten proposes that the RMs involve two polarized pairs—CS versus MP and AR versus EM—and that AR is linked to MP in an "agonic" couple, and EM with CS in a "hedonic" couple. Although this reduction of Fiske's tetrad to polarized pairings runs counter to his theory, it nevertheless makes heuristic connections. It is somewhat consistent with Francis and Jackson's (in press) finding that people are more emotionally vulnerable to discrepancies between their self-concepts and how others think they ought to be when they are in CS and AR rather than EM or MP relationships with these others. Psychoanalysis

Although Fiske (1991a, 2000) gives an account of the learning of the RMs, both in concrete terms of their order and age of emergence and in the abstract language of externalization, little theoretical or empirical work has been done on social development from a RMs perspective. The one exception is a paper by Forsyth (1995) that examines the mapping of the RMs onto psychoanalytic concepts. Consistent with the psychoanalytic emphasis on earlier stages in psychosocial development, Forsyth argues that CS and AR are the most fundamental social logics and presents EM and MP as essentially derivative. In view of their respective ages of appearance and their formal complexity, CS and AR are linked to preoedipal and oedipal themes, respectively, with the oedipal conflict itself corresponding to the overlaying of a rivalrous and asymmetrical triadic relationship onto the child's earlier dyadic relation with its mothering figure. Aside from these and other conceptual links, Forsyth makes the useful observation that individual differences in the salience of particular models may reflect developmental experiences. Trust Two pairs of theorists have explored the ways in which the RMs illuminate the study of trust. Earle and Cvetkovitch (1997,1999) used the RMs taxonomy to test a theory of "social trust" in the context of environmental risk management. Against a rationalist account of trust in institutions, according to which trust depends on empirically demonstrated competence and responsibility, they argued for a value similarity account by which trust is assigned according to cultural values that institutions narratively represent or espouse. Con-

50

HASLAM

sistent with general prediction, judgments of trust in a fictitious environmental agency depended on the match between people's environmental values, classified according to the RM taxonomy, and the cultural values expressed in messages issuing from the agency. This finding indicates that the RMs capture elements of environmental values and mediate responses to socialpsychological processes like the assignment of trust. Sheppard and Sherman (1998) go beyond this indirect influence of the models on trust, proposing that the very quality and level of trust varies according to the prevailing RM (see chap. 10, this volume). They suggest that the models are underpinned by bipolar dimensions involving the range, importance, and frequency of interaction (shallow vs. deep) on the one hand, and the degree of contingency between the persons within a relationship (dependent vs. interdependent). CS and AR relationships are relatively deep in these terms, and CS and EM relatively interdependent. Levels of trust and of interpersonal risk vary predictably between the models when they are located within this schema. Cultural Dimensions

One mapping between the RMs and another set of constructs has been proposed by Triandis and Gelfand (1998), who argue that the models are associated with the dimensions of individualism and collectivism. The former is understood as a social pattern involving loosely collected individuals driven by personal goals and independent self-understandings, the latter as a pattern of connectedness in which people give priority to collective norms, duties, and goals. These "cultural syndromes" have been influential in the comparative study of cultures and in organizational research, and afford a way for established differences among cultures to be reconceptualized as different degrees of implementation of the four RMs. Triandis and Gelfand distinguish vertical (ranked, unequal) and horizontal (egalitarian, similar) forms of both individualism and collectivism, and suggest that collectivism corresponds to CS, individualism to MP, the vertical dimension to AR, and the horizontal dimension to EM. In a study testing these predicted associations between the two theoretical schemes, Vodosek (2000c) examined consulting group members' judgments of relationships within the group and of the recommendations generated by the group ("outcomes"). Group members' scores on Triandis and Gelfand's (1998) scales for assessing vertical and horizontal individualism and collectivism generally failed to correlate strongly with the RMs. MP was correlated with vertical but not horizontal individualism, CS was associated with horizontal but not vertical collectivism, and EM was associated with horizontal collectivism but not individualism, and all of these findings obtained in judgments of group outcomes only. All associations between the

2. RELATIONAL MODELS: AN OVERVIEW

5I

two systems were weak as well as inconsistent, suggesting that there is less redundancy between them than Triandis and Gelfand (1998) proposed. However, limitations of the study, such as its correlation of general individualist versus collectivist dispositions with contextually bound RM judgments suggest that these associations are worth pursuing in future research. In a similarly cultural vein, Earley (1997, 1998) applied RMs theory to the study of "face," proposing that different dimensions of face are salient in societies whose social exchange practices are governed by particular models. A dimension involving the person's position within a social hierarchy is most prominent in societies with preponderantly AR and MP exchange practices, and a dimension involving conformity to rules for moral conduct is most prominent in societies in which exchange is governed more by CS and AR. SECTION 4: CONCLUSIONS

This chapter began with a list of general desiderata for scientific theories. How well has RMs theory measured up to these standards? Any assessment is provisional, given that the research literature is steadily growing, and needless to say this assessment is not impartial. Nevertheless, the research conducted to date allows some tentative judgments to be made. One of the basic requirements of any theory is that its explanatory concepts correspond to real, demonstrable entities or processes. As reviewed in Section 1, the structural component of RMs theory has accumulated a large amount of empirical support. A series of studies has confirmed and replicated all of the structural postulates of the theory. The models have been found to correspond to four distinct and discrete categories, and the descriptive features of each appear to cohere as hypothesized. Whether the four models are exhaustive is harder to establish, but there is evidence (Haslam & Fiske, 1992) that there is no major dimension of social relations captured by its alternatives that it does not encompass, and these alternatives tend to be less exhaustive. With only four explanatory entities, its coverage is not purchased at the expense of economy, another theoretical desideratum. In sum, the evidence for the structural adequacy of RMs theory is probably stronger than the evidence for any other relational taxonomy. Our review of the influences of the RMs in Section 2 allows a tentative assessment on some of the other criteria set out at the beginning of the chapter. The range of phenomena that they have been used to explain is quite plainly very extensive, ranging widely over individual and collective levels of analysis, from cognitive psychology to organizational behavior, and from relationships between multinational corporations to the oedipus complex. However compelling individual studies and conceptual analyses may be, this diversity demonstrates that theorists and researchers have found the

52

HASLAM

theory to be effective and versatile. The theory has shown itself capable of accounting for a large variety of phenomena as well as offering an integrative understanding of existing accounts that offer more partial understandings subsumable under one of the models. Its breadth of application also speaks to a desirable capacity to inspire a range of research programs, both empirical and theoretical. Our review of research also indicates that the theory has stood up quite well to comparisons with other theories. Although the field of relational taxonomies is somewhat sparse and comparative studies rare, RMs theory has yet to be clearly outperformed in an empirical test and has predicted a variety of phenomena—including slips, lay understandings of relational structure, and psychometric structures underlying personal relationships—at least as well as its alternatives. When the focus of comparison is changed to culture an equally positive conclusion can be drawn. Not only have the RMs been found to be readily applicable to ethnographic studies in West Africa and Papua New Guinea, not to mention a great deal of secondary ethnographic material worldwide, but several of their effects that were first established in the United States have been replicated in other cultures. The theory can therefore make a relatively strong case for universality. A theory that purports to synthesize a great deal of previous thought leaves itself open to the "old wine in new bottles" critique, and a hostile critic might be inclined to question how far RMs theory goes toward offering original understandings of familiar phenomena or falsifiable predictions about new ones. Ideally, of course, a theory ought to do both. It is true that a large proportion of research using RMs theory employs it in an interpretive rather than hypothesis-testing manner, and throws support back onto the theory by the cogency of its analyses rather than by the survival of Popperian tests. Nevertheless, the theory has shown itself not to be predictively empty, and has yielded a large number of empirical studies with overwhelmingly supportive results. The fair-minded reader would also grant, I think, that many of the more theoretical or conceptual analyses reviewed here also offer valuable reframings of existing fields of study, pointing to new ways of thinking about a range of phenomena or indicating that some existing theoretical accounts of particular phenomena rely implicitly on a single model. Revealing the limiting assumptions of existing theoretical accounts and the possibility that several perhaps equally plausible alternatives may exist is surely one way in which a theory can establish its ability to yield new understandings. Directions for Future Research In view of its youth, dispersion across disciplines, and wide range of application, research on the RMs has inevitably been thinly spread. It is therefore difficult to single out areas that seem especially promising for future

2. RELATIONAL MODELS: AN OVERVIEW

53

research efforts or to be in need of further refinement. To this reviewer, a few areas do stand out all the same. First, there has been almost no systematic work on the factors that influence the "choice" of model to be implemented in particular contexts. This question can be pitched at multiple levels, from why most societies tend to organize certain relationships according to particular models to why individuals of particular social positions or life histories tend to employ particular models on specific occasions. Research suggests that partial answers could be found at the levels of adaptive functionality, social ecology, learning history, political ideology, negotiation between initially discrepant interactants, and self-interested or strategic figuring. Whatever shape such analyses take, however, at present it is clear that we know with much more confidence that the models exist and what they influence than we do about when they are used and why. Second, we know precious little about the developmental aspects of the RMs, aside from Fiske's (1991a) speculations about their ages of origin and the "externalization" view of social learning, and Forsyth's (1995) psychoanalytic reconceptualization. The ontogenetic development of the RMs should be a research priority. As a side note, it is also notable that there has been no longitudinal research on the development (or decay) of relationships of different sorts. In theory, the dynamics of relationship change should show some interesting discontinuous features, when one predominating model gives way to another. Third, although RMs theory has unusually strong cultural credentials as a psychological theory, there has been no systematic research on cultural comparison in terms of the models. The cultural universality of the models and their role in social cognition and action has been supported in several studies, but we know little about how cultures differ in their "preferences" for particular models either in general or in relation to specific kinds of relationships. How such preferences map onto established dimensions of cultural comparison awaits further study. Fourth, the general area of relations among relations—how relationships governed by particular models combine, interlock, or conflict with one another—has yet to attract adequate theoretical or empirical attention. This is an important task in linking dyadic interpersonal relationships to social organization more broadly, and should extend and deepen our understanding of the relational embeddedness of social cognition. It might also draw fruitful connection to work on social networks whose disciplinary home base of sociology is underrepresented in research and theory on the RMs. Fifth, although substantial work has been addressed to the RMs themselves, little attention has been paid to specifying the nature and form of the implementation parameters that enable them to be realized in concrete relationships. Ideally, future research might begin the task of modeling and

54

HASLAM

giving a theoretical account of these parameters, without which they risk functioning in a manner akin to unanalyzed error terms within the theory. Other directions for further research readily come to mind. This reviewer has high hopes for continued work on individual differences, for applications of laboratory-based implicit cognition methodologies, and for a swelling of the encouraging voices within organization studies. The good news for prospective researchers is that so much remains to be done. All in all, this review of the research literature on the RMs presents a picture of growth and diversity. RMs theory has generated or seeded a wide variety of empirical studies and conceptual explorations, and these have, in turn, given a generally positive report on its standing as a theory. It seems reasonable to expect that its second decade sees further development and consolidation. REFERENCES Allen, N. B., & Gilbert, P. (1997). A social risk hypothesis of the evolution of depression. Unpublished manuscript, University of Melbourne. Baldwin, M. (2000, May). Relational schemas: The representation, activation, and application of interpersonal knowledge. Paper presented at the conference on relational models in the disciplines, Buffalo, NY. Caralis, D., & Haslam, N. (2002). Personality correlates of the relational models. Manuscript submitted for publication. Connelley, D. (1997). Hidden bias: The effect of relational models on assessments of human resource systems in a demographically diverse organization. Working paper No. 798, School of Management, State University of New York at Buffalo. Earle, T. C., & Cvetkovich, G. (1997). Culture, cosmopolitanism, and risk management. Risk Analysis, 17, 55-65. Earle, T. C., & Cvetkovitch, G. (1999). Social trust and culture in risk management. In G. Cvetkovich & R. Lofstedt (Eds.), Social trust and the management of risk (pp. 9-21). London: Earthscan. Earley, P. C. (1997). Face, harmony, and social structure: An analysis of organizational behavior across cultures. New York: Oxford University Press. Earley, P. C. (1998). Exploring new directions in cultural context: Applying Triandis' typology to organizational face theory. In J. L. C. Cheng & R. B. Peterson (Eds.), Advances in International Comparative Management, Vol. 12 (pp. 45-52). Stamford, CT: JAI Press. Fiske, A. P. (1990). Relativity within Moose ("Mossi") culture: Four incommensurable models for social relationships. Ethos, 18, 180-204. Fiske, A. P. (1991a). Structures of social life: The four elementary forms of human relations. New York: Free Press. Fiske, A. P. (1991b). The cultural relativity of selfish individualism: Anthropological evidence that humans are inherently social. In M. Clark (Ed.), Review of personality and social psychology, 12: Altruism and prosocial behavior (pp. 176-214). Beverly Hills, CA: Sage. Fiske, A. P. (1993). Social errors in four cultures: Evidence about universal forms of social relations. Journal of Cross-Cultural Psychology, 24, 463-494. Fiske, A. P. (1995). Social schemata for remembering people: Relationships and person attributes that affect clustering in free recall of acquaintances. Journal of Quantitative Anthropology, 5, 305-324.

2. RELATIONAL MODELS: AN OVERVIEW

55

Fiske, A. P. (2000). Complementarity theory: Why human social capacities evolved to require cultural complements. Personality and Social Psychology Review, 4, 76-94. Fiske, A. P., & Haslam, N. (1996). Social cognition is thinking about relationships. Current Directions in Psychological Science, 5, 143-148. Fiske, A. P., & Haslam, N. (1997a). Prerequisites for satisfactory relationships. In L. Meyer, M. Grenot-Scheyer, B. Harry, H.-S. Park, & I. Schwartz (Eds.), Making friends: Understanding the social worlds of children and youth. Baltimore: Paul H. Brookes. Fiske, A. P., & Haslam, N. (1997b). The structure of social substitutions: A test of relational models theory. European Journal of Social Psychology, 25, 725-729. Fiske, A. P., Haslam, N., & Fiske, S. (1991). Confusing one person with another: What errors reveal about the elementary forms of social relations. Journal of Personality and Social Psychology, 60, 656-674. Fiske, A. P., & Tetlock, P. E. (1997). Taboo tradeoffs: Reactions to transactions that transgress spheres of exchange. Political Psychology, 18, 255-297. Fiske, A. P., & Tetlock, P. E. (2000). Taboo tradeoffs: Constitutive prerequisites for political and social life. In S. A. Renshon & J. Duckitt (Eds.), Political psychology: Cultural and cross cultural perspectives (pp. 47-65). London: Macmillan. Foa, U. G., & Foa, E. B. (1974). Societal structures of the mind. Springfield, IL: Thomas. Folger, R., Sheppard, B. H., & Buttram, R. T. (1995). Equity, equality, and need: Three faces of social justice. In B. Bunker & J. Z. Zucker (Eds.), Conflict, cooperation, and justice (pp. 261-289). San Francisco, CA: Jossey-Bass. Forsyth, D. W. (1995). Commentary on Fiske's models of social relations. Psychoanalysis and Contemporary Thought, 18, 119-153. Francis, J. J., & Jackson, H. J. (in press). Dependency relations, not power relations, influence emotional vulnerability to self-discrepancies in the workplace. Personality and Social Psychology Bulletin. Goodnow, J. J. (1998). Beyond the overall balance: The significance of particular tasks and procedures for perceptions of fairness in household work distributions. Social Justice Research, 11, 359-376. Haslam, N. (1994a). Categories of social relationship. Cognition, 53, 59-90. Haslam, N. (1994b). Mental representation of social relationships: Dimensions, laws, or categories? Journal of Personality and Social Psychology, 67, 575-584. Haslam, N. (1995a). Factor structure of social relationships. Journal of Social and Personal Relationships, 12, 217-227. Haslam, N. (1995b). A grammar of social relations. Transcultural Psychiatric Research Review, 32, 41-57. Haslam, N. (1997a). Four grammars for primate social relations. In J. Simpson & D. Kenrick (Eds.), Evolutionary social psychology (pp. 293-312). Mahwah, NJ: Lawrence Erlbaum Associates. Haslam, N. (1997b). Personality disorders as social categories. Transcultural Psychiatry, 34, 473-479. Haslam, N. (1999). Taxometric and related methods in relationships research. Personal Relationships, 6, 519-534. Haslam, N., & Fiske, A. P. (1992). Implicit relational prototypes: Investigating five theories of the cognitive organization of social relationships. Journal of Experimental Social Psychology, 28, 441-474. Haslam, N., & Fiske, A. P. (1999). Relational models theory: A confirmatory factor analysis. Personal Relationships, 6, 241-250. Haslam, N., Reichert, T., & Fiske, A. P. (2002). Aberrant social relations in the personality disorders. Psychology and Psychotherapy: Theory, Research and Practice, 75, 19-31.

56

HASLAM

lacoboni, M., Lieberman, M. D., Knowlton, B. J., Molnar-Szakacs, I., Moritz, M., Throop, C. J., & Fiske, A. P. (2002). Watching people interact: The neural bases of understanding social relations. Unpublished manuscript, University of California, Los Angeles. Jackendoff, R. S. (1992). Is there a faculty of social cognition? In R. S. Jackendoff, Languages of the mind: Essays on mental representation. Cambridge, MA: MIT Press. Jackendoff, R. S. (1994). Patterns in the mind: Language and human nature. New York: Basic Books. Jackendoff, R. S. (1995). The conceptual structure of intending and volitional action. In H. Campos & P. Kempchinsky (Eds.), Evolution and revolution in linguistic theory. Washington, DC: Georgetown University Press. Jackendoff, R. S. (1999). The natural logic of rights and obligations. In R. S. Jackendoff & P. Bloom (Eds.), Language, logic, and concepts: Essays in memory of John Macnamara (pp. 67-95). Cambridge, MA: MIT Press. Kiesler, D. J. (1983). The 1982 interpersonal circle: A taxonomy for complementarity in human transactions. Psychological Review, 80, 185-214. Lickel, B., Hamilton, D. L, & Sherman, S. J. (2001). Elements of the lay theory of groups: Types of groups, relational styles and the perception of group entitativity. Personality and Social Psychology Review, 5, 141-155. MacCrimmon, K. R., & Messick, D. M. (1976). A framework for social motives. Behavioral Science, 21, 86-100. Meehl, P. E. (1992). Factors and taxa, traits and types, differences of degree and differences in kind. Journal of Personality, 60, 117-174. Mikula, G. (1999, July). Distribution of tasks and duties: Justice, efficiency and satisfaction. Paper presented at the General Meeting of the European Association of Experimental Social Psychology, Oxford, England. Mills, J., & Clark, M. S. (1984). Exchange and communal relationships. In L. Wheeler (Ed.), Review of Personality and Social Psychology (Vol. 3, pp. 121-144). Beverly Hills, CA: Sage. Parsons, T., & Shils, E. A. (Eds.). (1951). Toward a general theory of action. Cambridge, MA: Harvard University Press. Plutchik, R. (1980). Emotion: A psychoevolutionary synthesis. New York: Harper & Row. Poulson, B. (2000). Interpersonal conflict and relational models theory: A structural approach to injustice. Unpublished manuscript, City University of New York. Semedar, A. (2000). Relational models theory and depressive behaviour: Towards an interpersonal model of mood disorders. Unpublished B.A. thesis, University of Melbourne. Sheppard, B. H., & Sherman, D. M. (1998). The grammars of trust: A model and general implications. Academy of Management Review, 23, 422-437. Sheppard, B. H., & Tuchinsky, M. (1996). Interfirm relationships: A grammar of pairs. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational behavior (Vol. 18, pp. 331-373). Greenwich, CT: JAI Press. Sondak, H. (1998). Relational models and organizational studies: Applications to resource allocation and group process. In C. L. Cooper & D. M. Rousseau (Eds.), Trends in organizational behavior (Vol. 5, pp. 83-102). Chichester, England: Wiley. TenHouten, W. D. (1996). Outline of a socioevolutionary theory of the emotions. International Journal of Sociology and Social Policy, 16, 191-208. TenHouten, W. D. (1999a). Explorations in neurosociological theory: From the spectrum of affect to time consciousness. Social Perspectives on Emotion, 5, 41-80. TenHouten, W. D. (1999b). The four elementary forms of sociality, their biological bases, and their implications for affect and cognition. Advances in Human Ecology, 8, 253-284. TenHouten, W. D. (2000). On Durkheim's notions of time and mind in Australian aboriginal cosmology and social life. In R. Altschuler (Ed.), The living legacy of Marx, Durkheim and Weber: Vol. 2. Applications and analyses of classical sociological theory by modern social scientists. New York: Gordian Knot.

2. RELATIONAL MODELS: AN OVERVIEW

57

Tetlock, P. E., Kristel, 0. V., Elson, S. E., Green, M. C., & Lerner, J. S. (2000). The psychology of the unthinkable: Taboo trade-offs, forbidden base rates, and heretical counterfactuals. Journal of Personality and Social Psychology, 78, 853-870. Tetlock, P. E., McGraw, A. P., & Kristel, O. V. (2000). The limits of fungibility: Relational schemata and the meaning of things. Manuscript submitted for publication. Tetlock, P. E., Peterson, R., & Lerner, J. (1996). Revising the value pluralism model: Incorporating social content and context postulates. In C. Seligman, J. Olson, & M. Zanna (Eds.), Eighth Ontario Symposium on Personality and Social Psychology: Values. Mahwah, NJ: Lawrence Erlbaum Associates. Triandis, H. C., & Gelfand, M. (1998). Converging measurement of horizontal and vertical individualism and collectivism. Journal of Personality and Social Psychology, 74, 118-128. Vodosek, M. (2000a, August). Relational models and the outcomes of work groups: A confirmatory factor analysis. Paper presented at the Academy of Management Meetings, Toronto, Canada. Vodosek, M. (2000b, August). Relational models and their effects on relationship, process, and task conflict in work groups. Paper presented at the Academy of Management Meetings, Toronto, Canada. Vodosek, M. (2000c, August). The correspondence between relational models and individualism and collectivism: Evidence from groups. Paper presented at the Academy of Management Meetings, Toronto, Canada. Vodosek, M. (2003). Finding the right chemistry: Relational models, relationship process, and task conflict in culturally diverse research groups. Unpublished doctoral dissertation, University of Michigan. Whitehead, H. (1993). Morals, models, and motives in a different light: A rumination on Alan Fiske's Structures of Social Life. Ethos, 21, 319-356. Whitehead, H. (2000a). Food rules: Hunting, sharing, and tabooing game in Papua New Guinea. Ann Arbor: University of Michigan Press. Whitehead, H. (2000b). Eager subjects, reluctant powers: The irrelevance of ideology in a secret New Guinea male cult. In J. Mageo & B. Knauft (Eds.), Power and the self. New York: Cambridge University Press. Wiggins, J. (1980). Circumplex models of interpersonal behavior. In L. Wheeler (Ed.), Review of Personality and Social Psychology, Vol. 1 (pp. 265-293). Beverly Hills, CA: Sage. Wish, M., Deutsch, M., & Kaplan, S. B. (1976). Perceived dimensions of interpersonal relations. Journal of Personality and Social Psychology, 33, 409-420.

This page intentionally left blank

P A R T

II COGNITION AND CULTURE

This page intentionally left blank

C H A P T E R

3 Four Modes of Constituting Relationships: Consubstantial Assimilation; Space, Magnitude, Time, and Force; Concrete Procedures; Abstract Symbolism Alan Page Fiske

This chapter explores the question of how people constitute social relationships.1 The central thesis is that there are distinct natural ways to constitute each of the four elementary relational mods. The other thesis is that the manner in which people naturally constitute a type of relationship is closely linked to the medium in which they have the greatest facility in communicating about it, and to the primary form of cognitive representation of that type of relationship. These are also homologous with the channels through which children discover the local cultural implementations of that 'This chapter is the current formulation of observations I have been mulling over for about 20 years now. I have presented various versions and aspects of these ideas at the Philadelphia Anthropological Society in 1988; a symposium on Methods in Psychological Anthropology, sponsored by the Society for Psychological Anthropology at the 1988 meetings of the American Anthropological Association, Phoenix; the Sloan Cognitive Science Seminar at the University of Pennsylvania in 1989; the Social Psychology brown bag of the University of Massachusetts at Amherst in 1989; the Michigan Conference on Culture and Cognition in 1991; the UCLA Developmental Forum in 1997; and, in 2003, the Close Relationships Interest Group and the Cotson Institute of Archeology Pizza Talks at UCLA. Many people have provided valuable feedback along the way, notably John Lucy in his role as discussant at the 1988 SPA symposium. Some of the time I devoted to thinking and reading about all this was supported by NIMH grant MH43857.1 am indebted to Michael Sean Smith for persistent, insightful searches in the ethnographic literature and Cathy Lee for energetic and intelligent pilot work with diverse informants around UCLA. I thank Henning Brinckman for recommending and Lotte Thomsen for translating several Danish sources on land distribution. Nick Haslam, Roy D'Andrade, Barbara Fiske, Kevin Haley, and Lotte Thomsen read the manuscript carefully and made many helpful suggestions.

61

62

FISKE

relational mod, and, conversely, the mechanisms by which these implementations are culturally reproduced or transformed. The assumption behind this is that people have innately structured and motivated relational proclivities, called "mods" (Fiske, 2000; see also chap. 1, this volume). These mods are incomplete; they cannot be implemented to coordinate interaction without being linked to socially transmitted complements called "preos." Preos consist of prototypes or precedents indicating where, when, how, and with whom to implement the relational mods. In these terms, a crucial aspect of each mod is the type of preo that people "expect." That is, people use a specific medium to create and communicate about a mod, so natural selection and developmental experience tune people to use that medium when implicitly seeking the cultural preo for that mod. This medium, the channel for cultural transmission of the preos for each mod, resonates with a homologous mode of cognitive representation that informs and also derives from these constitutive and communicative practices. These abstractions will be easy to grasp when we come to palpable concrete facts. The question to begin with is, simply, how do people create relationships? Consider, for example, four constitutive processes: 1. A woman gives birth to her baby, nurses it, cleans it, carries it, and sleeps next to it. 2. A subject waits for permission to approach the king, who is seated on a dais in a great palace; the subject bows and addresses the king as "Your Highness"; the king responds using the royal "we." 3. Two football teams flip a coin to determine who kicks off and which goal they initially defend; when the clock indicates half of the time has elapsed, they switch. 4. Representatives of two corporations negotiate and then sign a detailed contract that explicitly spells out product specifications, quantities, and prices. One firm pays money for the other's products. Each of these sets of acts creates a meaningful framework for coordinating interaction. As these examples suggest, it is natural for people to constitute different kinds of relationships in different ways. There is a distinctive mode of constituting each of the four types of relational models that is universally intuitive and that young children innately expect. For children and adults, the social meaning of appropriate actions in this modality are immediately apprehended, inherently evocative, and experienced as normatively binding. That is, certain actions in a specific mode have unique constitutive affordances and motivational potential—they evoke emotions and elicit commitments that tend to establish, sustain, or transform the correspond-

3. CONSTITUTING RELATIONSHIPS

63

ing type of relationship.2 Each culture elaborates these natural modalities in particular ways and to different degrees, but people in most cultures create most of their relationships primarily with the media natural to the corresponding type of relationship. People constitute Communal Sharing (CS) relationships primarily through consubstantial assimilation: They connect themselves socially by connecting their material bodies. Consubstantial assimilation joins people indexically and enactively through the orifices and surfaces of their bodies, for example, by birth or commensalism. Likewise, nursing joins the mother and child through the expression and ingestion of breast milk, a bodily substance that connects them. Synchronous rhythmic movement also connects participants through their mirrored motion in dance, ritual, and military drill. In contrast, people constitute Authority Ranking (AR) relationships primarily by arranging persons in space, time, magnitude, and force. Higher rank is constituted by being above and in front, being more, coming earlier in time, and having greater "power" or "force." This physics of social relations is iconic, creating social relations by metaphorically mapping people onto position, quantity, and temporal order. For example, bowing, or addressing someone as "Your Highness," differentiates two people in space, putting the inferior below the superior. Equality Matching (EM) is constituted by balancing procedures that are "operational definitions" of equality. These concrete operations make participants equal by virtue of a procedure that evenly matches them. Such EM procedures include taking turns, flipping a coin, lining up on a starting line, voting, or distributing shares to participants with a one-to-one correspondence process. Finally, people constitute Market Pricing (MP) relations primarily through abstract symbols, especially numbers, semantic language, and the metasymbol, money. What is characteristic of MP symbols is their arbitrary meaning, depending almost purely on social convention. MP symbols include currency or digital accounts, numerals that indicate a price or an interest rate, and linguistic representations of the features of commodities or the terms of employment. Table 3.1 summarizes these constitutive modalities. For characterizing the distinctions among these modes of action I use a typology that builds on Peirce's (1931-1958, 1985) typology of sign systems. Peirce analyzed forms of cognitive and communicative representation without considering the use of signs to create and transform social relationships. However, his characterization of indexicality, iconicity, and symbol2

The concept of affordances comes from Gibson's (1979) ecological theory of perception; I use it here to indicate the constitutive and representational potential inherent in a given medium.

TABLE 3.1 Natural Modes of Constituting Relationships Communal Sharing

Authority Ranking

Equality Matching

Market Pricing

Consubstantial assimilation Indexical, metonymic

Social physics Iconic, metaphoric

Concrete operations Procedural

Arbitrary signs Symbolic

Common body essences Giving birth & nursing Sharing comestibles: cooked food (esp. meat) drinks stimulants & drugs sacrificial offerings Mixing & ingesting blood Scarification, body marking Rhythmic action in unison or ritual reenactment Body contact & sex Shared pain or bodily peril

Above In front Earlier Larger More numerous Greater force (more powerful)

Taking turns Reciprocating in kind Conducting a lottery 1:1 correspondence matching; counting out Aligning side-by-side Starting on a line Working in unison alongside Balloting Weighing in a pan balance

Money Prepositional language (e.g., bids & contracts) Gestural signs (as in bidding) Writing & graphic symbols Numbers & math Account books & digital accounts

Thesis: These are, respectively, the most motivating and emotionally evocative constitutive media for the four types of relationship, and induce the strongest moral commitments. The natural mode of constituting a type of relationship is integrated with the manner in which people communicate (represent, mark, signal) the relationship, and with its cognitive representation. The constitutive medium for each relationship is also closely linked to the manner in which children learn their culture, discovering the specific implementations of the elementary forms of sociality. This means that the constitutive modalities are selective filters that shape cultural invention, transmission, and transformation.

3. CONSTITUTING RELATIONSHIPS

65

ism is still very useful for understanding the constitutive aspects of CS, AR, and MP acts. Likewise, Piaget's (1932, 1945) analysis of concrete operations, although focused on stages of cognitive development, also illuminates the constitutive practices and communicative representation of EM. The core claim is that there is one distinct natural way to constitute each of the four fundamental types of relationship. True, people have innumerable ways of creating relationships, including a myriad of culture-specific mechanisms. But across cultures there is a pervasive tendency to use particular media to create particular relationships—and the natural mode of creating each type of relationship tends to have greater motivational and normative impact than other means. It is also easier for children, immigrants, and ethnographers to recognize. The impact and intuitiveness of the natural constitutive modes is what makes them more likely to endure and to diffuse across cultures. There are many theoretical and practical benefits to understanding how people constitute social relationships. The only source of moral obligation or emotional concern toward a person is a meaningful relationship with that person or with a third being. In the absence of such a relationship there is nothing to prevent simply using other people as a means to ulterior ends without intrinsic regard for their well-being. Understanding how social relationships are created might enable us to reduce the violence that results from lack of sociable relations. There are potential clinical applications as well. It is plausible to imagine that understanding how people constitute relationships would make it possible to develop interventions to help autistic patients and others who have difficulties forming meaningful social relationships.

Conformation The characterization of these four constitutive modes and examples of them comprise the body of this chapter. But before we get to that, I want briefly to indicate how the constitution of a relationship is integrated with other cognitive, communicative, developmental, and cultural processes. For our purposes, we can say that a social relationship consists of a structured coordination system that people use to generate, understand, plan, remember, evaluate, and sanction mutually oriented action. From a psychological perspective, social relationships are based on: 1. The cognitive deployment of a relational model to generate and interpret action, and hence to coordinate interaction. 2. A set of congruent emotions that motivate people to sustain interaction consistent with the model.

66

FISKE

3. Moral commitment to act in accord with the model, along with the right to impose sanctions for violating the model and implicit acknowledgment of the legitimacy of appropriate sanctions. Hence when people constitute a relationship, they bring into play a relational model to generate their own action, interpret others' actions, and evaluate their own and others' actions. They also evoke motivating emotions and invoke moral commitments to act in accord with the relational model, while implicitly acknowledging the legitimacy of appropriate sanctions. Furthermore, relationships are dynamic: Over time, relationships vary in intensity and in the manner in which they are implemented, and they end. Moreover, people may violate relationships, punish transgressions, and make amends. So the constitution of a relationship is not simply its initial creation—it also consists of the maintenance, reinforcement, modulation, transformation, sanctioning, redress, and termination of the relationship.3 Some acts, such as bearing a child or performing a blood bonding ritual, clearly constitute a relationship. Other acts might be better characterized as the conduct of the relationship, such as feeding the baby or the blood brother. However, I doubt that it generally makes sense to draw a sharp line between the constitution and the conduct of a relationship. Every relational act modulates the relationship, amplifying or attenuating, sustaining or jeopardizing the bond. For example, if a person offers to sell a car to a buyer for a specified sum of money, the monetary payment of a sum (equal to or different from the specified amount) is both constitution and conduct of the transaction. Drinking and eating together sustains a CS relationship among the commensals; refusing to eat together separates them. So, while we should keep this constitution-conduct dimension in the back of our minds and return to it at some point in the future, to simplify the analysis, in this chapter I am not going to distinguish these two aspects of relationships. But how do people constitute relationships? That is, what kinds of action make people perceive that a relationship exists, evoke the motivating emotions, and invoke the commitments that bind people to the relationship? And how do people sustain or change these cognitions, emotions, and commitments? That's our topic. In addressing this question, we need to keep in mind that relationships are constituted interactively through the coordination of human minds. And relationships are culture-specific implementations dependent on soUnderstanding how social relationships and societies are constituted is the core problem of social science. My formulation of the problem has many independent roots in philosophy, political science, psychology, and anthropology—roots that I discuss in Fiske (2004a).

3. CONSTITUTING RELATIONSHIPS

67

cially transmitted complements for innate proclivities (Fiske, 2000). This means that the constitution of social relationships entails four more processes. First, constituting a relationship requires communicating about it. So, when initiating or conducting a relationship, people signal to each other and to observers their social expectations, emotions, and obligations. Goffman calls these "tie signs" (1971, p. 194ff.) or "ritual displays" (Goffman, 1976—see especially p. 3, point #7). Second, the structure of each relational model is cognitively encoded in neural processes: People must store, process, and retrieve relational information. That is, they have a mental representation of a relationship. The typical manner of constituting a relationship informs people's mental representation of what the relationship is, and in turn depends on the form of that mental representation. Likewise, cognitive representations tend to shape and are in turn partially shaped by communicative representations. So the semiotic system through which people display or refer to a relationship informs their cognitive encoding of the relationship (cf. Vygotsky, 1934). At the same time, a person's communications, intended to inform the other's cognitive representation, are in turn products of his own mental encoding. Furthermore, in order to constitute any relationship, children must first discover the local cultural implementations of the indeterminate relational mods. That is, people complete their innate mods with socially transmitted complements—"preos"—that specify when, where, how, and with whom to implement the model (Fiske, 1991, 2000). Children discover these preos by observing people creating and signaling relationships; to a considerable extent, the child's mental representations of relationships consist of integrated memories of these constitutive and communicative acts. Finally, the individual child's development of the capacity to participate in culturally informed relationships can also be analyzed from the dynamic, historical, and collective perspective of communities and social networks. Cultural reproduction, diffusion, and transformation of social relations consist of population-level aggregations of mental representations and communications (see Richerson & Boyd, 2004). Hence cultural transmission is a product, integrated across the population, of individuals learning the preos that enable them to participate in the culture.4 Constitution, cognitive representation, communication, cultural completion, and cultural transmission are analytically separable. But they are in4

0f course, culture is also transmitted indirectly through representations in artifacts such as images, writing, recordings, programs, socially transformed landscapes, and tools embedded in technical routines. But insofar as we are concerned here with the transmission of cultural forms of social relations, for the time being we can focus on the principal "direct" person-to-person process, the core of which is observation and imitation. Furthermore, cultural transmission through artifacts is necessarily mediateds by the cultural meaning of the artifacts, that is, by human cognition and communication.

68

FISKE

herently connected in a multidirectional causal web. Because they are, to a considerable degree, mutually determinative, it will often be useful to treat all of these aspects of a social relationship as a coherent system. I refer to this system as the conformation of the relationship. The interdependence among these five components of the conformation of a relational model is not total; there are always important inconsistencies. To refer to these incongruencies we can borrow the geological term "unconformities." They result from distinct factors impinging on each component, distorting the mappings among them. For example, in a system of billions of neural connections there are distinctive neurophysiological and anatomical constraints and possibilities for encoding, storing, and retrieving social information. These factors affecting the neural system differ from the physical and semiotic constraints and possibilities inherent in human communication. Nevertheless, the causal integration among the five aspects of each conformation system tends to maintain close correspondence among them, so on the whole they will tend to be more or less homologous. To the extent they are homologous, they form a conformation system. Later in this chapter I make a more detailed theoretical case for the integration of the five aspects of conformation systems, although at present there is little empirical evidence about this integration. Indeed, there is not yet much evidence relevant to my claims about systems of cognitive representation, cultural completion, or transmission. However, there is ample and strikingly consistent ethnographic evidence regarding the distinct ways that people constitute and communicate each of the four relational mods. That will be our focus. The following sections of the chapter characterize the four conformations, illustrating them with examples from diverse cultures. Of course, these illustrations could be criticized as selective and I have extracted them from the cultural context which provides important aspects of their full meaning. But my thesis is precisely that the core conformations of social relationships are not essentially relative to the particular culture in which they function. On the contrary, I aim to show that there are many clear and pervasive patterns across the world's cultures. A thicker description of these conformations in one culture, the Moose of Burkina Faso, is in preparation (Fiske, 2004a). As it turns out in that case, at least, the thicker the description the stronger the evidence in support of this theory. My treatment of the four relational mods in this chapter is very unequal: There is quite substantial ethnographic material on Communal Sharing, a large quantity on Authority Ranking, some on Equality Matching, and much less on Market Pricing. To a considerable extent, this imbalance reflects biases in ethnographic reporting and, perhaps, limitations in my imagination regarding where to locate relevant materials in other social science literatures. For example, wider reading of the economic and political science lit-

3. CONSTITUTING RELATIONSHIPS

69

erature might have evened out the treatment of the four mods. But the inequality also appears to reflect an intriguing fact about social relations: There are substantial differences in the extent to which people have elaborated the basic media of the four conformations. It is easy to imagine explanations for this but difficult for me to evaluate their plausibility, so I will not speculate about it. Each of the following sections begins with a brief overview of the conformation system, followed by detailed ethnologic analyses.

CONSTITUTING COMMUNAL SHARING

The primary mode of enacting Communal Sharing is consubstantial assimilation: making the substances, surfaces, or motions of the body equivalent. The core of consubstantial assimilation consists of giving birth, nursing, and connecting people through comestible substances. This reflects and is reflected in the cognitive representation of people in a CS relationship as having some fundamental bodily essence in common. People may be unable to articulate the exact nature of this shared substance, but in some cultures this intuitive sensibility is "hyper-cognized"—elaborated into a salient discourse on the substances of sociality and the processes of their transmission. The cultural articulation of consubstantiation is particularly notable among the cultures of New Guinea (e.g., Herdt, 1982; Poole, 1985). The most pervasive, powerful forms of consubstantial assimilation include birth, nursing, feeding a person, eating and drinking commensally, sharing comestible substances such as tobacco, taking drugs together, or transferring blood in bonding rites. Under appropriate circumstances, merging bodies through sexual intercourse also creates powerful bonds of consubstantiality (see Pitt-Rivers, 1973, pp. 92-93). Conceiving a child together, followed by the woman's gestating and bearing a child, may also create powerful bonds between a man and woman (e.g., Evans-Pritchard, 1951). To a significant degree, affiliative body contact also is constitutive of CS. In addition, people make their bodies equivalent by modifying them in the same way, especially through circumcision, excision, and scarification, but also in somewhat lesser degree with the configuration of hair, face or body paint, clothing and other markers such as rings, feathers, or headgear. Of course, people often construct large-scale ethnic or racial CS relations out of less mutable body features such as skin color, facial morphology, or hair types. The prototype form of immutable, intense CS is close kinship, and kinship is virtually always understood to be based on some shared essential substance such as blood, bones, or genes (for the ancient Arab world, see Smith, 1885; on American cultural conceptions, see Schneider, 1980; for the ontogenetic basis of this, see Hirschfeld, in press).

70

FISKE

In addition, people make and mark CS by acting the same way: by imitating, by reproducing remembered rituals, or by moving in the rhythmic synchrony of dance, drill, political performances, or religious ceremonies. Initiation rites and other rituals connect people in CS through combinations of consubstantial acts and, typically, repetitive, stereotyped action in unison. In addition, initiations and other traumatic events may be powerful constitutors of CS through the shared experience of extreme fear, privation, or pain. What Is "Consubstantial Assimilation"? Consubstantial assimilation is the prototype of what Tylor (1871) called contact and Frazer (1890) termed contagion. People apprehend physical connection among people as transferring fundamental social properties, making people alike. This is the implicit idea involved in commensalism, mixing of blood, and simple body contact.5 In an interesting sense, consubstantial assimilation is metonymic, since particular aspects of the body—a part of the self—stand for the whole social person. Equivalence of specific body substances, surfaces, and actions represent the social equivalence of selves. Although we are looking at the entire conformation system, not just semiotics, it is also illuminating to adopt the perspective of Peirce's (1931-1958) typology of the bases of representation of signs. Peirce's analysis treats signs as the media for both communication and thought, which he regarded as intrinsically linked. In terms of Peirce's most important dimensions of signs, what is particularly prominent about consubstantial assimilation is its indexicality. The indexical aspect of a sign is the meaning conveyed by the causal relation between referent and sign, or, more generally, their natural spatiotemporal co-occurrence. Indexicality derives from the intrinsic co-occurrence of sign and referent, beyond the conventional association that results from human uses of the index. There are several aspects of this indexicality in consubstantial assimilation. Mammalian biology is the source of much of the indexicality of consubstantial assimilation. Mothers give birth to live young and nurse them; most mammals are in direct contact with their mother and/or siblings during much of their infancy. Moreover, primate infants generally cling to their mothers and are held by other kin in several species. Human infants have to be fed by their kin for many years and a child almost always eats commensally with at least its mother 5 While he wrote about magic in "primitive" cultures, Tylor believed that all humans had the same psychology. Indeed Rozin and Nemeroff (1994; Rozin, Nemeroff, Wane, & Sherrod, 1989) have experimentally demonstrated the pervasiveness of contiguity in the "magical" cognition of contemporary college students reacting to contamination or positive contact histories of food and clothing.

3. CONSTITUTING RELATIONSHIPS

7I

and young siblings. Close kin physically resemble each other and are typically groomed and dressed so they further resemble siblings and other family members. In short, these practices are natural results of the core prototype of CS, close kinship. Beginning in infancy, humans apprehend this causal connection of consubstantial assimilation with CS, experiencing it as a functionally necessary co-occurrence. (At the end of this section I return to the phylogenetic and ontogenetic sources of consubstantial assimilation, but here my point is that consubstantial assimilation is naturally connected to CS relationships through the experiences that are intrinsic to primate and specifically human ontogeny.) Action that connects bodies or makes them equivalent is indexical in a more particular sense: It is a token performance of the relationship. Nursing a baby is a crucial act of nurturance, creating a care-taking relationship by enacting its core entailment. More generally, when people participate in a commensal feast or merely share a comestible item such as tobacco, ingestion together of the shared substance is an instance of the sharing on other occasions and in other domains that the commensalism motivates and commits participants to continue. Likewise, initiation rites are tokens of a lifetime of shared hardship and common fate—tests of fortitude and performances of commitment. In the terms of a long-recognized distinction, consubstantial assimilation consists of practical motoric skills or habits, rather than articulate conceptual knowledge. Marett (1929/1909, p. xxxi) and his student James (1917) argued that religious rituals are emotional motoric practices, not theoretical statements. They criticized what they construed as Durkheim's (1893) intellectualist theorizing of social solidarity, but they agreed with Durkheim that rituals are acts of identity, representing and reconstituting collective cohesion. Sapir (1927) described how people unconsciously organize language, exchange, and accumulation with reference to a myriad of historically transmitted cultural patterns of behavior that they take for granted as given in the nature of things and which they cannot understand in explicit terms. Later Durkheim's nephew and student, Mauss (1936), implicitly accepted this critique in articulating the idea of habitus, that is, culturally transmitted bodily practices (see Bourdieu's elaboration in 1980). This same contrast was developed by philosophers and psychologists interested in memory, beginning with the distinction made by Bergson (1911) and Russell (1921) between motoric habits and recollections of specific events. Later Merleau-Ponty (1945) contrasted "praktognosia" with articulate conceptual understanding, while Ryle (1949) opposed "knowing how" with "knowing that." Schutz (1977/1951) contrasted social interactions involving communication through concepts (whose meaning can be grasped at a given moment in time) with communication based on meaning that is inherently temporal, based on a joint experience of the flux of activities that

72

FISKE

are articulated temporally in a step-by-step sequence. Bruner (1964, 1966) contrasted symbolic and iconic cognitive representation with sensorimotor practices, which consists of things people do, rather than ideas they express.6 Further cognitive science research has verified that humans use several distinct, dissociable systems for encoding experiences, including semantic memory for concepts, episodic memory for specific events, and procedural competence for habits and skilled performance (Tulving, 1995; Squire & Knowlton, 1995; Schacter, Wagner, & Buckner, 2000). This established distinction is useful here because consubstantial assimilation is a form of cultural transmission and a system of cognitive representation based primarily on tradition embodied in motoric practices. It also involves episodic biographical memories of key events such as giving birth or being initiated, along with typified experiences such as participating in identity-forming rituals or eating together. Explicitly articulated semantic propositions are not central to the conformation of CS in face-to-face communities, although semantic representations of CS can become significant in large-scale societies with highly developed systems of mass communication. Let us examine the most important of the myriad manifestations of consubstantial assimilation. Commensalism Those who eat and drink together are by this very act tied to one another by a bond of friendship and mutual obligation. (William Robertson Smith, 1889, p. 265)7 The very act of eating and drinking with a man was a symbol and a confirmation of fellowship and mutual social obligations [in early Arabian culture]. The one thing directly expressed in the sacrificial meal is that the god and his worshippers are commensals, but every other point in their mutual relations is included in what this involves. Those who sit at meat together are united for all social effects; those who do not eat together are aliens to one another, without fellowship in religion and without reciprocal social duties.... Among the Arabs every stranger whom one meets in the desert is a natural enemy, and has no protection against violence except his own strong hand or the fear that his tribe will avenge him if his blood be spilt. But if I have eaten the smallest morsel of food with a man, I have nothing further to fear from him; "there has been salt between us," and he is bound not only to do me no harm, but to help and defend me as if I were his brother. (Smith, 1889, pp. 269-270) 6

Bruner's typology is based largely on Piaget's; for sensorimotor schemas, see Piaget (1932, 1945). Anthropological evidence is cited here and below in the "ethnographic present": the historical moment which the ethnographer selects to characterize the society—sometimes a point prior to the date of fieldwork.

3. CONSTITUTING RELATIONSHIPS

73

To eat together signifies amity; to refuse, or to be prohibited from doing so, signifies its absence. (Meyer Fortes, 1969, p. 236) Sharing food is one of the fundamental ways that one can display, establish, and maintain personal intimacy. By contrast, lack of sharing expresses social distance. (Paul Rozin, 1999, p. 22)

From the very beginning of modern social science, as soon as a wide sample of ethnographic reports became available, comparative ethnologists immediately recognized the worldwide pattern that I am characterizing as constituting CS through consubstantial assimilation. As the quotations suggest, William Robertson Smith (1889) explored the origins of Christian communion rites, showing that sacrifice in early eastern Mediterranean societies is an act of eating together that creates a profound bond between worshiper and god. This religious communion derives from the meaning and moral effect of commensalism among humans. Smith described the kin group's collective responsibility for vengeance and their collective liability for homicide, because, he wrote, kinship is a common life that kin represent as entailed by their physical unity "as one living whole, a single animated mass of blood, flesh and bones, of which no member could be touched without all the members suffering" (p. 274). This connection is formed through birth from the mother's body, then by suckling, then by feeding, "so that commensality can be thought of (1) as confirming or even (2) constituting kinship in every real sense" (p. 274). Smith saw the representation of a physical connection in blood and nourishment as the source of the binding moral force of parenthood (p. 41). Ritual commensalism then has the potential to constitute bonds between non-kin such as blood brothers, because "any food which two men partake of together, so that the same substance enters into their flesh and blood, is enough to establish some sacred unity of life between them" (p. 313). Rites of mixing and ingesting blood have the same effect. Less frequently and often less emotively, transfer or contact with other bodily substances such as tears or hair can create connections (pp. 323-335). Extending this mechanism, people in early eastern Mediterranean societies also formed bonds by indexical acts such as transfer or contact with clothing (in whole or fragment), ornaments, or even personal weapons (pp. 335-336). Likewise, putting on the skin of a sacrificed animal connects the sacrificer to his god (pp. 435-438). But, Smith concludes, one of the most culturally pervasive, historically enduring, and morally powerful acts of connection among humans is the communion act of partaking in the flesh and libations of sacrifice, cementing bonds among the humans while it merges them in vital union with their god. Hartland's (1895) interpretation of ritual sacrifices corresponds with that of Smith. Hartland writes, "Their great aim is union with the deity. It is attained by placing in contact with him something already part of ourselves,

74

FISKE

as our blood, hair, clothing, or other property; or else the blood of a victim of which we are about to consume the remainder" (p. 243). James (1917) agreed that rituals and sacrifice in particular are enactments of social emotions, expressing a universal desire for union with the deity. The member of a totemic group who perceives himself to be descended from his totem "endeavors to come into communion with his totem by assimilating the flesh and blood of the sacrosanct animal" (p. 238). The Christian Eucharist exhibits this principle as the congregation joins in eating the body and blood of Christ, thereby uniting with Him and each other. Taking communion—consuming the body and blood of Christ—is a defining act that constitutes a person as Roman Catholic. Sacrificing an animal, totem or not, and pouring libations, are acts of commensalism that join people to their gods through the substances offered up and jointly consumed (see Fortes, 1945, chap. VII; and Barth, 1975, chap. 21 for good examples). Furthermore, sharing in the beer poured in libations and the meat of the animals offered-up at these sacrifices defines the group and instills a sense of membership in a collectivity (Durkheim, 1893). Hartland (1895) recognized that all traditional societies are organized along kinship lines, which people implicitly understand as the union of connected parts of one physical body. This collective cognitive representation, this cultural model, is reflected in conceptions of common blood resulting from descent, or from the mixing and joint consumption of blood to create ritual bonds (p. 237). Hartland observed that commensalism has the same meaning and effect as shared blood, whether in its usual function among kin, in hospitality, or in ritual incorporation of outsiders: "Eating together is—not merely on solemn occasions, as the sacrifice of the totem-beast, but in a lesser degree at other times—an act of communion. The sharing of a common substance as food unites those who partake of it in a common life: it makes them parts of one another: they incorporate one another's substance" (p. 249). Clan solidarity is marked and sustained by commensalism: "This unity of the clan is constantly renewed by the common meal, where the same food is partaken of, and becomes incorporated into the essence of all who share it. ... The common meal was thus the pledge and witness of the unity of the kin, because it was the chief means, if not of making, at least of repairing and renewing it" (pp. 277-278). Likewise, eating a ritual food together or feeding the spouse is a key component of marriage ceremonies joining husband and wife in many societies (e.g., Hartland, 1895 on ancient Greece; Whiting, 1941 on Kwoma of Papua New Guinea). Later, Crawley (1902) added to this insight, emphasizing the importance of bodily contact in creating social relationships: skin to skin touching, ingestion of one another's blood or saliva, commensal eating, and sexual union. Conversely, he pointed out, taboos on contact correspond to social difference. Concurring with Crawley, van Gennep (1908/1960) emphasized that

3. CONSTITUTING RELATIONSHIPS

75

"eating and drinking together . . . is clearly a rite of incorporation, of physical union" (p. 29). He also noted the prevalence of direct bodily contact, including coitus between members of two groups and collective sexual license, "the precise equivalent of a communal meal participated in by all members of a particular group" (p. 170). Van Gennep emphasized that rites of incorporation typically join a person to a group through actual contact (e.g., a slap, a handclasp), exchanging gifts of food or valuables, eating, drinking, smoking a pipe together, sacrificing animals, sprinkling water or blood, anointing, being attached to each other, being covered together, or sitting on the same seat. Indirect contact may occur . . . through touching simultaneously or one after the other a sacred object, (p. 28)

Although they rarely cite these early sources, many modern ethnographies confirm the observations of Smith, Hartland, Crawley, and van Gennep. While they eschew universalizing comparison, many ethnographies, even those written by ardent cultural relativists, report that, in their particular culture, commensalism is the medium for creating and sustaining the strongest, most intimate bonds (e.g., Fajens, 1985, pp. 379-381 on New Britain in Papua New Guinea; Lutz, 1988, p. 129 on Ifaluk in Micronesia). Rabain (1979; Zemplini-Rabain, 1973) points out the importance of body contact, posture, proximics, and the giving and sharing of food—especially nursing and egalitarian exchange of food among siblings—in the molding of social relations among Wolof children in Senegal. Breast-Feeding Giving birth and suckling a child constitute some of the strongest, most reliable and enduring CS relationships. Hirschfeld and Gelman have shown that American 4-year-olds intuitively believe that breast-feeding an adopted child transmits important physical qualities; 5-year-olds no longer think that way (cited in Hirschfeld, in press). Indeed, in a number of polygynous patrilineal cultures of Africa, the closest kinship relations are those among children of the same mother, and such people often characterize their relationship as having "shared the same breast." Indeed, nursing a baby has the potential to create very strong bonds even when the woman has not borne the child she nurses. In many African and Mediterranean cultures, women may suckle each other's children, which creates an important permanent CS bond between the children who have shared milk from the same breast (e.g., Smith, 1889, p. 274; 1903, pp. 57, 176ff.; Davis, 1977; Altorki, 1980; Dettwyler, 1988; Fildes, 1988). In nearly every such culture, this "milk-kinship" entails prohibitions on sexual relations and marriage between the woman and her nursling and, more practically significant, between people who

76

FISKE

nursed at the same breast. Codified by Mohammed, these prohibitions became part of Islamic law. Furthermore, in a number of societies, sharing human milk or exchanging infants to be suckled is a way to make indissoluble peace, and commit both sides to mutual aid (Hollis, 1905, pp. 321-322; Beidelman, 1963, p. 334; Khatib-Chahidi, 1992). In some of these societies, consuming each other's blood (see below) is an adjunct or an alternative to sharing human milk. In a few societies, a man can assure himself of safety and even protection from his enemy if he manages to suckle momentarily from his enemy's wife (Lindblom, 1920, p. 141; Pitt-Rivers, 1968; Jaymous, 1981, p. 213). Beyond the powerful evocative impact of sharing breast milk, the ethnographic literature is replete with accounts showing that other acts of commensalism create and enhance communal relationships (see Richards, 1932 for southern Africa and Watson, 1987 for Hong Kong). For example, describing the Baktaman of Papua New Guinea, Barth (1975) writes, "Those who eat together are of one sort, ritually; and those who share food literally place their life and fate in each other's hands. . . . There is a strong and clear imagery expressing unity and mutual trust between members of a sodality" (p. 197). Similarly, among the Kaluli of central New Guinea, Schieffelin (1976) writes that "Food as gift or hospitality is the main vehicle for expression of friendly relations to anyone, kinsman or acquaintance" (p. 27). Giving and sharing food is the most important medium of Kaluli relationships; "it conveys affection, familiarity, and good will" (p. 48) and "does not merely express a social relationship; it validates and develops it" (p. 63). Giving and sharing food are the principal constitutive processes establishing or adjusting relationships, Schieffelin writes, "actualizing" potential relationships so that they become "socially real." As in many other cultures, ritual sharing of cooked meat, in particular, expresses affection and creates a special bond of identification and solidarity; after ritual meat sharing, Kaluli participants subsequently address each other by the name of the food: for example, "my bandicoot" (p. 50). Hospitality and Sociability

The sacred duty of hospitality (^evirj) is a pivot of the Odyssey and Iliad (Kakridis, 1963, pp. 86-108; Reece, 1993; Yamagata, 1994, pp. 28-39, 40-41, 135-136, 163-164). In general, Homeric Greeks recognized no moral or legal obligations toward strangers, preying on anyone they encountered. However, it was a sacred duty to provide hospitality to supplicants; a host had to receive any stranger without asking his identity. Homer's (perhaps ideal) representation of ancient Greek hospitality indicates that the host feasts the anonymous guest with meat and wine and other foods, and then either guest or host may toast the other with a goblet of wine. This meal is charac-

3. CONSTITUTING RELATIONSHIPS

77

terized as having a sacred quality. After the stranger reveals his identity and exchanges information with the host, the host may provide entertainment. The guest may pronounce a blessing on the host, who may invite his guest to share in a libation or sacrifice. The host provides a bed and has the guest bathed by women of his household; the guest is offered clothing. When he departs, the host may initiate an enduring bond of guest-friendship (^eivoQ by giving his guest a special gift. The host may provide a departure meal before they pour a libation and exchange farewell blessings. This ritual of consubstantiation obliged the host, binding him to aid and protect his guest—even if they were otherwise enemies—and then to send or take him safely to his next destination. Zeus himself was the god of hospitality, avenging violations of such obligations and bonds. Comprehensive authoritative reviews of the ethnographic literature show that the constitutive efficacy of commensalism is universal. In a famous review, Sahlins (1965) analyzes patterns of exchange in traditional societies, noting that "food dealings are a delicate barometer, a ritual statement, as it were, of social relations, and food is thus employed instrumentally as a starting, a sustaining, or a destroying mechanism of sociability" (p. 215). Compared to other substances, Sahlins finds that food is more likely to be shared than to be exchanged in balanced reciprocity (EM), distributed up and down a hierarchy (AR), or transacted in antagonistic or selfish economic interactions (asocial and MP relationships). Sahlins cites numerous ethnographies that describe food sharing as constitutive of what he calls "generalized reciprocity" (CS) among kin or other intimates, or as a way of cementing friendly relations with others. Likewise, reviewing the literature on eating together, Sobal (2000) notes that "Commensality is widely recognized as an important component of initiating, building, confirming, deepening, and reaffirming social connections" (p. 124), while "commensal circles demarcate 'us' versus 'them' by consumption of 'our foods' and not 'their foods'" (p. 123). Merely simultaneous eating or drinking or smoking have little effect: What evokes the emotional bonds is ingesting the same substance from the same source, or feeding someone.8 Everyday meals or sharing of tobacco or alcohol are minimal enactments of this, while ritual meals and celebrations are more evocative. Very often, consumption of a particular key food or set of foods is uniquely important in creating ethnic identity and difference (Brown & Mussell, 1984). Eating crawfish is a core constituent of being Cajun (Gutierrez, 1984). Commensal consumption of the identity-defining food, such as kimchee in Korean culture, incorporates and sustains ethnic membership, often more effectively than language or legal citizenship (Lee, 8 A number of contemporary movies depict bonding through feeding or commensalism. A charming one is Mostly Martha (Bavaria Film; Pandora; distributed by Paramount Classics).

78

FISKE

2000). Strathern (1973, p. 29ff.) shows, based on New Guinean ethnographies, that eating the food grown on the same land can link people in kinship because they perceive it as giving them common substance. Degrees of Intimacy The communicative meaning, emotional evocativeness, and inclusivity of different ways of transferring food compose a spectrum—perhaps a Guttman scale—in the constitutive efficacy of food sharing (Miller, Rozin, & Fiske, 1998). Mere transfer of food as such does not create a CS relationship: Theft, or the purchase of a meal at a restaurant, has no effect. A gift of uncooked food may be weakly constitutive. Consuming the same food from distinct sources or eating in close proximity has a modest effect: Choosing to drink Guinness along with everyone else is a way of belonging, more intimate than drinking merlot while everyone else drinks Guinness. Consuming something provided by another person is generally a bit stronger: "What will you have? Try some of this single malt I just discovered." Depending on the context, consuming food or drink prepared by a person may be distinctly more meaningful, especially if the person prepared it especially for the recipient (e.g., a batch of cookies). Eating the same food divided into distinct portions has a moderate effect, but true commensalism generally marks and makes a stronger bond: Consider drinking out of the same glass or passing a joint around. (The Norman Rockwell print of an adolescent boy and girl sipping through straws from the same soda captures the feeling.) In many parts of the world, families and other close associates eat commensally, reaching into the common food bowls or passing a calabash of beer around. Putting food into another's mouth is an index of greater intimacy (see Goffman, 1976, p. 35), but Americans, at least, perceive still greater intimacy when the feeder has already partaken of the proffered food (e.g., a woman bites a French fry and feeds a man the second half; Miller, Rozin, & Fiske, 1998). In many traditional societies, mothers and other adults chew food and then give it to infants and toddlers. Still more intimate than this is ingesting the substance of the other's body, especially by suckling at the breast. In a few New Guinea societies, the ultimate expression of affection and desire for perpetual union with deceased close kin is to eat their bodies (Gillison, 1983; Lindenbaum, 1979; note Freud's, 1913, 1939 theories). Different comestibles may have differing constitutive efficacy. Douglas (1975, pp. 256-258) stated that among the English middle-class, eating together is more intimate than drinking together. Co-consumption of meat, alcoholic beverages, and psychoactive drugs are often more strongly constitutive than other substances. Beyond its creative and sustaining effects, commensalism operates as a mechanism for transforming or reorienting relationships. Fortes (1966, pp.

3. CONSTITUTING RELATIONSHIPS

79

9-10) describes an ideal type of life cycle in African and other traditional societies in which "changing structural relationships" are constituted by shifts in co-eating and co-sleeping. First the infant nurses at the mother's breast and is held in her arms; at the next stage, toddlers eat with her and sleep next to her. Older boys begin to eat and sleep with their father and brothers. In the next phase, boys and girls eat and sleep with same sex peers. There may be initiation rites [which involve new, distinctive commensal and co-sleeping sets], followed, we may note, by conjugal eating and sleeping. In Africa, marriage entails wives preparing food for their husbands to eat, but traditionally, adult men and women, even spouses, never eat or drink together in most African societies. Bonding With Blood or Spit Maternally derived "blood" and milk and comestibles commensally ingested are not the only ways of connecting through shared substance. In a large number of societies, mutual ingestion of blood creates indissoluble CS relationships. Based on textual sources, Smith (1903, pp. 56-62) described inviolable bonds in early Arabia formed when two men mixed and ingested each other's blood. Based on fieldwork among the Azande of Central Africa, EvansPritchard (1933) provided the classic ethnographic account of blood-brotherhood; de Dampierre (1967) later described similar practices among the neighboring and very closely related Nzakara. Azande and Nzakara men form an extremely strong bond by making incisions and consuming each other's blood. As they do this, they speak to the blood, invoking its power to destroy a blood-brother who harms his partner, and spelling out benefits they will provide to each other. They then toss down gifts, usually large knives, that begin the relationship. Blood-brothers should be hospitable, feeding each other and offering beer, giving gifts, never refusing any practicable request, and providing refuge to a blood-brother being pursued. However, blood-brothers refrain from making immoderate, unreasonable, or unnecessary demands on each other. A man has the right to the head of any animal that his blood-brother kills in the hunt (for food). "A man constantly eats meals at his blood-brother's homestead and is invited there to beer parties. When a man kills a large beast his blood-brothers come to ask him for a share of the meat" (Evans-Pritchard, 1933, p. 389). Blood-brothers should warn each other of danger (including impending action by chiefs), stand up for each other in quarrels, never speak ill of each other to a chief, pay fines to save each other from vengeance or execution, and give each other scarce ritual goods. In particular, an Azande or Nzakara blood-brother should not refuse his partner's request to give his daughter in marriage. Most of these covenants are between neighbors, and covenants may result in establishing adjacent residences. A man has free access to his blood-

80

FISKE

brother's homestead and is treated as a special guest there, and he may even take up residence there. Men also form these pacts with men from other, potentially hostile tribes to assure safe passage when traveling in dangerous territories, or to facilitate material exchanges. In short, this bonding ritual created the closest, most trusting, intensive, and extensive relationship in Azande or Nzakara society. And violations of this commitment were thought to destroy an entire kin group. Very similar blood-bonding relations have been important in numerous other African and world societies, where, likewise, they were often the closest, most trusted of all, relationships (see the comprehensive review by Tegnaeus, 1952; also Beidelman, 1963; Brain, 1976; Eisenstadt, 1956). Typically, blood partners were more reliable and more generous to each other than even close kin. Trumbell (1885) argued that the Christian communion ritual is historically derived from this kind of blood-sharing covenant. Even in cultures where blood-sharing bonds are not institutionalized, they are often spontaneously invented (see the fictional account in Wells, 1997, and the movie based on it). CS bonds can also be constituted with spit. Africans traditionally did not kiss, but in Guinea, Laye (1966) describes his mother spitting on him when he left for boarding school, an act of bonding and blessing practiced in some other societies as well (including Greece, and among English school boys at some times—see Hartland's, 1895, p. 259 citation of Henderson, 1891). Food and Sex Taboos People make their bodies alike by sharing foods or transferring body substances; conversely, they differentiate "we" from "they" by not eating what "they" eat or not having sex with "them." People constitute many important CS relationships by not eating the same thing, or abstaining from particular sexual relations (Freud, 1913, 1939). In effect, the purity of the shared substantial essence of the group depends on all members respecting the same taboos against incorporation of particular polluting substances or polluting sexual relations. Observing a food taboo marks a social distinction contrasting those who taboo with those who do consume the substance. In her book, Food Rules, anthropologist Harriet Whitehead (2000) writes, "The one point upon which every theorist of food taboos seems to agree is that the taboos will among other things mark out social identities. . . . In turn, commensality is deeply involved in 'we' feeling and group affiliation. The point is so obvious as to be considered boring in most academic discussions of the subject" (p. 59). Innumerable ethnographies report that clans constitute their identity by not eating a particular species of animal. (In some cultures there are special and dramatic ritual exceptions to this principle, when the clan consumes the totem animal together.)

3. CONSTITUTING RELATIONSHIPS

8I

Religious and caste identities often hinge, in part, on food taboos observed in practice or ideally. Abstinence from sex and from any form of eating, drinking, or other ingestion during daylight hours of the lunar month of Ramadan, along with not eating pork or consuming alcohol at any time, are important in constituting Islamic identity. Abstinence from food and drink on Yom Kippur, not eating pork or shellfish, and observing other kosher restrictions on food and food combinations are important in constituting conservative Jewish identity. Fasting during Lent is part of what defines being Christian. Hindus, and Brahmins above all, distinguish themselves by not eating meat (especially beef) or drinking alcohol. The impurity of untouchable castes is legitimated by the imputation that they eat beef, as well as animals that died without being slaughtered (Simoons, 1994, p. 114). South Asian castes further delineate themselves by not eating with, not marrying, and not having sexual relations with, members of lower castes. (Caste membership is constituted matrilineally, and female sexual relations with lower castes jeopardize caste status far more than the corresponding acts by male members of the caste.) These food taboos and periods of abstinence evoke an emotional solidarity that binds participants together and separates them from those who do not observe the taboos. Durkheim (1912) theorized that rituals and taboos create social groups and the distinctions among them. Douglas (1966) developed and elaborated this thesis: "When rituals express anxiety about the body's orifices the sociological counterpart of this anxiety is a care to protect the political and cultural unity of a minority group" (p. 124). "Rituals enact the form of social relations and in giving these relations visible expression they enable people to know their own society. The rituals work upon the body politic through the symbolic medium of the physical body" (p. 128). Douglas theorizes that concerns about bodily purity and the horror of pollution delineate social norms and social distinctions—and in fact are essential for the creation of social order. She writes of "a kind of sex pollution which expresses a desire to keep the body (physical and social) intact. Its rules are phrased to control entrances and exits. Another kind of sex pollution arises from the desire to keep straight the internal lines of the social system. ... Individual contacts which destroy these lines [include] adulteries, incests, and so forth" (p. 140). Douglas theorizes that "the body ... provides a basic scheme for all symbolism" (pp. 163-165). But her own examples and a wider review suggest that body boundaries are apt to be indexical signs for one particular kind of sociality, CS. It is, specifically, the separation or intimacy of persons that is coded and constituted through their avoidance, acceptance, or active desire for contact with the substances coming from other persons' bodies (Goffman, 1971, pp. 46-58; Rozin, Nemeroff, Wane, & Sherrod, 1989).

82

FISKE

Hence commensalism is a two-edged sword, excluding as it includes. Those who are not permitted to partake are set apart, outside the substantial boundary. Likewise, refusal to join in a meal or a drink separates. Furthermore, tabooing a comestible substance consumed by others precludes commensalism, thereby marking an asymmetrical social difference in which those consuming the substance that others reject are inferior to those who reject it. Often there are concentric circles in which the inner circles are marked by more extensive rejections (Dumont, 1980; Whitehead, 2000). In this way, successive limits to CS commensalism can create the linearly ordered AR of a caste system. Whitehead (2000) analyzes the many food taboos of the Seltamin of Papua New Guinea, arguing that people who live together are quite uncomfortable when they are unable to share cooked meat. In our terms, because commensalism is a core and powerful conformation of CS, leaving people out is extremely awkward, and being left out is extremely alienating and aggravating. Seltamin seem to have resolved this dilemma with taboos that effectively restrict consumption of each species to an appropriate-sized subset of community members: Everyone for whom the animal is not taboo may eat it together. Very small species that Seltamin hunters are likely to bring home singly are taboo to many categories, for example, while large animals or ones typically caught in large numbers are taboo to few people, or edible by all. All kin groups constitute themselves in part by observance of incest taboos. That is, a kin group is a group of people who may not have sexual relations with each other. Likewise, in the human psyche, communion with gods is so closely linked to ascetic renunciation of sex, as well as to fasting and food restrictions, that intense desire for union with a god spontaneously generates powerful taboos that develop into constitutive entailments (Banks, 1996; Brown, 1988; Lester, 1995). That is, observing appropriate sex and food taboos becomes a requirement for forming or sustaining the CS relationship with gods. Thus asceticism and celibacy are often core constituents of priesthood and monasticism, creating a community identified with a god. Body Marking: Genital Modification, Scarification, Tattooing, Piercing, Hair, Clothing, and Adornment The essential substance of the body is invisible, but people can make their bodies visibly similar by marking their surfaces. Such markings tend to be less powerfully evocative than conformations involving body substances, but surface markings permit connection with a more extensive category of persons. People typically express and form a common collective identity by making their body surfaces the same through body modification, body marking,

3. CONSTITUTING RELATIONSHIPS

83

dress, hair, and adornment. Most evocatively, people mark their identities on their bodies with circumcision, clitoridectomy, facial and body scarring, piercing of the nose, lips, or ears; tattoos; face-painting, hair arrangements or hair removal. For example, among the Kayapo of Central Brazil, Terence Turner (1980) reports that in certain rituals (based on singing and dancing in unison) Kayapo grandfathers and uncles pass on particular styles of feather or shell ornaments to their grandchildren and nephews, who wear them as emblems of their relationship with the bestower. Turner also describes the corporate identity marked by uniform patterns of body painting, including the distinctive patterns all the men in the men's house of a village paint on their bodies when they gather for rituals or other collective action, or all the women in one of the women's societies paint on themselves. Body paint, hair and ear plug styles, initiated men's penis sheaths, and men's lip plug type define membership in each of the fundamental categories of Kayapo society. Furthermore, "variations in coiffure have become the principal visible means of distinguishing one tribe from another" all across Central Brazil (p. 116). Body marking often defines much more widely inclusive forms of consubstantiation than those created by commensalism. Furthermore, people often treat specific pre-existing qualities of skin, hair, or faces as constitutors of the most extensive CS groupings ("races"). As van Gennep (1908, p. 74) observed, genital and other body mutilations are procedures of [permanent] collective differentiation, while clothing and dress provide mutable differentiations. Military uniforms, monks' robes, adolescent fashions, class rings and wedding bands are signs of equivalence groups. All of these mechanisms join people through the identity of a feature on the body surface. Body Contact A strong desire to touch the beloved person is commonly felt; and love is expressed by this means more plainly than by any other. (Darwin, 1872, p. 213)

People also constitute CS with body contact, such as when a caretaker carries a baby or when people sleep together. Hugging, caressing, shaking hands, kissing, nuzzling or touching the other's body with one's nose, grooming, sitting in a lap, carrying a baby, holding hands, walking arm-in-arm or arms around shoulders, and even shaking hands potentially mark and make CS (Darwin, 1872, pp. 213-214; Goffman, 1971, pp. 226-237, 1976, pp. 54-56, 75-81; Morris, 1977, pp. 92-101; Montagu, 1979; Mehrabian, 1981, p. 23; EiblEibesfeldt, 1989, pp. 138-139, 209-234). So does sleeping alongside (Shweder, Jensen, & Goldstein, 1995). Carrying a lock of a person's hair may be a way of staying connected, like the practice in some Melanesian societies of wearing

84

F1SKE

the deceased husband's bones. Wearing another person's clothing or jewelry also connects people (Rozin & Nemeroff, 1994). As Hartland (1895, p. 332) wrote, the desire for connection with those we loved when they were alive leads us to "cherish the relics left to us of their bodily presence and think of the departed as yet about us while we hold these treasures," and to hope that ultimately "the dust which has been dearest will be that which mingles with our own." Similarly, some Christians sustain links with Christ or saints by possessing a fragment—no matter how small—of their body or some material that has been in contact with their body (Brown, 1981). Imitation, Synchronized Rhythmic Motion, and Ritual Repetition A society is a group of people who exhibit many resemblances among themselves produced by imitation or by counter-imitation. (Tarde, 1900, p. xii; emphasis in original)

Three of the founders of social psychology, Tarde (1900/1890), Baldwin (1897), and McDougal (1908), regarded imitation as the source of society, socialization, and the formation of the self (see Fiske, 2004b). Baldwin (1897, 1911) developed his theory of imitation into a sophisticated theory of cultural transmission ("social heredity"), the constitution of social norms, and what we now call agency. More recently Carol Eckerman has shown that imitation is, indeed, the principal medium for American toddlers' social coordination and is the matrix in which verbal conversation takes form (among many studies, see Eckerman, Davis, & Didow, 1989; Eckerman & Didow, 1996). People act like those with whom they identify, reproducing their actions as a way of becoming like them. Identity is expressed and enhanced when an individual reproduces the earlier actions of another. A stronger identification results when the imitation is mutual, synchronous, and rhythmic. Using Australian ethnographic reports Durkheim (1912) concluded that the synchronization of action in collective ritual generates the cohesion which is the primary basis for social solidarity—for the emotional sense of the unity of the community and the common identity of the society. Durkheim theorized that collective rites reflect an awareness of interdependence, the transcendent necessity for social cooperation, and the emergent power of social groups. At the same time, collective rituals elicit and enhance the motives that make social integration possible, creating sentiments of solidarity and moral commitments that enable people to coordinate their personal perceptions and transcend their individual self-interest. Participating in these rituals, people communicate and greatly enhance their social sentiments, uniting and renewing primary groups (see Connerton, 1989, pp. 68-72 for a recent, explicitly psychological restatement of these ideas).

3. CONSTITUTING RELATIONSHIPS

85

Building on Durkheim (1912) and van Gennep (1908), Victor Turner (1967, 1969, 1973, 1975) illuminated the subjective experiences involved in this ritual process. Based on his extensive fieldwork among the Ndembu of Central Africa, Turner showed that during the liminal stage of rites of passage, participants experience an emotionally intense CS relationship that he called "communitas," in which participants feel a sense of undifferentiated unity and mutual identification, without any sense of the AR that is normally omnipresent in such African societies. These processes operate in an implicit, automatic, procedural manner— not at the semantic level. People performing rituals often cannot articulate their reasons for so acting much beyond their sense that "We are doing what our ancestors did; this is what we do—it makes us, us." Without reference to Durkheim, McNeill (1995) investigated the history of close order military drill, concluding that the synchronized rhythmic movement of military drill effectively bonds recruits who are initially strangers to each other. Each time and every place where military drill was invented, the resulting solidarity was a major determinant of success in warfare. McNeill argues that prolonged movement in unison heightens positive emotions, which "binds the community more firmly together and makes cooperative efforts of every kind easier to carry through" (p. 39). McNeill provides a history of the bonding effects of rhythmic movement in community dances, work gangs, religious rituals, calisthenics, synchronized collective exercise in East Asian businesses, political rallies, and political marches. Movement in unison produces "a blurring of self-awareness and the heightening of fellowfeeling" (p. 8), meeting the human need to define identity and orient loyalties, answering the question "To what, and with whom do we belong?" (p. 156). Coordinated rhythmic movement on the dance floor is important in modern courtship, where it may catalyze initial connections. Furthermore, contemporary raves consist of large crowds dancing energetically and nearly continuously for several hours in extremely close proximity very late at night to extremely loud, repetitive, highly rhythmic electronic music. Raves produce a dramatic CS emotion (Holland, 2001b; Reynolds, 1998; Rushkoff, 2001; Takahashi & Olaveson, 2003). Rave participants report intense, transcendent spiritual experience described as love, unity, connectedness, belonging, togetherness, melding of selves, openness, and dissolution of interpersonal boundaries with each other. This CS sentiment encompasses participants who, before the event, were anonymous strangers. Consistent with the thesis that the conformation of CS is nonsemantic consubstantial assimilation, ravers also report that this "vibe" they experience is indescribable, "a kind of energy or pulse which cannot be expressed or understood in words but as something which can only be physically experienced" (Takahashi & Olaveson, 2003, p. 81). Rave participants often feel

86

FISKE

transformed and many later aspire to sustain and spread the Utopian, tribalistic community connectedness they experience. Rave participants commonly take psychoactive drugs, especially MDMA (Ecstasy: 3,4-methylenedioxymethamphetamine; see below). MDMA itself is a powerful source of CS experience (Bravo, 2001; Holland, 2001a, pp. 59-61). But while MDMA substantially enhances the euphoric CS limerance of raving, MDMA is apparently not entirely necessary to it: Raving without MDMA can also produce a powerful communitas through synchronous rhythmic collective movement. Music itself, especially percussion, has a synergistic role in kindling CS at raves, as it does in other rituals, particularly in conjunction with dance. Initiation: Shared Fear, Privation, Pain, and Body-Modification There is more to ritual than synchronous rhythmic movement. Indeed, initiation rites and other life-cycle transitions create new identities and bonds through combinations of all these forms of consubstantial assimilation. In such rites around the world, dancing (especially to percussion) is very common (Dulaney & Fiske, 1994; Fiske & Haslam, 1997). Concerns about contact with or ingestion of polluting substances, especially body wastes or secretions, are also very widespread in all rituals. Food and other taboos are often quite salient in initiation rituals, and most rituals involve ceremonial commensal eating, especially meat. Violence is common in initiation rites, including beating and cutting the bodies of the initiates (e.g., Herdt, 1982; see Girard, 1977 on violence in ritual). Shaving or dressing of hair, applying pigments or other substances to the body, and special clothing or adornments are very common components of initiation. Many initiation rites revolve around modifications of the body, especially genital operations but also other kinds of scarring.9 For example, all of the above features occur repeatedly in the seven-stage series of initiation rites that Barth (1975) describes in central Papua New Guinea. Elimination of polluting "female" substance by bleeding, extrusion of pus, and vomiting are core activities in the initiation rites of many New Guinea cultures (Herdt, 1982). In several New Guinea cultures, initiation also involves fellatio, because boys need to repeatedly ingest semen in order to complete their development or acquire the supply of semen they will need as adults. "Initiated together, emotional and social bonding may occur that lasts for a lifetime and that involves social cooperation and respect" (Ottenberg, 1994, p. 354). Indeed, when Young (1965) coded ethnographies from 54 cul9

In a number of societies, initiates form life-long ritual bonds with the person who holds them during circumcision.

3. CONSTITUTING RELATIONSHIPS

87

tures, he found that male solidarity is greater in societies with more elaborate initiation rites, and female solidarity is greater in societies with more elaborate female initiations. Other researchers (Burton & Whiting, 1961; Herdt, 1982; Whiting, Kluckhohn, & Anthony, 1958) have shown that male initiation typically breaks down the boy's identification and intimacy with his mother, substituting male identity and solidarity with peers and adult males. Male initiations are more prevalent and often more intense than female initiation rites. They can be extraordinarily severe in cultures where warfare or raiding is frequent and men must be courageously loyal to each other (e.g., Herdt, 1982). In many pastoralist societies of East Africa (where warfare and raiding was constant), men form their strongest CS bonds with their age mates who were initiated during the same span of years; age mates share food and even wives, pay each other's fines, and risk their lives for each other (Baxter & Almagor, 1978; Spencer, 1985). Initiation rites in tribal societies are not the only instances where strong CS bonds result from fearful threats to the body, shared pain, and corporal privation. Boot camp and combat forge intense links, sometimes even to the point where men will sacrifice their lives for each other. Even hostages whose lives are at stake sometimes come to identify with their captors (the "Stockholm syndrome"). Suffering from the same disease or affliction is a potential connector as well. Other rituals, too, are typically composed of these forms of consubstantial assimilation (Dulaney & Fiske, 1994; Fiske & Haslam, 1997). The additive—or perhaps multiplicative—psychosocial impact of these conjoined components explains Durkheim's (1912) observation that rituals represent and reinforce social solidarity. I have emphasized the consistency and prevalence of consubstantial assimilation across cultures. But there are innumerable intriguing cultural differences in the ways that people constitute CS and, of course, people can form CS relationships with reference to anything they have in common, not just their bodies (Tajfel, 1978). However, the pattern evident across cultures is that people are strongly disposed to conforming CS by making their bodies the same in some way. There is great cultural diversity in how this is done and each culture has important unique features. For example, milk kinship and bond friendship through mutual ingestion of blood are not culturally defined and elaborated in most societies. But they are intuitive and potentially emotive to most people, which leads to repeated independent invention of analogous concepts. And this constitutive affordance results in general susceptibility to cultural reproduction and diffusion of these particular conformations (when they are compatible with other local practices, norms, and institutions). This CS conformation is often inaccessible to conscious reflection or articulate expression. In contrast to organizations based primarily on AR, EM,

88

FISKE

or especially MP, CS organizations articulate few explicit rules, often eschewing written codes or manuals (Henry, 1988; Rothschild-Whitt, 1979). This reflects a sense that abstract language does not readily capture the fundamental essence of the CS relationship. (Imagine falling in love according to the terms of a contract.) It is characteristic of CS relationships that people often feel that such relationships simply are, without being able to articulate precisely why they are as they are. However, when CS, AR, or EM relations are contested or otherwise problematic, a discourse usually develops that translates the ineffable primary conformation into articulate ideology. Then people develop declarative accounts of them, formulating complex narratives and propositions about these relationships. But these semantic ideological representations are derivative of the fundamental conformations: CS, AR, and EM relationships do not require conceptual formulation or explicit declarative expression. People simply "do" Communal relationships, especially without having to analyze them abstractly. Sources of Consubstantial Assimilation

Formed by natural selection, consubstantial assimilation is an adaptation whose function is to coordinate CS relationships. The human genome, neurophysiology, neuroanatomy, ontogeny, and life history mediate the constitutive effects of consubstantial assimilation on CS. Conversely, the genome, neurophysiology, neuroanatomy, ontogeny, and life history mediate the expression of CS through consubstantial assimilation. More simply, consubstantial assimilation is a biological cause and consequence of CS, integrating participants' psychologies so they jointly create mutually adaptive social relationships. Close kin share genes and hence inclusive fitness, so animals have a variety of mechanisms for kin recognition (Hepper, 1991), for pair bonding and maternal-infant imprinting, and for parental care. The details of the neurochemistry of human consubstantial assimilation are not known yet, but some aspects of possibly homologous processes have been extensively investigated. Much is known about mammalian maternal behavior (particularly in rats), the bonding of ewes to their lambs, and pair bonding in monogamous species of voles. These mechanisms are relevant since such neurochemical processes are likely to be phylogenetically conserved and subsequently elaborated into more complex social adaptations. Oxytocin appears to play an especially pivotal role in maternal behavior and also in female pair bonding in "monogamous" species of mammals. A closely related neuropeptide, arginine vasopressin, appears to mediate pair bonding and perhaps paternal care in males (Carter, DeVries, Taymans, Roberts, Williams, & Getz, 1999; de Bono, 2003; Insel & Young, 2001; Kendrick, 2000; Uvnas-Moberg, 1998; Young, Lim, Gingrich, & Insel, 2001). The behavioral ef-

3. CONSTITUTING RELATIONSHIPS

89

fects of vasopressin evidently depend on the distribution in the brain of its Vla receptors. This distribution varies greatly across closely related species that exhibit different degrees and patterns of affiliative sociality (Insel, Young, & Wang, 1999). Conversely, species with similar patterns of affiliation have similar distributions of Vla receptors in the brain. A gene has been identified that affects the distribution of Vla receptors. Most species of mice are not sociable, and mice normally do not respond affiliatively to cerebral vasopressin. When the Vla distribution gene is taken from an affiliative species of vole and introduced into the mouse genome, the transgenic mice respond affiliatively to vasopressin (Young, Nilsen, Waymire, MacGregor, & Insel, 1999). However, the endogenous release of these pituitary peptides during gestation and parturition is not sufficient for the subsequent maintenance of mammalian maternal behavior, which requires continued close contact with offspring (Rosenblatt, 1994). In human mothers, delivering a baby, skincontact with the baby, the infant's massaging of the breast, and nursing all release oxytocin from the posterior pituitary, mediated by the paraventricular nucleus of the hypothalamus (Matthiesen, Ransjo-Arvidson, Nissen, Uvnas-Moberg, 2001; Panksepp, 1998, pp. 246-260). It is plausible to imagine that this oxytocin release, and perhaps social bonding more generally, may be mediated partly by a specific kind of slow-conducting unmyelinated nerve that responds to slow stroking (Olausson et al., 2002). It is also notable that the intense CS emotions produced by MDMA (Ecstasy) involve strong desires for mutual physical contact (e.g., back rubs, hugging, hand-holding; Beck & Rosenbaum, 1986; Klam, 2001). The synergy in raves between MDMA and prolonged dancing to very rhythmic music is also interesting. "The community of ravers loses its sense of self in favor of a larger, stronger, and perhaps healthier entity. . . . Ravers are bonded to one another as members of a single tribe dancing around the fire" (Holland, 2001b, pp. 346-347). MDMA clearly activates a very specific neural mechanism for CS bonding. People report that when they take MDMA, they feel extremely close to everyone around them; the feeling is extremely sensuous, but usually not erotic (Reynolds, 1998; Bravo, 2001; Holland, 2001a, 2001b; Klam, 2001; see also the MAPS literature reviews at http://www.maps.org/research/mdma/ litupdates/). In one of the very few systematic quasi-clinical trials, all 27 participants who took MDMA in the presence of others "felt closer and more intimate with anyone present" (Greer & Tolbert, 1986, p. 320). People on MDMA experience "a hunger for human connection and a desire to empathize" (Klam, 2001, p. 40). The CS effect occurs in diverse settings, with or without prior expectations. MDMA makes people feel this way about everyone: partners, friends, casual acquaintances, strangers, even rival gangs (Beck & Rosenbaum, 1986; Reynolds, 1998; Klam, 2001). An appreciable

90

FISKE

sense of communion often lasts for days or weeks after taking MDMA (Greer & Tolbert, 1986). It is not known how or where in the brain MDMA produces CS, but it would enhance our understanding of both entities to have further research on the neurochemistry of MDMA and the mechanisms by which its psychopharmacology converges and resonates with the constitutive effects of consubstantial assimilation.10 Effects of MDMA11

Love (agape, not eros}, intimacy, tenderness, deep sharing, warm friendship. Hunger for human connection, urge-to-merge. Connectedness, oneness, cohesion, community, merging with collective, group mind, collective consciousness, oneness with the universe. Unconditional acceptance, openness, kindness, empathy, understanding, tolerance, conciliation, niceness to everyone. Greater ease in relating across boundaries, respect, improved relationships. Enhanced, more open communication, greater self-revelation. Trust, loss of defensiveness, breakdown of social inhibitions, forgiveness, overcoming conflicts, pity for others' suffering. Desire to touch others and to dance collectively. Euphoria, exhilaration, joy, self-acceptance, reconnection with oneself, self-love, comfort, peace, tranquility. Sources: Greer & Tolbert, 1986; Beck & Rosenbaum, 1994; Reynolds, 1998; Bravo, 2001; Holland, 2001a, 2001b; Rushkoff, 2001; Klam, 2001, Olaveson, in press. Clearly the conformation of a social relationship is integrated into a much more extensive system of genes, experience, epigenesis, neural architecture, and hormones. Endogenous proclivities, pervasive commonalities in human developmental experience, and socially transmitted practices 10

MDMA is a Schedule 1 drug in the United States and illegal in most other countries as well; the 1912 Merck patent has long expired, making it unprofitable for any pharmaceutical manufacturer to conduct clinical trials. "People develop tolerance to MDMA: With repeated doses, the effects gradually diminish and often eventually disappear; also, the effects of a dose are greatly reduced if it is ingested within a few days of the previous dose. MDMA also has a number of typical side effects, including initial nausea, loss of appetite, hyperthermia, teeth grinding and jaw clenching, amphetamine-like hyperactivity, and subsequent "hangover"; severe side effects occur but apparently are rare (Henry & Rella, 2001).

3. CONSTITUTING RELATIONSHIPS

91

generally resonate, amplifying or modulating each other. These other components of human sociality feed in and out of the conformation subsystem: Some activate or orient the conformation, some mediate its operation, and some are mediated by the conformation system. Functionally, they are components—or aspects—of the human social relationship systems. Hence one way to explore conformation theory is to attempt to identify the set of genes, which, in normal epigenesis, are expressed in neural architecture linked to neurophysiology in specific ways. In appropriate developmental, social, cultural, and physiological conditions, conformation theory predicts that actions such as sharing food, being caressed, or undergoing an initiation activate sensory and central nervous processes that result in the release of peptides such as oxytocin or vasopressin, along with cortisol, resulting in activation of further cognitive processes and motivational mechanisms also mediated by these and other neuropeptides, gonadal steroids, other hormones, dopamines, and opioids. This process enhances emotions and motives that create and sustain CS relationships. In many mammals (and birds), close contact with mother, siblings, or father is crucial for keeping infants warm, and may be nearly continuous for some days or weeks after birth. Many (perhaps all?) mammalian mothers lick newborns (in some cases, thereby stimulating the onset of respiration) and some mothers also clean older infants with their tongues. Some mammals, such as dogs, continue to lick and rub their bodies against adults to express affection (Darwin, 1872, pp. 118-119, 124). Nearly all mammalian infants are in close body contact with their mothers, siblings, or father much of the time, which is often crucial for keeping infants warm.12 Primate infants cling to their mother's body and are in direct contact with her continuously for the first few months; as they grow, infants of some species may ride on the mother, and all move off her body more and more but return to the mother for safety. This body contact appears to be crucial to both sides of the infant-mother bond (see Harlow, 1958; Harlow, Harlow, & Hansen, 1963). Many primates groom each other frequently, apparently creating and maintaining bonds by touching each other, going through the fur to find— and eat—ectoparasites (Dugatkin, 1997, pp. 116-124; Dunbar, 1996). In many primate species, juveniles and adults, particularly females, are very attracted to infants, seeking to approach, touch, and carry them. To reconcile after a conflict or to prevent potential aggression, chimpanzees "hug" and "kiss" (de Waal, 2001). In addition to nursing and grooming, other uses of body orifices to create and sustain close relationships are also important in some primates. One of the closest relatives to humans, the bonobo, very frequently uses genital contact and other forms of sex across and within gender to redress 12 Young mammals often have a distinct pelage color or pattern which may also affect parental and other adult behavior toward them.

92

FISKE

and perhaps to reinforce relationships; capuchins also use sex in this constitutive manner (de Waal, 2001; Hohmann & Fruth, 2000; Manson, Perry, & Parish, 1997). Capuchin monkeys have the biggest brain-body ratio of any nonhuman primate, although they are new world monkeys not closely related to the great apes and humans. Capuchins often invent and learn from each other a variety of practices that seem to reinforce trusting relationships (Perry et al., 2003). Capuchins are almost frenetically active, yet typical socially transmitted capuchin practices include: sitting placidly with fingers in each other's nostrils; placing tail or fingers in the mouth of another capuchin, who holds on firmly while the first animal tugs and, if able to withdraw the appendage, replaces it in the animal's mouth again; or biting out tufts of the other's hair and passing the tuft back and forth until it is depleted, whereupon one capuchin starts over by biting out another tuft to pass on. This primate licking, nursing, clinging and holding, grooming, adjacent sleeping, sociable sexual contact, and body orifice or hair behaviors are precursors that set the phylogenetic stage for human proclivities for consubstantial assimilation. In most human societies, infants and young children are carried on their mother's or sibling's body most of the day; the carried infant necessarily moves in synchrony with the caretaker's body. It is a common sight in many societies to see a mother walking, digging, hoeing, grinding, pounding, stirring, or dancing, with her infant or toddler tight against her body, joined in her rhythms. Infants in almost all cultures sleep next to their mother at night; older children sleep next to a parent or other family members (Whiting, 1961). In addition to nursing, humans (but not other primates) actively feed their children, and young children typically eat commensally with their mother and siblings, at least. To the extent that maternal-child relations are the most intense instantiation of CS and early relations with mother and siblings are the ontogenetic prototype for later CS relationships, ingestion of body substances, feeding, commensalism, and close body contact are thus natural constitutors of CS. Human children presumably generalize from nursing, feeding, nurturant body contact, and synchronous rhythmic motion to later forms of feeding, food sharing, body contact, and movement in unison. Most mammals do not share food, although some canids bring food back to their den and regurgitate for offspring, mates, and siblings. Chimpanzees, bonobos, and capuchins beg food successfully from each other, especially meat; actively initiated giving of meat occurs, but is infrequent (de Waal, 1996, pp. 136-150, 244). However, in virtually all contemporary human hunter-gatherer societies, hunters who kill any sizable animal share the meat with the entire band; indeed, in most such societies the hunter turns it over to someone else to divide, and may himself get a very modest share (Hawkes, O'Connell, & Blurton Jones, 2001; Testart, 1987).

3. CONSTITUTING RELATIONSHIPS

93

Recent research has identified crucial components of the neural substrates of imitation and its relation to empathy and identification. At least some primates have neural systems which represent goal-directed action; moreover, certain sets of "mirror neurons" fire both when the monkey perceives a particular action (visually or auditorily) and when the monkey performs that particular action, even if it acts in complete darkness and cannot see its own actions (Gallese, 2003; Rizzolatti, Fogassi, & Gallese, 2001). Likewise, the same regions in the human superior temporal, posterior parietal, and inferior frontal cortices are activated both when the subject performs a specific action and when the subject observes that same action (Carr, lacoboni, Dubeau, Mazziotta, & Lenzi, 2003; Heiser, lacoboni, Maeda, Marcus, & Maziotta, 2003). This shows that monkeys and humans have agentindependent neural representations of actions. This is likely to be the basis for the extraordinary human capacity to imitate—a capacity that no other animal (including other apes) possesses to anywhere near the same degree. In humans, at least, the capacity to imitate appears to be innate, since newborns are able to imitate a variety of tongue protrusions, despite being unable to observe their own tongues (Meltzoff & Decety, 2003). Human infants are also aware when they are being imitated. Gallese (2003), Heiser et al. (2003), and Meltzoff and Decety (2003) plausibly infer that these agentindependent action representations and innate human capacities for imitating and recognizing imitation provide a sense that certain others are "like me," and hence are the bases for empathy and identification. It is plausible to imagine that this neural imitation-identification-empathy system partially mediates the constitutive effects of repetitive and synchronous action in ritual. Imitating the actions of predecessors and peers may enhance participants' sense of social identity because of their common representation of their own and others' actions: Ritual participants perceive their bodies as kinesthetically equivalent. Conversely, well before age two, toddlers around the world selectively and persistently imitate the social roles of people with whom they identify. Indeed, they use their intrinsically motivated imitative capacities as the principal means to develop the competence to participate in culturally coordinated social relations of each community (Fiske, 2004b). All this is simply to say that the social processes involved in consubstantial assimilation (and other conformation systems) operate via neural systems that evolved out of mammalian and primate antecedents. This phylogeny must have potentiated and constrained the biology of human sociality, a biology which in turn informs the interpersonal and cultural processes. Understanding this biology is crucial for understanding the sociocultural processes—but understanding the sociocultural processes is also crucial for understanding what it is the biological systems are doing, why they evolved as they did, and why they function as they do. The phylogeny,

94

FISKE

genetics, epigenetic processes, and neurobiology shape and are shaped by sociocultural processes. We need to understand how they dynamically inform each other. Researchers who can imagine that the mind is a blank slate might suppose that in the absence of innate proclivities, universal aspects of child experiences would be sufficient to foster adult dispositions to constitute and cognize CS through food sharing, body contact, and rhythmic synchrony. This supposition can be tested by examining cultures where many children are not nursed, are fed separately from adults or fed distinct foods from separate dishes, sleep alone, and are infrequently carried on the body. Until recently, this describes middle-class North American culture, where consubstantial assimilation nonetheless has always been the salient conformation system for CS. So it is hard to doubt that there are endogenous proclivities for consubstantial assimilation. Hence this conformation system must have some source in human evolution, and indeed been bootstrapped somehow from the kind of human life history experiences that are typical of most cultures and presumably were typical of foraging societies during most of human prehistory. Baldwin (1896; see Fiske, 2000) theorized just such a bootstrap mechanism for genetic assimilation of adaptive learning. He pointed out that when it is consistently beneficial for an organism to learn something, natural selection will make that learning more rapid and reliable. If certain cues consistently index a potential benefit, the organism will evolve so that the adaptive actions become automatic responses to the cues. Thus Baldwinian selection produces innate proclivities for responses that are typically adaptive. This is the evolutionary mechanism that, building on mammalian precursors, probably molded consubstantial assimilation into the conformation system for CS.

CONSTITUTING AUTHORITY RANKING Every political system incorporates elements of hierarchy, i.e., relations of 'superiority-inferiority'. The very words we use in English indicate how deeply ingrained is the idea that this kind of social relationship can be represented in 'the logic of the concrete' by differences of relative level, above-below. This idea certainly comes fairly close to being a human universal. The political leader is 'above' and 'in front of those whom he 'leads'. (Leach, 1972, p. 335)

People generally constitute hierarchical relations in the medium of a social physics in which space, mass, time, and force function iconically to represent and create rank. Compared to persons of inferior rank, people represent superiors as

3. CONSTITUTING RELATIONSHIPS

95

above; in front; earlier in time; larger; more numerous; and more powerful (exerting more force, stronger). In short, the conformation of AR maps orderings in space, time, magnitude, and force onto social hierarchies, and vice versa. People tend to think, constitute, and communicate AR with this social physics, and arrive in the world expecting to find and use it. Each culture uses varying and particular space, time, magnitude, and force metaphors for rank, but these local implementations are universally intuitive and usually immediately apprehended by children and newcomers, including ethnographers. In effect, the ordinalities of space, time, and magnitude are models for the social ordering of humans. Humans naturally use metaphors mapping one configuration onto another, and not just in relation to social hierarchies, of course (Lakoff & Johnson, 1980). Indeed, the founders of anthropology collected evidence from around the world to show that all humans not only think with metaphors but act metaphorically. That is, they manipulate one set of elements to effect homologous transformations in another set. Tylor (1871) called this manner of action similarity or resemblance; Frazer (1890) developed Tylor's ideas, variously characterizing this kind of action as sympathetic magic, imitation, or homeopathy. Frazer and Tylor described myriad practices in which people manipulate one configuration to effect similar transformations in a homologous configuration. Their discovery is illustrated by the way Moose, when they are striving to gain high rank, ingest potions compounded to contain the essence of "highness," using combinations of ingredients such as rocks from the tallest mountain peaks and epiphytes growing atop the highest trees. That is, they use superiority in elevation of objects to bring about social superiority. This metaphoric principle of similarity or imitation is involved in using physics to think about, represent, constitute, communicate about, recognize, and transmit cultural implementations of AR. In a Peircean semiotic analysis, the physics of AR is predominantly iconic. An icon is similar in pattern to its referent—the arrangement of its components relative to each other resembles the configuration of its referent. Drawings, schematic diagrams, and maps are common nonsocial icons. People are operating iconically when they use their spatial or temporal positions to confer and signal positions in a social ranking. Iconicity is the semiotic form when beings with differing social weight are "greater" or "inferior," and get offices, abodes, or temples of corresponding magnitude. Elders or persons with

96

FISKE

seniority may be accorded prestige, and people may show deference by giving others temporal precedence. A monarch assuming office "ascends" the throne. People who have social control over others are conceived and spoken of as "powerful." An effective leader is "strong," regardless of how much she can bench-press. A recognized traditional leader is one who is said to be the "first man" (mae ulu). A powerful funei is not only "big" (bi o), he is figuratively "on top" (au faklignd) or "high" (au fahaghe). Accordingly, followers of such a person show respect by "looking up" (filo fahaghe), as in the English idiom.... The cultural importance of the theme of dominance is reflected in the frequent use of these spatial metaphors of size (BIG/SMALL) and position (ABOVE/BELOW, IN FRONT/BEHIND) to express the relative positioning of social actors. (White, 1985, p. 345)

What White says of the Solomon Islands is true in most other cultures, but has been most extensively studied in Pacific island cultures. One of the first clear references to this is by Malinowski (1935), describing the Trobriand Islands: "Rank . . . entails a definite ceremonial, the main principle of which is that elevation must be commensurate to rank" (p. 34). In a classic paper, Firth (1970) describes the use of space to constitute AR in Tikopia. At public events, being seated in front of someone indicates higher rank: "The general Tikopia proposition is, 'My front to the rear of a person indicates he is of higher status than I. My rear to his front indicates that I am of higher status than he'" (p. 195). Relative elevation is even more important: Standing indicates superiority over those seated. Sitting up conveys superiority over those crouching, squatting, kneeling, or crawling at a lower level. On ceremonial occasions, only chiefs sit on stools, raising them above those sitting on the ground. The gods are thought to be above humans and vertically differentiated among themselves. This comes out in the Tikopian lexicon: The term man runga indicates higher in either physical or social position. Certain Tikopian gestures are also based on elevation. Firth (1970) describes how a chief's bowing his head or raising an offering cup indicate respect to the gods. He also describes how CS and AR are constituted simultaneously by touching the nose to another person, often while gently inhaling (sniffing). This nasal contact establishes an emotional CS connection, while the vertical point of contact on the other person's body indicates relative status. Equals touch nose to nose or cheek; subordinates touch their nose to the other's wrist, knee, or (very rarely) foot, according to the degree of differentiation and the formality of the occasion. These vertical points of contact also are used to indicate the abjectness of an apology. This elevation system operates again in the recipient's response to being sniffed, so that the person who receives a nose-to-knee, for example, usually responds

3. CONSTITUTING RELATIONSHIPS

97

by raising the other person's head and touching nose to nose, "making the face good." Likewise, Garvin and Riesenberg (1952) and Keating (2000) report that relative rank is constituted and communicated on ritual occasions in Pohnpei (Micronesia) by several uses of vertical differentiation: entering directly onto, and sitting on, the raised platform in the feast house (rather than entering at ground level and sitting on the ground); sitting up on a chair (rather than on the ground); standing above others who are seated; having one's head above others (who keep below, unless they are serving or greeting); having others walk on their knees in one's presence; receiving bows from others; and the position of the deities on a second, higher platform. Pohnpeians also use a morpheme meaning "heaven/sky" as a royal honorific. In addition, like Tikopians, Pohnpeians rely heavily on horizontal spacing to constitute rank: Higher status persons sit nearer to the end of the feast house where the deities are located. Temporally, rank is conveyed in Pohnpei by being served food and ceremonial drink earlier and (in modern commercial bars) being entitled to start drinking first. Magnitude is also important: Occupying a larger feast house and receiving a larger share of food produce and result from higher status. Higher ranking persons also receive choicer portions (containing more fish and chicken). A similar system of elevation, temporal precedence, and magnitude operates in Fiji, where people build their houses on earthen platform foundations whose relative height corresponds to the owner's rank (Toren, 1990). Fijians seat themselves in rank order toward the front or back; they refer to these positions as "above" and "below." When they walk past others who are seated, they stoop and lower their heads to different degrees according to their relative rank. Fijians also mark rank temporally: They begin eating and they serve their traditional social drink in order of rank. Senior men get the biggest and best portions, leaving some for those seated "below" to finish off. Those who are older thereby rank higher than their juniors, and are entitled to command and physically discipline them. At the same time, of course, "eating together marks out the household, and is at the same time a defining feature of kinship in the widest sense" (Toren, 1990, p. 58; see also p. 168). The physics of rank construction is not limited to Oceania, of course. As Firth (1970) points out, postures and platforms differentiate hierarchical positions in China, Britain, and Roman Catholic practice. As Spencer (1896, pp. 116-143), Leach (1972), Goffman (1976, pp. 40-41), and Morris (1977, pp. 142-147) note, lowering the body by bowing or prostrating shows deference/subordination/submissiveness in most cultures (while, as Goffman observes, superiority is marked by standing erect with the head held high). My own brief exploration shows that vertical postural marking of rank is at-

98

FISKE

tested, for example, in ancient Egypt and Hawaii; premodern Europe, and contemporary formal European court etiquette; Moose, Hausa, and Yoruba in West Africa; and China, Korea, and Japan. In his survey of 212 sources on postures in 480 cultures around the world, Hewes (1955) reports that the use of stools is often restricted to persons of high rank or authority, and in many cases, to men. Schwartz (1981) reviews observations of several sociologists commenting on the prevalence of vertical metaphors in American and other cultures' cognizing of rank. Schwartz further discusses the meaning of elevation in American architecture, where offices on higher floors or the elevation of the judge's seat in the courtroom confer high status relative to those below. Mayans in Chiapas (Mexico) use up, ahead, and earlier to rank men in their religious hierarchy (Rosaldo, 1968). Mayan men avidly seek ranked offices in weekly rituals honoring a local saint. Rosaldo characterizes these rituals as undramatic, highly redundant performances of ranking. Hierarchy is repeatedly displayed and created when lower ranking officeholders bow to men who occupy senior offices, men distribute drinks in order of rank, and they process from place to place in inverted order of seniority. Seating at table is also linearly ordered by the rank of the man's office. All officeholders bow to the altar, acknowledging their position below the saint and God. In everyday life, also, these Zinacantans mark rank by serving drinks to men in order of age, by younger men bowing to older men, and by walking along the narrow paths in age order. It is notable that the Mayans of Chiapas can provide no exegesis or verbal rationalization of their system of constituting rank, nor do they have any mythology for it: It is simply "custom." That is, there is no formulated semantic representation of these iconic practices. Darwin (1872, p. 263) briefly suggested that people convey superiority by erect posture, puffing up to appear large, and gazing disdainfully downward on inferiors. In fact, there is direct experimental evidence that posture and spatial position communicate hierarchical position. Spiegel and Machotka (1974) asked American undergraduates to rate attributes of men depicted in line drawings. One figure was depicted standing on a surface that appeared approximately 18 inches higher than a lower figure. Compared to the lower figure, the higher figure was rated higher on a scale going from superordinate to subordinate, higher on a scale going from important to insignificant, and higher on a scale going from haughty to humble. If the figure was gazing out straight ahead over the other figure or directly at the other, these superiority ratings increased, while the lower figure was rated lower on these dimensions if he was also looking down. In another set of male figures, one with his left arm raised in gesticulation was rated the most superordinate and important. The same interpretations of spatial positions appeared when Schwartz, Tesser, and Powell (1982) asked a sample of un-

3. CONSTITUTING RELATIONSHIPS

99

dergraduates to designate the "socially dominant" person in 64 drawings of dyads. Subjects were strongly biased to attribute dominance to the elevated member of the dyad, and also showed smaller biases to attribute dominance to the person in front, and to the person standing next to a seated person. As Spiegel and Machotka (1974) point out, in the movie The Great Dictator, Charlie Chaplin caricatured Hitler's and Mussolini's concern about their short stature and their relative superiority by representing them in a barber shop, successively raising their chairs to stay above the other until they reach the ceiling. With a throne rather than a barber chair, in Paris in 1537, Francis I used his elevation to display and establish his superiority over Parliament, and to indicate the relative rank of other participants: He sat on a dais at the highest level, surrounded by other nobles and officials arrayed on steps or seats at levels indicative of their rank (Hanley, 1985, pp. 79-80). Goffman (1976) interpreted magazine advertisements to show that relative elevation in space indicates high social position (pp. 43-45), and "social weight—power, authority, rank, office, renown" is also conveyed by size, especially stature (p. 28). Indeed, psychological studies have shown that relative social rank distorts American subjects' perceptions of their own and others' heights: People perceive higher ranking persons to be taller than they really are and judge lower ranking persons to be shorter than they are (Dannenmaier & Thumin, 1964; Higham & Garment, 1992). Moreover, taller people are more likely to be selected for corporate leadership and to be elected president (Hensley & Cooper, 1987; Melamed & Bozionelos, 1992; see also research reviewed in Schwartz, 1981, p. 60). Traditional rulers and nobility around the world often wear robes that greatly expand their apparent size and coiffures or headgear that make them appear taller (e.g., for renaissance European monarchs, see images in Strong, 1973, and depictions of Louis IV in Giesey, 1985, p. 61, including Thackeray's satire showing the small man beneath the grand attire that renders him king). Languages from many different families constitute and communicate AR with words whose primary reference is space and magnitude. Schwartz (1981) reviews incidental sociological observations on the ubiquity of vertical metaphors for rank in English, and interviewed native speakers of 17 languages from around the world. All indicated that their languages tend to encode social power and authority along a vertical dimension. Interviews with language faculty at UCLA and a few other native speakers confirm that terms denoting up and down are commonly used to convey social status in Ancient Egyptian, Korean, Mandarin Chinese, English, French; my own experience confirms this for the Moore language of the Moose. The Japanese also use vertical metaphors for hierarchy (Wetzel, 1993). UCLA interviews and personal knowledge also indicate that words for physical magnitude are commonly used to denote rank in Ancient Egyptian,

100

FISKE

Mandarin Chinese, English, French, Moore, Swahili, Wolof, and Hausa. Indeed in many languages, it requires circumlocutions to refer to AR without using terms such as above and below, high(er) and low(er), bigCger) and small(er). In addition, a great many languages from Indo-European, Native American, Pacific Island, Asian, and various African language families use plurals in address or reference to indicate deference or subordination (e.g., French vous vs. tu, and French and English royal nous and we; Mandarin and Korean nin vs. ni; Moore yamb and tondo; see reviews by Agha, 1994; P. Brown & Levinson, 1979, p. 203; R. Brown & Oilman, 1960; Garvin & Riesenberg, 1952). Furthermore, diverse languages refer to social authority as power or force. These linguistic practices are more than simply lexical metaphors; they are collective cognitive representations of what AR consists of: Authority is being above, greater, and more powerful. Languages mark status in diverse ways, including titles and speech registers, but these metaphors of social physics are the most widespread and often the most basic. Magnitude in the conformation of AR is also exhibited when people of high rank are accorded greater social space than people of lower rank. They have larger abodes, wider ritual zones, and more extensive domains. Executives have offices that consist of architecturally reified personal spaces, while lower functionaries may have cubicles and subordinate workers have no individual space of their own. Higher ranking executives are often entitled to bigger desks, bigger cars, and so forth. Moreover, their "personal" space is greater: They have more social volume. The higher a person's rank, the further away others are kept; subordinates must stop before entering the superior's space. Perhaps the most extreme examples of this are the Imperial Palace in Tokyo and the Forbidden City in Beijing. In general, to gain access to very high ranking persons, one must go through at least one secretary, chamberlain, or assistant. However, spatial approach is usually asymmetrical: A person of lower rank enters the space of a higher ranking person only with permission, while the converse is not usually so important (Mehrabian, 1981, pp. 58-65). For example, to reach the President of the United States an outsider has to gain permission to pass a number of physical and social boundaries, but the President may readily enter the space of subordinates and approach them without asking their permission. Around the world and over millennia, the vast majority of dominant gods are located above humans (Laponce, 1981, pp. 69-92). The gods are often represented as intimidating giants: massive statues of corpulent Buddhas, giant stone heads on Easter Island, enormous statues in ancient Egypt, or the towering masks of West Africa and New Guinea (see, e.g., Morris, 1977, p. 152). And in places of worship such as enormous cathedrals, mosques, and temples, or icons such as the Ka'ba of Mecca, humans are commonly made to feel small in the presence of their god(s). For millennia, rulers have

3. CONSTITUTING RELATIONSHIPS

101

built great monuments to themselves, too, such as the Egyptian pyramids whose mass is an icon for the ruler's "greatness." Memorials have the same meaning: Consider the height of the Washington monument and the massiveness of the Lincoln memorial. Obelisks and arches indicate the grandeur of the rulers who erect them, while larger-than-life sculptures of heroes are often set on high pillars or pedestals. High deities are also represented by elevation, as in towering church steeples, Mayan stellae, or Northwest Indian totem poles. Likewise, altars are often raised above the congregation or set upon massive bases, such as the Mayan pyramids. With the advent of steel girder construction, tall buildings became secular icons of status among corporations, cities, and even nations. The human proclivity to perceive rank in spatial orderings and magnitude is so strong that special efforts may be necessary to preclude unwanted ranking. Corporations that intend to minimize ranking among managers may need to build flat, single-story office buildings—perhaps ringshaped, with executives on the windowed perimeter in same-sized offices with identical desks. Likewise, according to legend, the Round Table was created to avert conflicts over precedence in seating King Arthur's knights. Rank is also constituted by temporal precedence. Subordinate military personnel must salute first and hold the salute until the higher ranking person returns the salute. When naval ships pass, the ship commanded by the junior officer renders honors first. More generally, to receive the first share in a distribution, or to be entitled to speak first, is to rank first. Superiority is also signaled when others have to wait for the arrival of a person before beginning an activity, wait for their departure before leaving, or wait for a person to begin to eat or drink. Naval officers getting into small boats enter in inverse rank order and debark in rank order (so the higher your rank, the less time you wait for others). Similarly, at formal military and diplomatic dinners, people are supposed to arrive in inverse order of rank, so that junior officers wait for their seniors. Their departure is in rank order, such that those who rank below should wait until their superiors leave. Sometimes the initial position is the only one clearly demarcated, but often people follow an ordinal temporal sequence more or less all the way down the hierarchy. On a longer time scale, order of birth often creates rank ordering. In virtually every society where people are ranked, older siblings are superior to younger siblings, and, other attributes aside, seniority in age denotes relative social status (e.g., Toren, 1990). In modern bureaucracies and other contexts, seniority on the job often translates into formal or informal status, as well as pay and promotion. Temporality can be a crucial determinant: Within any given military rank, officers are ranked in order of their date of commission; even a day's difference is sufficient to give the senior officer command.

102

FISKE

To posit that space, time, and magnitude operate as the primary conformation system for AR does not imply that any set of people must occupy immutable positions on these dimensions, of course. People in a given dyad or network may activate many potential relational models; strategic maneuvering and shifting among potential relationships result in varying configurations. People contest their ranking through the conformations analyzed here, so that spatial positioning, order, and magnitudes vary. Moreover, spatial and temporal positions and magnitudes associated with persons only constitute or communicate rank in certain contexts, notably in the highly structured, predictably redundant occasions that we call rituals. It is an interesting question how people recognize the frames that demarcate constitutive rituals—but we all do recognize them. In some Pacific island cultures, a commoner must request permission to climb a palm tree above a chief; usually, however, in most cultures there is a tacit but clear understanding of whether rank is being constituted. In other contexts subordinates may precede superiors, or be above or ahead of them, without this position marking their rank. The origins of this iconic conformation of AR are presumably both phylogenetic and ontogenetic. In some contexts, such as foraging, all animals can judge vertical position, relative magnitudes, and temporal ordering; but not all animals have dominance hierarchies, and only humans use physical iconicity so extensively in the conformation of their social relations. However, every mammal has the experience of starting out smaller and weaker than its parents. Furthermore, in a number of species of primate and other taxa, dominance rank among juveniles and adults is partly determined by the number of coalition partners (often close kin) who support an animal in agonistic confrontations. Likewise, conflicts between groups are determined partly by relative numbers—as champanzee participants recognize (Wilson, Hauser, & Wrangham, 2001). Furthermore, in many species, agonistic encounters such as competitions over resources and females are decided in part by the animals' judgments of who is larger; in some instances, the outcome of conflicts is partly determined by relative strength. Hence Baldwinian selection must have resulted in innate predispositions to defer to bigger individuals backed up by more numerous allies. The facts of human social life contribute, as well. Wealthy and socially influential people are able to acquire larger dwellings, accumulate more possessions, accrue more followers, attract more mates, and sire or bear and raise more children. Where food is limited, concomitant nutritional benefits tend to result in greater stature and strength. In some mammalian species, dominance displays end with submission gestures in which one contestant signals subordination by lowering itself into a vulnerable position (see Spencer, 1896). Also, in human and nonhuman conflict, combatants who occupy the high ground enjoy a tactical ad-

3. CONSTITUTING RELATIONSHIPS

103

vantage. These factors might facilitate the representation of "above" as socially superior. In a threatening encounter, the individual behind (e.g., mate, offspring) is shielded or protected by the individual in front (see Goffman, 1976, pp. 72-74). The position in front thus tends to convey social assuredness or combative aggressiveness, which may tend to be correlated with age and rank. Temporal constitution of rank has similar antecedents. In some primates, each female tends to assume a rank immediately below her mother and above all elder sisters. Moreover, individual animals and groups often defer to "prior possession": The animal currently in possession of a valuable resource (food, nest site, consort) is more likely to retain it, prevailing over interlopers who arrive on the scene later. These two factors might predispose human use of temporal precedence to constitute AR. Overall, developmental experience generally resonates with and amplifies innate expectations; reciprocally but slowly, through Baldwinian selection, innate expectations are shaped by the adaptive benefits of rapid and reliable learning of objective contingencies, along with the benefits of competent participations in social relationships constituted by prevalent, enduring, socially transmitted mechanisms. Together, innate proclivities and experience in turn filter cultural reproduction over the long run so that socially transmitted representations tend to be congruent with them. Thus cultural mechanisms, while variable, for the most part further reinforce evolved and experientially acquired proclivities to constitute AR in space, time, and magnitude. CONSTITUTING EQUALITY MATCHING

People typically constitute Equality Matching relations in the medium of concrete matching operations. These operations demonstrably produce one-to-one correspondence, an operational definition of even balance. Examples include taking turns, comparison by side-by-side alignment, flipping a coin or conducting a lottery, using counting-out rhymes to choose who goes first, and voting. There is nothing in the Peircean tripartite typology corresponding to the conformation of EM, because Peirce apparently did not recognize the representational properties of procedures, and his taxonomy was not concerned with constitutive processes. However, Piaget's (1952) concept of concrete operations is helpful in understanding this conformation. Although it was intended as an ontogenetic stage theory, we can nonetheless apply Piaget's concept without concerning ourselves with the problematic theory of cognitive stages. Concrete operations are manipulative procedures for assessing physical properties such as mass, volume, or number. Piaget explores how children use concrete operations for judging equality or inequality.

104

FISKE

Such procedures include aligning objects in one-to-one correspondence arrays, juxtaposing solids to see if their lengths match, and comparing the height of liquids poured into containers. Given that concrete operations are essentially procedures for determining whether quantities are equal, it seems reasonable to ask whether the primary adaptive function of these capacities is social. However, Piaget was primarily interested in how children come to understand the physical world. Very early in his investigations he also applied his stage framework to moral development and social comprehension (1932, 1941-1951), but he did not explicitly discuss concrete operations for constructing socially balanced relationships. Nevertheless, the application to social relations is quite direct. One of the most pervasive concrete procedures for constituting EM is taking turns. When an entity is indivisible, even balance is achieved by rounds of successive possession. This fundamental operation is the procedure for rotating the Chairship of the United Nations Security Council, standing watches on a ship, working the night shift in some organizations, alternating turns in the favored seat in the car, making moves in checkers or tic-tac-toe, or dealing cards. Participants in a seminar or powwow may take turns holding the floor, going around the group one or more times. In pick-up sports contests, the captains of two teams may alternate choosing players for their teams. Opponents in ball sports alternate the goals they defend at halftime, with a timekeeper using a clock to match the number of minutes. In tennis, opponents switch sides of the court after odd-numbered games. In baseball or cricket, the two teams alternate at bat and in the field. In football, teams alternate in two other ways: They give up possession if they fail to make a first down and kick off when they score. Virtually all board games entail taking turns playing. The practical problem in turn-taking is to identify and agree on what a turn consists of, and when each party has properly taken a turn. Counting discrete actions (such as swings on a playground swing) and timekeeping procedures are the most prevalent devices to solve this problem. It is notable that in a long series of turns, people may not have any idea how many turns they took, and may not even recall who started, although it is patently obvious whose turn comes next in the alternation. When turns consist of doing something for, or giving something to, another party, this back and forth grades into balanced in-kind reciprocity. We call it "reciprocity" when a turn consists of re-turning the same thing: Helping someone with work balances previous help, while inviting someone to a feast is a return for attending their feast. In-kind reciprocity need not be bilateral; several parties—individuals or groups—may participate; the number is limited by the number who can attend an event, the feasibility of keeping track of participation, and the delay that is acceptable before the relationship is balanced. The reciprocated acts need not be benevolent: Tit-

3. CONSTITUTING RELATIONSHIPS

105

for-tat operates in aggression and violence, where harm is matched to harm. People also perform cycles of one-to-one correspondence to constitute EM. Participants' shares are ostensively matched by giving "one for you, one for you, and one for me; one for you...." Thus when children divide up cookies or proceeds from a lemonade stand, they may take one type of item and give one to each person, then go around again and again until the item is exhausted. Each type of item is handed out in a separate set of cycles: They set out pennies, one for each participant, going around as many complete rounds as they can; then they count out nickels, one to each; then dimes; then quarters. Whatever coins are left over may be given to Mom, set aside for the next occasion, or used to buy gum which can be evenly divided. Using the same ostensive matching procedure, a group of, say, six Moose age-mates will divide up a pile of kola nuts they receive, conducting cycles of one-to-one correspondence to create six equal piles. The remainder is broken into small pieces to permit a further complete cycle. The person or persons who do the distribution get last pick among the shares they have set up. No one pays attention to the absolute amount received, but they attend carefully to the procedure, which makes the shares ostensively equal. Likewise, Moose peers dividing a calabash of beer pass it around, each taking a drink, cycle after cycle, until they finish it off. This practice of passing out objects into one-to-one correspondence with the participants, going around as many rounds as necessary, is readily used by people who cannot do long division, serving as an operational definition of equality (Hunting & Sharpley, 1988; APF Moose field notes). Note that, like taking turns, a one-to-one correspondence operation does not depend on—or indicate—the cardinal quantity; performing or observing such an operation, participants may be unaware of how much they received, yet know that the distribution is equal.13 Other procedures exist for dividing up material objects into equal shares. One is to display or specifically align the objects to be distributed so their evenness is apparent by visual inspection. Mechanical devices such as a balance scale may be used; balance between the two pans operationally defines equality. Hence blindfolded Justice holding a pan balance is a meaningful symbol of equal treatment under the law. However, instead of matching one set of objects to another set of objects, each set can be matched to a cardinal sequence, and the end point of each sequence compared. This is counting. When enumerating items, actions, or other entities for each participant, the cardinal numbers mediate the correspondence: If you have 6 and she has 6 and I have 7, then our "Likewise, mathematical proofs to show that two infinite sets are equal are based on demonstrating that there is an operation for putting them in one-to-one correspondence.

106

FISKE

shares are not equal. If each of us flies 100 missions, the risks have been evenly divided. When we each have 11 players on the field, it's a fair match. This is accounting by tale: enumeration where each unit is treated as interchangeable, without regard for its specific features. Thus we put aside the fact that we faced different numbers of fighter attacks during our bombing missions, or the fact that the players on the two teams have unequal skills. When there is one indivisible role or entity to be allocated on a unique occasion, a common procedure is a fair lottery. Children have many traditions to manage this, including counting-out (or choosing up) rhymes to pick who is it, such as the English "Eenie, meenie, miney, mo...." There are several collections of counting-out rhymes by folklorists: Bolton (1888) cites rhymes from 22 cultures; Abrams and Rankin (1980) list 582 English counting-out rhymes with many variants. As children, my friends and I used another widespread procedure to choose up for first pick in forming softball teams. One person tossed a baseball bat to the other, who caught and held on; the tosser then grasped the bat immediately above the catcher's hand; then the catcher grasped it above the tosser; and so on, until the last person who could hold the bat by his finger tips was entitled to the first pick. Then the two captains would alternate picking players from among those present. (See Chase, 1976, for a description of this procedure in New York City stickball.) My children currently use pick a number: A third person secretly selects a number, say, between 1 and 20. Each of the others picks a number, and the one who comes closest is chosen. A group needing to select one person fairly—perhaps to undertake a dangerous task—may draw straws. One person holds a set of straws whose tops are aligned but whose ends are concealed, while the others choose; the person drawing the short straw is selected. Chance is often combined with turn-taking. Chess players assign color, with first move, by having one person hold a white pawn concealed in one fist, a black pawn in the other; the opponent chooses a fist, taking the color held there. Card players may use six or more concrete operations for evenly balancing the chances of each player: They shuffle, cut the deck, deal cards in cycles around the table, bid in cycles around the table, and rotate as dealer. They may also take turns hosting their games or supplying new decks. If there are many unique and dissimilar items to be allocated between two or more parties, for example in a divorce, the parties can make a complete list, then use another procedure such as a coin flip to decide who gets first choice. The two or more parties make alternating selections from the list, either in person or based on priority lists previously drawn up. If there is only one item, or if the issue is who gets the first turn, it may also be allocated by equal opportunity to stake a claim according to a recognized procedure. In American preadolescent culture, for example, the first person to call out "shotgun.*" claims the right to sit in the front passen-

3. CONSTITUTING RELATIONSHIPS

107

ger seat. The same procedure provides concrete equality of opportunity when people form a queue to buy scarce tickets or to get into a store holding a remarkable sale. Gambling may also function as a leveling device by redistributing wealth or goods (see Woodburn, 1982, on Hadza hunter-gatherers in East Africa and Mitchell, 1988, on the Wape of Papua New Guinea). However, in many cultures the leveling so produced is unintended and not directly constitutive, although the act of wagering against a random outcome may, in some sense, put the gamblers in an evenly balanced relationship. Simultaneity of action is another practice for matching. Starting a race at the same moment, evenly lined up along a line, makes it fair. Similarly, at the firing of a signal gun, homesteaders galloped into Oklahoma Territory to stake their claims. Eight-year-old children in my community use a rhyme and alternation procedure to synchronize the start of a "thumb war" in which the contestants grasp each other's fingers and try to pin each other's thumb. While swinging thumbs left and right of each other, the two contestants recite, "One, two, three, four, I declare a thumb war. Five, six, seven, eight, try to keep your thumb straight"—at that moment, the contest begins. Simultaneity may be combined with turns. In a Moose threshing bee, the threshers work alongside each other, swinging their flails in time to a drum; participants rotate through the threshing line, taking a turn doing less arduous parts of the work or resting before re-entering the threshing line. Similarly, both Moose and Tiv farmers convene weeding bees at which they line up and start together, hoeing across the field side by side to the rhythm of a drum or song (Bohannan, 1968, p. 72). If there is simultaneity in the beginning and end points, durations are matched: In a pie-eating contest, all contestants must begin and end at the same points in time. Likewise, working alongside others, beginning and ending at the same time, constitutes equality on work days in cooperatives. In these procedures, there is no need for measurement of the magnitudes involved, and the actual amount of time or work devoted to the activity may be unknown to participants, because the magnitude is socially irrelevant: Equivalence, operationally defined, is what matters. In events of this type, the participants' opportunities or contributions are not objectively equivalent in many other respects, as everyone recognizes—but by performing these procedures, they are operationally constituted as matched for the purposes at hand. Even when time is measured, the absolute magnitude is not crucial; what really matters is matching of times. At academic meetings and legislative hearings, session chairs attempt to keep all speakers to an identical time allotment. In many sports, each team defends each goal for the same time interval, measured in playing time. Chess manifests a more complex instrumentation, in which both players have the same total time for considering moves, but may distribute this time across the game as

108

FISKE

they please, punching the timer after each move. The timer is an operational mechanism for matching opportunity. Weighing is a specialized but sometimes important balancing procedure similar to timekeeping. In some races, horses have to carry weights to counterbalance the unequal weights of their jockeys. Wrestlers and boxers weigh in to assure that they compete against opponents in the same weight class. Automobile and motorcycle races are governed by maximum limits on displacement and minimum limits on weight, along with many other technical rules designed to equalize the engineering constraints. In some of these procedures, people transfer and hold material mnemonic tokens that function as place holders; these objects serve as manifest signs to aid memory and avert conflict over what happened. The person in possession of a turn may hold a ritual object, display distinctive insignia, or occupy a marked location. In a powwow, the person holding the pipe is in possession of the turn as speaker—the object is a visual, tactile reminder of whose turn it is. Passing it on completes the turn, demonstrably transferring the role to the next person. This tends to prevent interruptions. Concrete tokens are especially crucial when the number of participants exceeds two, when the actions to be balanced are not performed in front of all participants, or when the delays between act and reciprocation are long. Place-holding tokens are almost essential when the balancing is indirect, rather than bilateral (see Malinowski, 1922). Thus in a baby-sitting cooperative with many participants, chits can serve as concrete tokens for turns: Holding and transferring chits, it is not necessary for anyone to keep track of everyone's hours, much less for everyone to be informed of everyone's specific acts of baby sitting. Bride-wealth is the best studied example of place holding in the ethnographic literature. In most traditional cultures, kin groups give their daughters in marriage to other groups, expecting to receive brides in return, eventually. Sometimes the reciprocation is immediate, but when it is not, the delays can be substantial: Sometimes more than a generation passes. Moreover, in what Levi-Strauss (1949) called "generalized" exchange, Group A gives brides to Group B, but receives brides from Group C: The reciprocation is indirect and cyclical. (The loops can be closed in a variety of network paths.) When the reciprocation of brides is indirect or long-delayed, cultures almost invariably make use of place-holders: Before, during, or after the bestowal of the bride, the recipients transfer to the bride's kin group valuables such as cattle, ritual tools or weapons, or other objects. Typically the bride-wealth items are used only, or primarily, within a very restricted sphere of exchange, limited to the transfer of brides and perhaps compensation for homicides; they rarely or never enter into market transactions (Firth, 1938, pp. 340-344; Evans-Pritchard, 1940; Bohannan, 1955; Douglas, 1963; Meillassoux, 1975; Fiske, 1991).

3. CONSTITUTING RELATIONSHIPS

109

Moose sometimes make even-quantity exchanges of different goods, where a unit container serves as the standard of equality. Moose call this tekre, distinguishing it from gift giving, purchase, appropriation, and other forms of exchange. A woman goes to her neighbor with a basket full of, say, peanuts. If the neighbor is willing, they empty the peanuts onto a mat and fill the basket to the same level with, say, cotton. Although the neighbors are perfectly aware that the market price of these items differs, they are still content with the even exchange, just as Americans don't worry about equalizing the costs of the dinner party they gave and the one they attended. Likewise, in the Bolivian Andes, one practice among the Laymis in the Likina region is to exchange an item for the quantity of another substance that fills up the container. "For example, earthenware pots are exchanged against the amount of grain or ch 'unu (dehydrated potato) which they contain. When quality cloth is sought by Likina households they offer in exchange the cloth's capacity in maize grain" (Harris, 1982, p. 77). These operations manifestly level the transaction. A variety of political, legal, and organizational processes to assure "equal treatment" have been institutionalized, especially in Western democracies. In many contexts, people feel that a decision is legitimate if meets standards of procedural justice, such as permitting all parties to be heard and to correct invalid decisions (Tyler & Lind, 2001).14 For example, in American court proceedings, each side may exclude a certain number of potential jurors, and each party is entitled to take a turn presenting their case and questioning each witness. In political debates, procedural justice consists of giving each candidate equal time to make an opening statement (in an order determined by coin toss) and to reply to each question. Of course, they take turns answering questions. Procedures for EM may be borrowed from other contexts, and people create ad hoc inventions shaped by the natural proclivity for concrete operational conformation of EM. At my barber shop, the available barbers take turns cutting the hair of walk-in customers. However, since barbers may be on break, going out for coffee, or at various stages of cutting other customers' hair, there would often be ambiguity about who should take the next turn; simple errors of memory might also lead to disagreements. The mechanism they use is a clothespin labeled with the name of each barber, clipped to a wire. A barber coming on duty puts his or her clothespin at the left end of the row. The available barber whose clothespin is furthest to the right is entitled to cut the hair of the next customer who arrives; when 14

Tyler and Lind are concerned with the EM basis of satisfaction and perceptions of legitimacy regarding actions of authorities. In most of the contexts explored by research on procedural justice, EM relations are embedded under AR relations. Other examples of this embedding include the equal treatment expected from authorities by employees, students, co-wives, and children.

I 10

FISKE

the customer sits down in that barber's chair, the barber moves his or her clothespin to the left end of the wire. A glance at the line of clothespins provides an ostensive answer to who gets each customer. An undergraduate informant reported an ad hoc device she and her roommates had once used to organize house cleaning. A set of four roommates used a "chore wheel" consisting of a circular piece of laminated paper divided into quarters and labeled: "toilet," "floor," "shower," and "sink and mirror." In the center of this chore wheel rotated a smaller circle with the names of the four people who shared the bathroom. Every month, the inner circle turned, indicating the task assigned to each person that month. These matching procedures are often substituted for each other, and grade indistinguishably into one another. For example, first choice in choosing up teams can be decided by bat toss, coin flip, or first arrival on the field. These concrete operations also often co-occur in combinations. For example, alignment and simultaneity of starting are complementary components of starting a race. Sports teams flip a coin to allocate ends of the playing field or court for the first interval, then take turns; the football referee counts to make sure there are no more than 11 players on the field from each team. The operations of the most complex EM institutions, rotating credit associations, may entail complex combinations of ostensive matching (Ardner, 1964; Firth & Yamey, 1964). At regular intervals, participants each put in an identical sum; one participant takes home the combined total each time. The participant who takes the money may be preselected, or determined on each occasion by a lottery among those still eligible. If a participant drops out before the cycle is complete, then the balance is restored either by reimbursement between the drop-out and a replacement, or by reimbursements of the drop-out by each of the participants who have received a distribution. In some credit groups, some "participants" may be groups whose members equally split one share's contributions, and their distribution. Other combinations of operations enable people in traditional societies to divide arable land according to EM. They may begin by delineating plots that are each homogenous with respect to quality and crop-suitability. Then they measure each of these plots into equal parcels and assign one parcel in each plot to each household or kin group, often by a lottery or rotation (Ethiopia: Hoben, 1963, pp. 48-49; Taiwan: Barnett, 1971, p. 109; Andes: Harris, 1982, p. 74; premodern Denmark: Andersen, 1989). Moose and Yakut use a similar procedure when dividing a slaughtered cow or bullock into equal shares: They butcher the animal, separating it into distinct types of tissue (e.g., leg meat, ribs, intestines, fat); then they divide each respective type of meat into equal-sized piles, side by side; shares of each type are separate from the piles of other types of meat (APF field notes; Sieroszewski, 1896).

3. CONSTITUTING RELATIONSHIPS

I I I

These concrete procedures are operational definitions of EM: When duly performed, the interactive episode is balanced—it becomes EM. For example, when eligible voters cast ballots in a legitimately conducted election and votes are properly counted, then, ipso facto, the outcome constitutes a valid EM decision. When the national draft board selects numbers in a fair lottery, the selected persons are legitimately conscripted. When you invite me to dinner in return for dinner last month at my house, the hosting is balanced. When Moose neighbors fill the same basket in turn with the items they are exchanging (e.g., cotton for peanuts), this matching procedure makes the tekre fair and valid. In each case, participants recognize that, viewed in other frameworks, the costs or consequences are not the same for everyone—but within the framework of the procedure that constitutes the particular EM event, those differences are effectively irrelevant. The concreteness of the operation makes participants demonstrably even. That is, the balance is ostensively manifest in a procedure that operationally defines equality. There are no obvious phylogenetic precursors of concrete matching operations. Although there is some nonquantitative quid-pro-quo reciprocity in some primates' grooming (de Waal, 1996), no nonhuman animal has been observed to employ any of the foregoing procedures. Nor are there obvious ontogenetic sources in human infancy. Human conversation involves "turntaking" from the earliest babbling stage, but researchers have not noticed (or looked for) additive matching or efforts to attain even balance. It is revealing to see that Australian children aged 3; 10 to 4; 10 are unable to divide continuous substances into even shares and show little awareness of the meaning of an experimenter's request to do so (Hunting & Sharpley, 1988). Yet many of them have no problem with discrete objects; they can divide crackers among three dolls by allocating one to each doll in a cyclical procedure. This suggests that children understand a basic matching procedure for EM before they understand the semantic concepts of fraction or equality, let alone the symbols for them.

CONSTITUTING MARKET PRICING [Money is] a pure symbol, an abstract expression of abstract relationships. (Emile Durkheim, 1901, characterizing Georg Simmel's, 1900, message) Money is a system of symbols similar to language, writing, or weights and measures. Symbols do not merely "represent" something. They are material, oral, visual, or purely imaginary signs that form part of the definite situation in which they participate; thus they acquire their meaning. (Karl Polanyi, 1957, pp. 175,177)

I I2

FISKE

Money is ... a completely generalized symbolic medium. (Talcott Parsons, 1971, p. 27)

Market Pricing relies on the third Peircean aspect of signs, symbolism. In Peirce's typology, the symbolic aspect of a sign is the conventional component of its meaning, the meaning it derives from being used for a particular representational purpose. In a negative sense, Peircean symbols are said to be arbitrary, meaning just that they are neither iconic nor indexical—they do not resemble their referents or tend to co-occur with them (aside from their communicative or cognitive juxtaposition). Thus phonetic, written, and binary magnetic encoding of ideas are symbolic. Money is the quintessential symbol, since money has meaning only by convention—insofar as people accept money to mediate certain types of relationships. Modern forms of money have no intrinsic value. As Polanyi (1957), Einzig (1966), and Grierson (1978) point out, money acquires its meaning by virtue of its use in payment of obligations, as a standard of value, or as a means of indirect exchange. (Historically and ethnographically, different monies have often been used in each of these domains.) As a unit of account, money is the common denominator in ratio calculations of value. As a unit of exchange, money is the medium for proportional transactions. As a measure of outstanding claims and payment settlements, money is a common standard that mediates diverse MP obligations. Because of these functions, money can also operate as a device for storing and accumulating value, making it possible to link within the same coordinated field innumerable interactions over long periods of time. In short, whether as a scale for evaluating personal injuries or for representing commodities and services, money is an arbitrary standard whose value depends on social practices in a given community or network.15 Direct barter between two parties exchanging two commodities is possible without a tangible medium of exchange, but it requires the mediation of a symbolic cognitive representation of proportion—say, 2 apples are worth 3 oranges. Furthermore, archeology, history, and ethnography show that more extensive MP transactions in larger networks ineluctably give rise to a material medium of exchange. Initially, the medium is often a core subsistence commodity or basic tool, in terms of which other commodities and services are priced and debts settled (Dalton, 1965; Einzig, 1966). If trade, 15 Grierson (1978) offers plausible evidence suggesting that money may often have originated as an abstract representation for the practical consequences, and, especially, the sentiments of anger and grief, occasioned by the loss or alienation of persons—slaves, brides, or victims of homicide. Some of the earliest Western legal codes consist of conventional specifications establishing the precise compensation for death and for loss of specific body parts and specified chattel.

3. CONSTITUTING RELATIONSHIPS

I I3

rents, and wage labor expand, a symbolic system consisting of an arbitrary, conventional unit of account soon follows: money. With digital encoding in electronic networks, the material medium transforms into a purely abstract symbolic accounting. MP is defined as coordination with reference to ratios, and the very idea of a ratio depends on symbolic representation. There is virtually no way to communicate clearly about ratios without abstract numeral or prepositional expression. Numerical symbolism and prepositional articulation are essential to calculating costs in relation to benefits; inputs in relation to outputs; prices; interest; rents; wages; or taxes. Common discrete examples of symbolic coordination devices mediating MP are price tags, receipts, invoices, bank checks, stock quotes, and train tickets. More complex examples include commercial contracts, classified ads, budgets, ledgers, and the language of haggling in the market place. Symbolic constitution of MP is not limited to the use of money, numerals, or verbal language, however: Consider the hand signals used for bidding in auctions and making offers in commodity exchanges, or the auctioneer ending bidding with the bang of a gavel. Prices, wages, interest, and so forth represent the values of commodities relative to other commodities—as ratios. These ratios of value are denominated in terms of money. Thus a price is an abstract representation of the ratio of the MP value of a given quantity of commodity (or, for wages, interest, and rents, a given entity per unit of time) in proportion to all other commodities in the system. Monetary denominations are abstract units of these proportions. Money has no iconic or intrinsic causal relation to the commodities and services for which it can be exchanged or the functions it has in, say, tithes, rents, or payments of interest. Indeed, money has no necessary, intrinsic features of its own, even within any one culture (Simmel, 1900): Currency in diverse denominations, demand deposits, credit card balances and debt limits, lines of credit, electronic funds transfers, bank ledgers, and digitally encoded accounts have no common attributes except that they are forms of money. Money has no manifest history—no one knows or cares where the buyer got his money. As Simmel (1900) put it, money colorlessly and indifferently represents innumerable previous and potential objects, services and relations, most of which are remote in time or space. More precisely, money represents an abstract common quality combining exchange values, obligations, and account values, thereby relating all things commodified to all other commodities. Thus it is a purely symbolic link mediating an indefinite number of heterogeneous commodities, persons, and transactions that are otherwise remote from each other. In Parson's (1967) terms, money is a generalized symbolic medium of social inter-

I I4

FISKE

action, simultaneously a form of communication and a form of influence or inducement.16 The constitutive value of money depends purely on its acceptance by people who take for granted that others will generally accept it as a medium for exchanging and storing the capacity to pay fines or taxes, make restitution, or procure labor, land, goods, and other human artifacts, material and otherwise. This conventional cultural basis for the value of an arbitrarily designated set of monetary signs is what is meant by calling it a symbol system, in a Peircean sense (see Polanyi, 1957; Codere, 1968). However, money is not merely a representation; its meaning inheres in its mediation of the conduct of MP relations.17 Money functions constitutively, coordinating social relations. Of course, referential symbols such as verbal or written prices, taxes, wages, rents, or interest rates are signs that denote monetary values—they are symbolic representations of money, which in turn is the symbolic medium for the conduct of MP relations. In terms of the classical contrast between practical competence and semantic knowledge (Marett, 1909; Bergson, 1911; James, 1917; Russell, 1921; Merleau-Ponty, 1945; Ryle, 1949), obviously MP is based on abstract, propositionally formulated concepts. MP typically involves declarative propositions concerning value-relevant features of commodities. When advertising or haggling, these claims are intended to affect the price of the kola nuts, the baseball team, or the contract for gardening services. For example, when people negotiate contracts, they verbally articulate the reasons for the values of the various attributes of the entities they are transacting. To set a price to sell your car, you translate the make, model, year, mileage, features, and condition into a numerical price. Likewise, doing income taxes is an exercise in abstract semantics, interpreting propositionally stated rules to arrive at a numerical representation of a monetary obligation. Money is not the only symbolic medium for MP. Other semantic abstractions mediate some MP relations. Cost-benefit analysis entails drawing on semantic knowledge to produce an abstract articulation of the factors entering into the decision calculus. The numerator or denominator may repre16 By "generalized," Parsons did not mean that money is the medium of all social interaction, but rather that money is impersonal, without intrinsic connection to any person, event, or thing, so that the meaning of money derives from its mediation of diverse market exchanges and its omnipotent relation to all commodities. A nice overview of theories of money is Zelizer's (1994) first chapter. In counterpoint to Parsons, she shows that in many circumstances people do imbue money with circumscribed meanings that restrict its intended use to specific social functions. For a guide to other discourses on the symbolism of money, see GanfSmann's (1988) review of modern German social theorists. 17 Even when paper bank notes nominally represented gold, the gold itself had little or no representational meaning. Gold was an arbitrary conventional standard of value whose social significance depended on its potential use in mediating social relations.

3. CONSTITUTING RELATIONSHIPS

II5

sent any sort of quantity. Sometimes the units of account are vaguely defined utility, such as subjective effort or pleasure. A guy may break up with his girlfriend because she is "too high maintenance"—the investment in emotional energy is not worth the relationship payoffs. A family decides to get a dog after weighing diverse expected emotional costs and benefits for family members. In these everyday, informal group decisions, there are no explicit numbers, let alone ratios, but the interactions involve more or less articulate prepositional statements of the factors and likelihoods entering the subjective utility equation. To decide whether to take a job or plant tomatoes to sell, you have to consider the returns per unit of work, along with the investments and ongoing costs you will incur, temporally discounted. In such cases, you need to weigh alternatives, using proportions combined according to the distributive law. Unless you are a well-run corporation you probably use heuristics rather than creating a spreadsheet of numerical values, but to make a rational economic decision you must do something functionally equivalent to a cost-benefit analysis. Such decisions are, at the very least, enormously facilitated by the algebra of expected utility, which entails the manipulation of symbols such as prepositional statements about value-relevant features, numerical values, probabilities or frequency rates, opportunity costs, temporal discounting, and expected value. Thus a person might recruit friends to join a workout routine by making claims about the number of calories burned per hour. Baseball players are evaluated by their earned run averages and batting averages. Academics may be evaluated according to the number of citations to their work per year or per article. Universities take pride in their ratio of applicants to entering class size. A scientist might decide to write a popular book because she could reach more readers with a year's work than by spending the year writing for scientific journals. Family planning policymakers orient their activities toward birth rates, and under Robert McNamara, the U.S. Department of Defense seems to have made (or justified) decisions about Vietnam war strategy according to kill ratios. Significantly, deployment of the symbols that mediate MP creates commitments to (the terms of) relationships: Making a claim in a sales pitch, posting a price, gesturing to bid in an auction, and signing a contract are binding acts. Accepting payment for a commodity creates an obligation to provide it. Violations of these symbolic acts evoke strong emotions, arouse moral sentiments, and motivate sanctions. Very little is known about the phylogenetic or ontogenetic sources of symbolic constitution of social relations. Whereas animals with very simple nervous systems can calculate rates in the context of foraging and other learning (Gallistel, 1990), only humans use symbolically represented rates to coordinate social relationships. There has been a great deal of speculation, especially since Durkheim (1912), about the prehistoric origins of symbolic

I I6

FISKE

capacities. Most of the speculation has focused on language, although some theorists take into account prehistoric art and technology (e.g., Mithen, 1996). Aside from the Durkheimian tradition, most theories about the origins of symbols have considered only the representation/referential aspect of symbols (e.g., Deacon, 1997) without considering their constitutive ("performative" or "illocutionary") functions. Yet the constitutive function of symbols in organizing social coordination may be their original function and principal adaptive benefit. In all nonhuman animals, in any case, most types of communication function principally to coordinate social relations, not merely to convey referential information about nonsocial matters. Piaget claimed that children (even when schooled) do not attain mastery of formal operations, including declarative proficiency with proportions, until late middle childhood (Piaget, 1945, 1952). However, some initial understanding of formal operations with symbols, including fractions, emerges in social contexts around age 7 (Empson, 1999; Mix, Levine, & Huttenlocher, 1999; Watson, Campbell, & Collis, 1999).

HOW AND WHY CONFORMATION SYSTEMS FORM The Interdependent Components of Conformation Systems Having described the distinctive conformations of each of the four respective relational mods, I would like to return to the analysis of the components of these conformation systems. Let us begin by considering their connections, and then review each of them separately. For many analytic purposes, it is helpful to distinguish among five relational processes, according to their function: constitution (and conduct); communication; cognition; the development of the child's capacity to participate in social relations through cultural completion of innate but generic proclivities; and the historical transmission and diffusion of culture. But in practice these five processes are integrated, often to the point of being aspects of the same act. People relating in an intense CS mode, such as mother and child, implicitly think of each other as having some bodily essence in common; this cognitive representation is both cause and consequence of the proclivity to create CS by connecting bodies. Children are born with the implicit expectation that there are equivalence classes in their social world, but they cannot know, a priori, how or when these groups come into play, what defines those categories, who belongs to which ones, and which ones they themselves belong to. Children actively seek the cultural preos that answer these implicit but pressing questions, and they know where to look. They intuitively know that CS groups are

3. CONSTITUTING RELATIONSHIPS

II7

marked in the medium of consubstantial assimilation. They discern the CS groups to which they belong by the experience of being nursed, fed, held, and having others sleep alongside them. These experiences are intrinsically rewarding and elicit specific emotional bonds. Mother and child communicate through suckling, for example: They connect and comfort each other. Nursing thus indexically represents the same bond that it enactively contributes to creating. These acts also communicate the CS relationship to all concerned and are closely linked to the meaning and the normative effect of adult acts of connection such as eating and drinking commensally, hugging, caressing, or hand-shaking. Such acts of communion or bodily connection elicit distinct emotions and entail moral commitments that tend to sustain CS relationships. In short, there are homologies and mutual causal connections among all aspects of the conformation of CS. Likewise, the constitution of an AR relationship through bowing corresponds to the way that people cognize hierarchy as a vertical dimension, with people above or below each other. Abasing oneself also communicates to participants and observers the existence of a relationship: It is an iconic semiotic representation of ranking. Children, migrants, and ethnographers are attuned to attend to this communicative code and recognize its constitutive effect. Innate, implicit intuitions about the spatial encoding of AR permit people to identify the local forms and contexts of hierarchy, enabling these newcomers to participate in culturally informed status relations. Furthermore, ceremonially lowering one's body to a superior often elicits motivating emotions and evokes a legitimating normative framework for the relationship. Most social acts are multifaceted, with each facet of a conformation shaped by its intersection with the others. When the referee flips a coin at the start of a football match, that act is part of what constitutes it as a fair match. Flipping a coin makes participants and observers feel that the turntaking is fair, and obligates them to abide by the outcome. It also commits each team to defending particular goals and to a particular team kicking-off to start the first half or the second. Yet flipping a coin would be pointless unless the captains observed the result and communicated it to their teams. The coin flip also communicates to the team captains and the audience that each team has had an equal chance of kicking off: The coin flip is a representation of equality in practice. For the fans, seeing that the coin was tossed ostensively persuades them that the match is fair—imagine the outcry if the referee simply assigned goals and picked the receiving team without a coin toss. Concrete operations such as this are intuitively meaningful, readily understood devices that resonate with the EM model: Any 4year-old, and anyone from any culture, can readily appreciate what a coinflip does. So children and immigrants watching the coin flip from the stands discover that the event is structured as EM, and more specifically, that in

I I8

FISKE

American team sports, EM can be implemented with a coin toss. Each observed performance of a coin toss continues the reproduction of this implementation of EM, thereby potentiating and constraining future EM relations. Similarly, all of these aspects of MP are linked to each other. The evaluative and emotional components are necessarily joined with the communicative components. Buyers may be pleased or frustrated at the price a seller calls out, and a seller who proposes a price that is much too high commits a certain kind of infraction. Overcharging, usury, and exploitation are normative violations that angry buyers judge harshly; conversely, a fair price and especially a bargain motivate repeat business. Additionally, calling out a price commits the seller to convey the commodity at that price. Buyers will surely be offended if the seller refuses to follow through, and they have a cause for action. Announcing a price is manifestly informative: It is a signal representing a ratio of currency to commodity. Young children do not understand MP, but by early adolescence they develop the capacity to discern the prices of goods and services in their community. And they are attentive to these interesting and necessary cultural complements to their intuitive idea of socially meaningful ratios, rates, and proportions. In short, the medium in which people constitute each type of relationship is likely to correspond to the semiotic sign system for communicating about the relationship, the form of cognitive representation of the relationship, the way people learn the cultural implementations of the relational mod, and the processes of cultural transmission. These five components of social relations tend to be homologous (although they are never perfectly congruent). In order to see the causal interdependence among these five aspects of the conformation of social relationships, however, it is helpful to distinguish them analytically. So let's consider each aspect of conformation separately, and how it affects and is affected by the others. Constitution. The most fundamental aspect of sociality consists of the establishment of a framework for coordination. In relational models theory, to constitute a relationship is for some set of persons to adopt a relational mod for some aspect of their interaction, utilizing definite cultural complements (preos) that specify how to implement the mod in particular circumstances. Constitution includes initiating new relationships, but also modifying, rejecting, contesting, renewing, transgressing, redressing, and terminating relationships.18 18 As indicated earlier in this chapter, some actions have more impact than others in initially establishing or transforming relationships. So it is sometimes heuristically useful to distinguish between the constitution and the conduct of a relationship (see Rawls, 1955). For example, a birth or a wedding evoke special emotions and create moral commitments that provide the foundation for subsequent interaction. Similarly, a contract establishes the framework for ensuing MP relations; an election or a coin flip sets up an EM relationship that people then carry on

3. CONSTITUTING RELATIONSHIPS

I I9

To constitute a relationship is to evoke motives and emotions that modulate the interaction, to legitimate evaluations and moral judgments about it, and to acknowledge the validity of appropriate sanctions. To constitute a relationship is also to invoke commitment (Fiske, 2002; Nesse, 2001). All this is what D'Andrade (1984) calls the "directive force" of cultural models. Part of my thesis here is that the natural constitutive mechanism specific to each type of relationship intrinsically evokes strong corresponding emotions, motivations, and normative commitments. Sharing each other's blood in a brotherhood ritual, or undergoing together the fear and pain of brutal initiation rites, motivate people to commitments so powerful that they feel obligated to care for each other without limit, sometimes even sacrificing their lives for each other. Signing a contract rarely has this effect, but it has other, far-reaching effects; it may mobilize investment of vast sums of money and time, and motivate people to produce products or provide services. People are uniquely responsive to cues that fit the type of social relationship. Which has the greater impact: a girl saying to a boy, "I like you," or her brushing up against him and giving a quick kiss? On the other hand, a kiss from the Vice President for Human Resources may be less pleasing to most employees than his verbal assent to and subsequent signature on a union contract. Or consider which makes people feel that a race is fair: lining up in positions staggered according to some meticulously calculated and checked concentric elliptical distance formula, or simply lining up side by side on the same starting line? A starting line is ostensively even, while the staggered starting blocks on an oval track are less persuasive. Communication—Semiotics. People refer to relational events; represent to each other the existence, status, and history of relationships; convey evaluations; and signal social intentions. Communications are often about constitutive events, and often affect the quality of relationships. Communicating (about) a relationship means acting to influence another person's mental representations of it, often with the intention to shape the relationship itself. Conversely, communications are usually meant to indicate the communicator's cognitions: Signs are interpersonal representations of mental representations. Although signs and cognition are distinct systems, during the appropriate term or match. Likewise, building offices of different sizes on different floors with desks of varying sizes, and assigning people to those offices, creates a context for future action. In general, some acts institute a relational model that future acts presuppose. However, there is a fluid line between the constitution of a relationship and the conduct of the relationship: witness common-law marriage, or the constant maintenance of equal numbers of players during the substitution of players in a football match. In both cases, a pattern of conduct constitutes a relationship. Furthermore, people often constitute CS by refraining from violating the entailed food and sex taboos, which means that the relationship is continuously constituted.

120

FISKE

they tend to shape each other (see review by Lucy, 1992). Indeed, Peirce's (1931-1958) analysis of sign types integrates both types of representation. To a considerable degree, the process of developing competence to participate in culturally informed relationships consists of learning to understand communications about relationships and to communicate effectively about them. Although people do not explicitly tell children most of what they need to know about social relations, children actively seek to discover, monitor, and construct relationships (Fiske, 2004b). They do this by observing and participating in communicative processes—thereby reproducing (or transforming) the culture. Cognitive Representation. People attend to, perceive, interpret, categorize, make inferences, remember, and plan relationships (often without reflective awareness, or course). There are multiple heuristics for mental processing involving different systems of cognitive representation; all are selective, biased, and constructive. Because social relationships are crucial for human fitness, humans are likely to have evolved more or less domainspecific cognitive algorithms with specialized forms of encoding different basic types of sociality (Fiske, 1991; chap. 7, this volume). Cultural Complementation. In the framework of complementarity theory (Fiske, 2000), relational mods are taken to be innate and universal, but incomplete and indeterminate: Relational mods are necessary for social coordination but not sufficient. To organize meaningfully coordinated interaction, people must discover the local cultural parameters for implementing universal mods. These shared preos specify when, where, with whom, and how to implement, for example, the linear ordering that comprises the AR mod. Some anthropologists take it as axiomatic that culture consists of symbolic systems, where the meanings of symbols are arbitrary (by Peirce's definition). But if culture consisted only of symbols and the meanings of symbols were arbitrary—in the sense that the links between sign and referent category were random and purely conventional—then children, immigrants, and anthropologists might never discover those meanings. Children and other newcomers do discover the local cultural meanings of social acts because they are attuned to the specific media in which each mod is constituted and communicated: People are innately and intuitively responsive to the natural medium for each mod. For example, a person entering a society attends to space, time, magnitude, and force to discover the local cultural complements essential for implementing AR. Of course, in every culture people use multiple sign systems to display and construct each of the four relational mods. Having discovered the general cultural patterns and some of the relationship-specific features of the

3. CONSTITUTING RELATIONSHIPS

I2 I

conformation of AR in a particular culture, the novice can easily use this knowledge to learn the culture-specific signs for rank. Among Moose, for example, these include ceremonially dressing in "white" robes and carrying spears. No innate constraints would lead the novice to expect Moose to greet chiefs "Yztig be neere!" (May your place be beautiful); no innate focus of attention would lead novices to tune in directly to the significance of rectangular houses with ostrich eggs at the roof corners. But the newcomer learns that the people greeted and sometimes housed in this way are the people, who, at ceremonies, sit in front of others, are greeted before anyone else, are referred to as "big" and "strong," are addressed and referred to with plural pronouns, are accorded exceptionally extended personal space, receive the first and largest shares in distributions, and so forth. Any child or stranger immediately recognizes the relationship constituted and displayed when people stop, crouch, and await permission to approach a man in billowing robes and a big hat, and then lie on the ground and touch their foreheads to the dust in front of him (as Moose do when approaching a high chief). Newcomers expect that the conformation system for CS will be consubstantial assimilation. In every culture this implicit attunement is sufficient, and probably necessary, to discover how to culturally complete the innately structured and motivated but indeterminate CS mod. "The child has to learn through the carrying out of the daily routines of mealtimes the structure and unity of the family group, and its relations to the other households" of the community (Richards, 1932, p. 62). By observing with whom and in what manner people share food, children discover the core solidarities: "It is by lessons such as these that the whole kinship structure is felt rather than explained. Family sentiment is imprinted by a series of daily [food] habits rather than taught by any definite lesson or rule" (Richards, 1932, p. 66; see also p. 212). Likewise, "In just such a concrete fashion, the segregation of different groups at mealtimes makes [the child] aware of the interrelation of the various households comprising the village or clan" (p. 68). "The food routine teaches the child his place in a complicated social structure" (p. 72). In wider ritual contexts, food preparation and shared consumption also evoke and reflect sentiments of solidarity linking humans with their ancestors (Richards, 1932, pp. 79, 173, 186, 190). Children innately desire to relate to others; they need to form relationships. In particular, children expect to belong to CS groups and actively seek to form CS relationships. But children cannot know innately what primary or secondary groups operate in their own community, or who belongs to them. So children are looking for preos that complete their relational mods, enabling them to translate their relational proclivities into actual relationships. What they must do—what they do in fact—is attend to the particular constitutive and communicative media that permit them to connect with people in specific types of relationships. For CS, this is consubstantial

122

FISKE

assimilation. This is the medium that connects innate CS proclivities to the local culturally informed CS practices. To discover the cultural complements that specify how to realize indeterminate relational mods, a child needs to attend to and have an intuitive incipient understanding of others' communications about social relationships. The child needs to know that grooming, caressing, hugging, food sharing, and feeding are displays of CS. Moreover, children need to recognize the mechanisms for constituting relationships: The child needs to understand actions intended to create relationships with her, and grasp the relationships that other people are conducting among themselves. She needs to map what she perceives onto a cognitive representation that enables her to encode, categorize, process, store, and retrieve social memories. In turn, from birth, her mental representations of relationships selectively filter the social acts she notices and how she interprets them. This integration of cognitive and communicative representation, cultural complementation, and constitutive mechanisms is what makes it possible for humans to coordinate so closely, complexly, and diversely. Cultural Transmission. "Culture" is the set of aspects of human life that characterize a set of people because of their participation in a particular community or network. That is, culture is what people acquire by social transmission. The core of the social aspects of culture consists of the conformations described here. That is, with regard to CS, a culture is the local forms of consubstantial assimilation. For AR, a culture is a particular system of social physics, and, with respect to EM, a culture is the concrete operations people use for matching. For MP, a culture is the symbols for ratio systems of coordination. Innate expectations about how people construct or signify relationships affect what humans notice, learn, remember, and re-enact from the innumerable social constructions and signs people encounter. People will reproduce the expected ones most readily and reliably. Signs and constructions that do not conform to innate expectations are more likely to be ignored, incomprehended, forgotten—or transformed so that they do conform. This is a necessary consequence of the differential salience, learnability, memorability, and performability of different human practices: All cultural ideas, practices, and goals are not equally compatible with the human mind, human perceptual systems, human anatomy, and physiology. The compatible ones are more likely to be invented and retained over the generations. The human mind, along with the rest of the body, is the major filter and selective copier for cultural practices. It is the major selective force effecting differential invention, reproduction, and reshaping of cultural variants (Sperber, 1994, 1996; Deacon, 1997; Fiske, 2000). Hence if humans innately "expect" certain ways of constituting relationships, then human culture will tend to be informed accordingly. People

3. CONSTITUTING RELATIONSHIPS

123

will sometimes invent and transmit constitutive devices that do not conform to these expectations, but such inventions will be comparatively rare and will tend not to persist. What it comes down to is that children need to form relationships, and they will be able to relate to people only if they are innately predisposed to recognize and utilize the systems that people in their culture use to constitute relationships. Conversely, cultures have to be learnable and appealing to human children: Children will not be able to adopt ways of constituting relationships that they do not apprehend and they will tend not to use constitutive systems that do not resonate with the ways they think. Furthermore, when an existing cultural constitutive device fails to elicit emotional and moral commitments sufficient to sustain relationships, the relationships so constituted will not endure and that constitutive device will no longer be sustained; it will fail to reproduce. In short, there is a natural way to communicate about and constitute each relational model. The primary, natural conformation for each relational model is more effective in creating or modifying that kind of relationship: It has unique constitutive efficacy (and tends to override conflicting messages in other modalities). Most people in most cultures use these default conformations more often than others. These conformations are psychologically easier for people to use; people can understand, learn, and remember a relational model's primary conformation better than other modes of representing that model. Therefore the primary conformations are also the ones that tend to be most sustained and elaborated in most cultures. Indexes, icons, operations, and symbols have distinctive properties of transmission. What is preserved when an index is passed on is not what is preserved in the transmission of an icon. When children imitate a balancing procedure, the nature of the isomorphism between the model operation and the copy of it contrasts with the isomorphism between an original symbol in one generation and its reproduction in the next. Furthermore, for each conformation medium there are distinctive varieties of likely errors in reproduction, as well as systematic sorts of intentional stylistic variation. Also, for each conformation medium there are characteristic patterns of elaboration, reduction, or decay. And when people reject a conformation, there are contrasting sorts of antitheses for each medium. Exploring these patterns of cultural transmission and transformation would require a vast historical project which I have not undertaken, but an illuminating model for such an enterprise is Whitehouse's (2000) analysis of contrasting patterns of transmission of religious traditions encoded in episodic versus semantic media. Another speculative but stimulating comparison of media is Donald's (1991, 1993) theory of cultural evolution through a succession of distinct modes of cultural transmission.

124

FISKE

Conformation Systems. In sum, people most readily use and understand constitutive/communicative modalities that correspond to the mode of cognition they naturally tend to use to represent the corresponding type of relationship. Conversely, people are prone to adopt systems of cognitive representation for social relationships that are congruent with the semiotic systems people are using to conduct those relationships. This congruence is not logically necessary nor is it empirically complete or invariant. But the closer the fit between mental representation and social signs, the easier it is to conduct, cognize, and communicate social relationships. Communicating with other minds entails constraints (and opportunities) that differ from neural encoding and mental processing. Coordination of multiple attentional, mental, motivational, and action systems within an individual mind involves different problems than coordination or contestation of action, evaluation, and thought across the minds of multiple persons. But the closer the homology linking the cognitive and the communicative modalities, the easier the translation at the interface points of perception and articulation. Congruent semiotic and mental systems facilitate thought and social interaction and the integration of the two. Moreover, people's implicit processes of social thought are shaped by their perceptions of what a relationship is, as a function of how the relationship is socially marked; and people's actions to construct, or conduct, or comment on, or refer to a relationship are products of their implicit thoughts about the nature or essence of such a relationship. For example, if people think of hierarchy as inherently and fundamentally amounting to superiority (that is, aboveness), then they will readily tend to abase themselves to show deference and elevate themselves to show dominance. Conversely, if people display rank by positioning people above or below others, it is natural to think about rank as relative elevation. In short, this theory posits that there are mutual causal links and structural homologies among these five aspects of the conformation of social relations.19 The Relational Logic of the Conformation Systems Why do the relational mods have these particular conformations? I have mentioned some unique factors that may partially account for the respective conformations of each mod, but there is also a systematic determinant. 19

Because this theory of conformation systems closely links the constitution of social relationships, cognition, communication, and culture, it is somewhat Durkheimian. However, Durkheim posited that all complex cognition grows out of social cognition, a position that is untenable in light of what is now known about even solitary animals' capacities to categorize, navigate, and act in accord with ratios (Gallistel, 1990). Nonetheless, Durkheim was prescient in recognizing the interdependence among these systems, which pervasively inform each other.

3. CONSTITUTING RELATIONSHIPS

125

Each relational mod consists of certain relations and operations (Fiske, 1991, pp. 207-220), and each conformation system has distinct affordances that are congruent with the relations and operations of the corresponding mod. That is, given basic perceptual and cognitive capacities, each conformation system lends itself naturally to representing and constituting certain relations. Ordinal position is perceptually obvious in time, vertical elevation, position in front or behind, magnitude, and force. Linear ordering "jumps out" of these dimensions. Likewise, it is manifest to human perception that two bodies are connected when an infant emerges from the mother, when substances flow out of a body into another body, or when a set of people eat or imbibe commensally. Anyone experiencing birth or partaking in a commensal meal can readily appreciate the consubstantial assimilation that has occurred. Even-matching is "directly" evident when two arrays align in one-to-one correspondence, as when people reciprocate or take turns. Finally, symbols are virtually essential for MP: There are simply no good, clear, nonsymbolic, nonpropositional ways to represent ratios. Ratios are abstract ideas whose conceptualization and communication require semantic conventions for prepositional formulation. A ratio is a relation between two numbers, a relation readily represented in symbols. Other representations of ratios are cumbersome and ambiguous. In short, people use the particular conformation medium whose affordances are intrinsically congruent with the corresponding social relational structure. Let us examine the nature of these four congruencies. CS is an equivalence relation, a set of social categories. That is, CS consists of making binary discriminations among people: Two persons are either equivalent, or they are different. The equivalence of people in the same group inheres in sharing the same substance, markings, body surfaces, or genital modifications. In most societies, primary groups are often composed of commensals who observe the same food and sex taboos, go through initiation together, and conduct the same rituals, imitating each other's actions or acting in rhythmic synchrony. These conformation media inherently divide people into groups, because in each medium people are either equivalent, or different. People either share food, or they do not. What makes consubstantial assimilation apt for CS is that it creates equivalence classes, organizing relationships into a categorical system like a nominal scale. The physics of rank naturally creates linear orderings, like an ordinal scale. In relation to each other person, a person is above or below, in front or behind, bigger or smaller, more numerous or fewer, earlier or later, of greater or lesser force. Each dimension is inherently asymmetrical, inherently generating the transitive ordering that is AR. The structure of EM is that of an interval scale, or, more technically, an ordered Abelian field (see Fiske, 1991, pp. 209, 215-216). This structure consists of addition and its inverse, subtraction. Many of the conformation pro-

126

FISKE

cedures of EM consist of concrete operations of addition and subtraction. In particular, in-kind reciprocity amounts to adding and then subtracting the same quantity: Our group gives you a bride (-1 for us, +1 for you); then your group gives us a bride (+1 for us, -1 for you). Likewise, when I do you a favor and then you do me a favor, I've subtracted 1 and then added 1, while you've added 1 and then subtracted 1. In an ordered Abelian field, the equality or inequality remains unchanged if the same number is added to both sides of an equation. Thus cycles of distribution by one-to-one correspondence operationalize this structure: When each person gets one kola nut in each cycle, the relations are balanced. Likewise, if I return from a trip with a present for each child, this leaves unchanged the equality among them in relation to me. Taking turns can be seen as adding an element to each participant, so that when each person gets a turn, they remain equal. An ordered Abelian field must contain an element, zero, which is the additive identity (adding zero to an element yields the same element). Aligning along a starting line and simultaneity of action are zero elements: We begin at the same point. This relational structure can be conveyed symbolically (as I just did), but it is much easier to grasp and manipulate using concrete operations. Practically speaking, procedural coordination of EM works better than coordination based on indexes, icons, or symbols. An interval scale is composed of equal units, and the units in EM are often represented in tangibly matched tokens. In EM exchanges with long delays, place-holding valuables function as concrete units, representing an obligation to restore balance by giving a bride, for example, to the holders of the bride-wealth. Similarly, in a lottery or a game the objects or features representing outcomes must be manifestly equal (e.g., the sides of a coin, faces of a die, the bull's-eyes in the targets or the goals at the ends of the field). Many of the characteristic operations of EM are based on natural units or preexisting artifacts that are treated as even for this social function: a bride, a homicide, a meal, a ballot, a bid, the throw of the dice, a basket measure for Moose tekre. Many other EM operations are procedures to establish a set of equal entities, events, or other units. For example, hoeing a field beginning together in a line on one side and working side by side across to the other edge operationally defines the participants' work as equal. The chess clock, punched after each move, mechanically tracks consumption of the total interval each player has for deliberation. Material equality is established by visually inspecting allotments, and weighing them in the hands or in a pan balance. The structure of MP can be described as an Archimedean ordered field; it closely resembles a ratio scale of measurement. A symbolic system of propositions and abstract logical operations makes possible complex manipulation of ratios. The factors that affect these ratios can only be clearly formulated in semantic terms, so discussion and negotiation about them must be con-

3. CONSTITUTING RELATIONSHIPS

127

ducted in propositions. In short, only a symbolic conformation is suitable for the ratio scale relations and operations that comprise MP. Thus each of the four conformation systems is intrinsically congruent with the relations and operations of just one relational mod. In an ontological sense, this correspondence is the ultimate logical, natural determinant of the affinities between the relational mods and their respective modes of conformation. Each medium is aptly suited to just one mod.

IN REALITY Cultural Particulars In summary, the core thesis is that humans have universal predilections for these specific conformations of each relational mod: People are disposed to cognize each mod in a particular mode, culturally naive persons expect these conformations, and people resonate with them emotionally. The natural conformation of each mod is most readily learned, used, and understood; tends to evoke the most powerful sentiments; and inherently generates the strongest moral commitment. Indexical enactments are more natural and more efficacious than other means of constituting CS. Iconic physics is uniquely suited to AR. Concrete operations or procedures work better than other ways of constituting EM. MP has a special affinity with abstract symbolism. As a result of these predispositions, these are the most prevalent conformations within and across cultures. However, it is obvious that these predispositions are not the only predispositions in the human mind, and these conformations are certainly not the only factors in human social experience. Humans have quite a variety of evolved predispositions, motives, and capacities. Furthermore, a myriad of experiences influence human action, cognition, communication. Humans have extraordinary general intelligence and creative capacities. The history of any community and the biography of each person in it affect their social relationships. It hardly needs to be stated that for these reasons human sociality is not entirely determined by these conformation systems, and the present theory does not offer an exhaustive, neatly complete account of human sociality. Consider an analogy from physics. Newton's law of gravity is a universal and powerful determinant of the motion of objects, but trajectories are also affected by friction, the flow of any fluids in which objects move (e.g., wind), conservation of angular momentum, and of course other forces such as magnetic fields. Hence the motions of baseballs, balloons, airplanes, and even comets do not correspond exactly to the predictions of the law of gravity, alone. This does not mean that gravity is not universal, it just means that it isn't the only factor affecting motion. Likewise, when peo-

128

FISKE

pie constitute relationships or think about them, they are not limited to the primary conformation; the natural conformations are ubiquitous, but do not exhaust the possibilities. The patterns of conformation described here always affect human sociality, but obviously they are never the sole determinants of human social action.20 Moreover, in observing human action in everyday life, we must remember that people know their relationships and tend to take them for granted, without constantly deploying all of the constitutive devices they use on ceremonial or emblematic occasions. People are born only once and do not indefinitely continue to nurse at their mother's breast, but the bonds so created tend to be immutable. Hence the media for constituting relationships are not incessantly in evidence in the conduct of social life. When Prince Charles is playing polo, his opponents do not bow to him; in private, I presume that his girlfriend does not address him as "Your Highness" and he does not refer to himself in the plural when he speaks to her: It's hard to imagine him saying, "We are going to brush our teeth." But when people form, reinforce, redress, or contest social relationships, they generally use the specific media that are natural to the type of relationship. Every culture formulates unique aspects and combinations of the basic forms of social relationships. My aim is not to blindly disavow the diversity and particularity of cultures, but to describe the patterns in this diversity. Examining any feature of human action across the world's cultures, we find a wide distribution, but that distribution is never random. Any distribution has variance and outliers, but there must be some limit to the dispersion and an identifiable mean and mode. We can discern the shape of the distribution along with its central tendencies, and we can characterize the parameters of the variation. My thesis is simply that the primary conformation is the most prevalent across the world's cultures, and in most cultures is the most salient conformation. If you go to any culture, you can discern the local implementations of the four relational mods by relying on these constitutive media. That is how people discover how to relate across cultures. But people are capable of using, and do use, all kinds of procedures for constructing every kind of relationship. So in conjunction with social physics, there are a variety of other mechanisms that people use to co-constitute rank, including 20

The thesis that each relational mod operates in a distinct medium does not imply the converse, of course. Each medium has many other uses beyond the conformation of a particular relational mod. For example, the cognitive representation of space, time, magnitude, and force has many nonsocial functions. Even animals that live mostly solitary lives have such cognitive representations, which they use for foraging and navigating. Likewise, humans have multitudes of uses for arbitrary symbols, as the present essay, consisting of symbols, exemplifies. The use of symbols does not imply that people are Market Pricing.

3. CONSTITUTING RELATIONSHIPS

I 29

gender, descent, and charisma. Furthermore, there will always be additional, culturally particular and context-specific ways of constituting relationships: A rectangular house with ostrich eggs displayed on the roof corners signifies chieftainship only in a few West African cultures. Only modern militaries use stripes and stars on the shoulder to denote rank. And in any given culture, some of the specific component mechanisms may be absent—and occasionally even reversed for dramatic or dissimulative effect. For example, the Pope may get down on his knees to wash someone's feet. There are also exceptions that reflect particular local cultural concerns: When a Moose chief or elder travels ceremonially, he does not lead the procession: A young boy and sometimes others precede him on the trail so that they will step on and absorb any magical poison placed on the path to kill him. The most widespread practice incongruent with this theory is the use of symbolic systems, especially language, in the conformation of CS, EM, and particularly, AR. People use language to communicate about all relationships, to gossip, to declaim ideologies, or to analyze relationships (as we are doing here). And people sometimes think about non-MP relationships in a more-or-less prepositional medium. Also, people sometimes transmit cultural models of diverse relationships to children or peers through prepositional or imperative speech. This semantic transmission of cultural models is less common than commonly supposed (Fiske, 2004b), but it certainly occurs; in many cultures, elders give commands or instructions about manners, if nothing else. Furthermore, in Western and modern societies, authorities appoint people to offices verbally or in writing (Austin, 1961, 1962; Rawls, 1955). People also swear oaths, take vows, curse and utter imprecations, thereby forming or transforming relationships. These uses of language reflect the generic meta-cognitive and metapragmatic functions of language (and other symbol systems). Abstract symbolism and language in particular have broader representational potential than iconic and indexical signs or procedural operations. In that sense, these "exceptions" perhaps are more cogently analyzed as properties of symbol systems as such, rather than being viewed as a characteristic specific to the social uses of language. That is, while social relations are especially interesting, verbal reports and commentary are not limited to discussion of relationships; people talk about everything. People regularly use language to transform the world, whether as magical spells, imprecations, or taking an oath. Symbolism, and language in particular, are ubiquitously significant, but MP relations have a unique affinity with them because, as I have argued, MP relations depend on symbolic and prepositional conformation. In contrast, language and symbolism are not essential to CS, AR, or EM relationships.

130

FISKE

Combinations

As we have seen, there are numerous specific conformation devices within the media natural to each of the four basic types of relationship. This multiplicity means that individuals, institutions, and cultures usually have several compatible options. It would be interesting to explore how people make choices among these options. For example, when does the conformation of AR utilize above-below, when do people differentiate by magnitude, and when do people select earlier-later? Another consequence of the multiplicity of specific devices is that they can be used in combinations. These combinations may reinforce each other, as occurs in many initiation rites, for example.21 Or a graded series of devices, sometimes a Guttman scale based on progressive inclusion, may indicate degrees of CS intensity or AR differentiation (see Goodenough, 1944, 1965; also Firth, 1970 and Keating, 2000). Moreover, the devices a person deploys may be inconsistent—people may act discordantly, for example by including someone with one CS device while excluding them with another. A girl on a date may dance with a boy and poke popcorn in his mouth but refuse to kiss him. Furthermore, as Bateson (Bateson, Jackson, Haley, & Weakland, 1956) pointed out, verbal and nonverbal cues may communicate discordant messages requiring mutually exclusive responses. A mother might say "I love you" but refuse to cuddle her child. People may perform discordant constitutive or communicative acts in order to create ambiguity. Or they may wish to convey mixed messages that imply role distance, such as the implicit message "I do as I am supposed to, but I don't personally will it to be so." Or the discrepancy may indicate that, for example, once the basic fact of authority is established, the higher ranking person wishes to show that although he is, in fact, superior, he also values the others and wishes to honor them. Sometimes the inconsistencies become culturally standard schemas. In AR, a king may sit on a dais and receive a curtsy from a lady, yet usher her through the doorway into dinner ahead of himself and insist that she start to eat without waiting for him. This combinatorial potential also permits subtlety in social relations and people make use of it to maneuver. For example, Duranti (1994, chap. 3) highlights the constitutive/communicative heteroglossalia of multiple temporal orderings and spatial configurations among persons in Samoan gatherings. Agha (1994) reviews several studies that consider the ways that people combine linguistic honorifics. 21 It would be interesting to know in a given context and in general: Are the effects additive, multiplicative, or redundant? Only one study that I know of asked this question; it found additive effects of higher elevation, standing versus sitting, and front versus back, with very small interaction terms (Schwartz, Tesser, & Powell, 1982).

3. CONSTITUTING RELATIONSHIPS

131

The four types of conformation may be used simultaneously to constitute multiple concurrent relationships. A common example is the ordered positions at table at a very formal dinner, or the order of serving in an African meal or drinking event, in conjunction with the commensal consumption of the food and drink. Another example occurs, I believe, at some American Indian powwows. Participants pass a pipe around the circle, joining the group commensally as they each smoke. At the same time, passing the pipe is an EM place-holding operation: Each participant speaks in turn while they hold the pipe. But even when linked in such ways to create complex practices and institutions with unique properties, the two conformations are nonetheless distinct, as hydrogen and oxygen are distinct in H20, or protons and electrons are distinct in the hydrogen atom. Implications for Research and Practice

Conformation theory has a multitude of scientific, epistemological, and applied implications. I mention a few that have occurred to me. 1. Perhaps the most obvious implication is that there are natural channels for discovering the social processes in other cultures: Researchers (somewhat like immigrants or children) begin with highly probable a priori clues about the meaning of key forms of social actions. For archeology, this theory offers the prospect of better grounded frameworks for interpreting artifacts, architecture, and settlement plans of people we cannot know directly. 2. Moreover, this theory implies that to create and communicate about three of the most important types of social relationships—let alone to think about them—language is not at all necessary. Thus it need not be problematic to form CS, AR, and EM relationships with people who speak other languages. And this theory identifies the bases for relating to people who become aphasic or who have congenital language or symbolic impairmentsincluding people with autism. Many kinds of complex reports, critiques, and debates about relationships certainly do depend on abstract language, but language is not crucial for the basic conduct of CS, AR, or EM relationships. 3. Erving Goffman's engaging books and many contemporary ethnographies aim to reveal the subtle tactics required to "bring off" everyday interaction. Postmodern studies seek to reveal the strategies of hegemony, resistance, identity, and agency. This work has been greatly hampered by incomplete, unsystematic, and inexplicit understanding of the mechanisms for forming and modulating relationships. The present theory aims to delineate the key mechanisms, which should greatly facilitate research on social strategy and interaction tactics. It is impossible to fully analyze strategy or tactics without a clear conceptualization of the means that may be deployed.

132

FISKE

4. Each conformation system necessarily has distinct limitations and potentials. These virtues and vulnerabilities must affect the likelihood of sustaining each basic type of relationship in a given context. This is a promising, untouched frontier for research. A few examples indicate the sort of hypotheses that can be derived from the theory. Insofar as the core conformation of CS involves the passage of substances in and out of human bodies, CS relationships will tend to be stronger and more readily maintained in face-to-face interaction than when interacting at a distance or indirectly. In contrast, the abstract symbolic conformation of MP relationships imposes no such constraints, if people have ways of communicating symbolically over long distance and to large audiences, and if they can store the communications. Thus technology differentially affects the potential for each type of relationship. Hence electronic and other forms of mass communication facilitate the formation of symbolically mediated MP relationships and, to a considerable but much lesser degree, iconically mediated AR relations. Mass media and digital communication will have much less impact on the expansion of CS relationships beyond face-to-face networks. Likewise, proportional representative democracy depends on the institution of censussing, which in turn may depend on effective bureaucracies, literacy, computer technology, and statistical capabilities. In a similar manner, technological and procedural inventions can affect the scope and ease of EM relationships. The pan balance makes equality of weight ostensively manifest. The use of chits facilitate baby-sitting co-ops, while place holders or knots on a string may function as legitimating mnemonics that permit long-deferred reciprocity. Paper ballots make it possible for unlimited numbers of people to vote without being copresent—but leave ambiguity about what counts as an "X" or how to count hanging chads. The primary conformation of EM is concrete operations to add, subtract, and equalize material entities and units of action. Hence the theory implies that the potential for forming EM relationships in any given context depends on the availability of salient, readily manipulable units with naturally equatable affordances. So, for example, children who cannot make an even distribution of fluids or continuous substances may readily do so with discrete objects such as crackers (Hunting & Sharpley, 1988). A concomitant of the spatial and temporal conformation of AR is that people can gain authority and prestige if they create the perception of force, great magnitude and number, standing above and ahead. Political history shows the impact of such iconic devices (e.g., Strong, 1973). The archeological evidence suggests that complex societies are associated with the construction of imposing palaces, temples, tombs, monumentsedifices that perhaps impress upon the population the "greatness" of the ruler. In short, the prospects for social relationships of each type are functions of the feasibility of constituting, communicating, and learning them in the modalities that are natural to each. Environmental, technological,

3. CONSTITUTING RELATIONSHIPS

133

psychological, ontogenetic, and other factors that affect the feasibility of deploying a conformation system thereby affect the potential for constructing and sustaining the corresponding type of relationship. 5. Given what is known already about how maternal behavior, maternal bonding, and pair bonding are mediated by neuropeptides, endorphins, cortisol, and gonadal hormones, it would be informative to study the biochemistry of human responses to the consubstantial acts posited to constitute CS relationships. When more is learned about the biochemistry of nonhuman dominance hierarchies, similar studies could be done on the mechanisms for constituting AR. This kind of research might shed light on biochemical factors in personality disorders, psychopathy, and autism, and help explain the biochemical basis of narcotic addiction (cf. Insel & Young, 2001). Conversely, it would be equally informative to explore the other side of this mediation, looking at how changing levels of neuroactive chemicals such as MDMA, oxytocin, and arginine vasopressin affect dispositions to perform constitutive actions. 6. Constituting relationships involves invoking a relational modelthinking, acting, feeling with reference to a model. Ordinarily, when actors constitute relationships they intend that their partners (or third parties) will interpret and evaluate their actions with reference to the same model the actor is using. Each conformation medium specifies different kinds of information, with concomitant indeterminacy for the coordination of the relationship. That is, in each medium there are distinctive forms of latitude and ambiguity in the coordination of relationships. This affects the nature of the variation from interaction to interaction, from dyad to dyad, from community to community, and from generation to generation: What counts as a cultural reproduction of "the same" relationship depends on the medium of its conformation. For example, within MP, the units used for calculating ratios are of little practical importance. Thus there is nearly indefinite latitude for the monetary units, provided that everyone expects everyone else to value that symbol in future transactions; currency can be renamed, or the units reconfigured, without substantially affecting the structure of MP relationships. In CS, the ethnographic record on ritual covenants ("blood brotherhood") shows that within and across cultures there is considerable variation in the practices used to extract, mix, and ingest blood; indeed, sometimes the partners share a meal or comestibles, rather than mixing blood. The essential feature is the sense of having some shared substance in both partners' bodies. Within this constraint, there is wide latitude for variation as this institution diffuses across cultures or is transmitted across generations. In EM, there is indefinite diversity in the particular counting-out rhymes used as selection lotteries; the words don't matter much so long as the syl-

134

FISKE

lables are distinct and the rhyme is memorable. So such rhymes vary, while retaining certain operational features. Likewise, there are many demonstrable units of action that can count as a turn, and so long as an action is phenomenologically bounded, people can use it as a turn. Similarly, in material EM transactions, the affordances of any readily manipulable physically discrete objects or organisms provide natural units for matching. In AR, the affordances of space, magnitude, time and force offer many possible icons for social ordering, limited only by the mobility of the human body and the material constraints on human engineering. But color and shape are less likely to be invented or adopted in the conformation of AR, and, when introduced, are less likely to diffuse and endure. The nature of all these affordances, their constraints and latitude, need to be explored. 7. Another corollary of conformation theory is that there may be cognitive consequences of cultural differences in the prevalence of relational models. It seems likely that the mastery of a conformation system developed in social contexts may support corresponding capabilities in nonsocial domains. Where a given relational model is very extensively and complexly implemented, the population may develop greater proficiency with the corresponding form of cognition. That is, expertise in constituting, cognizing, and communicating a relational model in a particular mode may foster general faculties for cognizing and communicating in that mode. Conversely, competence developed in technical or artistic contexts might affect aptitudes for social cognition and communication. So, for example, exceptionally elaborated AR relationships in a culture might facilitate mental and communicative representations of space, time, magnitude, and force that could be applied in manufacture, mapping, or science; conversely, craftsmen's development of technical capacities to represent space, time, magnitude, and force might facilitate their development of fluency in representing AR. Thus conformation theory offers a systematic extension of Durkheim's (1912) theory of the social basis for cognition, an extension that goes beyond the currently popular contrast between cognition in independent societies versus cognition in interdependent societies (Nisbett, 2003).

Conclusion

Consubstantial assimilation, the medium for constituting Communal Sharing, is distinctive to CS. In contrast, people conduct Authority Ranking in a "social physics" of space, time, magnitude, number, and force. Equality Matching operates in another medium, concrete operations: To conduct EM, people use procedures that ostensively balance the relationship. Arbi-

3. CONSTITUTING RELATIONSHIPS

135

trary symbols are the prototypical medium of Market Pricing: numeric representations of prices, currency, the abstract language of bank checks and contracts, and money, the metasymbol. Conformation systems are the media of human sociality. They shape subjective perceptions and interpretation of relationships, along with memory and expectations about them. And they are the means through which people recognize each other's relational intentions and resistance. Furthermore, each basic type of social relationship is constrained and potentiated by the conformation system used to learn, discern, constitute and conduct it. Conformation systems also affect the nature of the variation in "the same" relationship within and across societies, as well as the relational errors and misunderstandings that arise. They are probably involved in aspects of pathology and other breakdowns in social coordination. This chapter offers theories at five levels. The more general hypotheses I advance could be true even if the more specific ones turn out not to be. Hence I conclude by distinguishing the eight successively nested theoretical propositions. 1. People use certain modes of action to constitute and conduct social relationships. 2. There are a number of distinct modes of constituting and conducting relationships. 3. Certain modes of action are more natural and more effective than others in constituting social relationships. 4. Different media are specifically effective for constituting different types of social relationships. 5. These differences are based on the natural affordances of different media. 6. These differences are also based on innate human proclivities and commonalities in human developmental experience. 7. People naturally tend to constitute Communal Sharing through consubstantial assimilation; Authority Ranking through space, time, magnitude, and force; Equality Matching through concrete operations that are ostensive balancing procedures; and Market Pricing through arbitrary symbols and propositional language. 8. Human sociality is based on an integrated conformation system composed of homologous subsystems for constitution and conduct, communication, cognitive representation, cultural completion, and channels of cultural transmission of relational models.

136

FISKE

REFERENCES Abrams, R., & Rankin, L. (Eds.). (1980). Counting-out rhymes: A dictionary. Publications of the American Folklore Society, Bibliographical and Special Series, Vol. 31. Austin & London: University of Texas Press. Agha, A. (1994). Honorification. Annual Review of Anthropology, 23, 277-302. Altorki, S. (1980). Milk-kinship in Arab society: An unexplored problem in the ethnography of marriage. Ethnology, 19, 283-244. Andersen, B. S. (1989). Plov og K01: Danskernes Virkelighed 1720-1814 [Plough and keel: The Danes' reality 1720-1814.] Copenhagen: Gyldendal. Ardner, S. (1964). The comparative study of rotating credit associations. Journal of the Royal Anthropological Institute, 94, 201-209. Austin, J. L. (1961). Performative utterances. In Philosophical papers. J. L. Austin (pp. 220-239). Oxford, England: Clarendon Press. Austin, J. L. (1962). How to do things with words. Oxford, England: Oxford University Press. Baldwin, J. M. (1896). A new factor in evolution. The American Naturalist, 30, 441-451,536-553. Baldwin, J. M. (1897). Social and ethical interpretations in mental development: A study in social psychology. New York and London: Macmillan. Baldwin, J. M. (1911). The individual and society; or psychology and sociology. Boston: Gorham Press. Banks, C. G. (1996). "There is no fat in heaven": Religious asceticism and the meaning of anorexia nervosa. Ethos, 24, 107-135. Barnett, W. K. (1971). An ethnographic description of Sanlei Ts'un, Taiwan, with emphasis on women's roles: Overcoming research problems caused by the presence of a great tradition. Master's thesis, Michigan State University. Barth, F. (1975). Ritual and knowledge among the Baktaman of New Guinea. Oslo: Universitetsforlaget; New Haven: Yale University Press. Bateson, G., Jackson, D. D., Haley, J., & Weakland, J. H. (1956/1972). Toward a theory of schizophrenia. Behavioral Science, 1, 251-264. Reprinted in G. Bateson, Steps to an ecology of mind; Collected essays in anthropology, psychiatry, evolution, and epistemology (pp. 201-227). San Francisco: Chandler. Baxter, P. T. W., & Almagor, U. (Eds.). (1978). Age, generation and time: Some features of East African age organizations. New York: St. Martin's Press. Beck, J., & Rosenbaum, M. (1994). The pursuit of ecstasy: The MDMA experience. Albany: State University of New York. Beidelman, T. 0. (1963). The blood covenant and the concept of blood in Ukaguru. Africa, 33, 321-342. Bergson, H. (1911). Matter and memory (N. M. Paul & W. S. Palmer [pseud.], Trans.). London: Allen & Unwin. Bohannan, P. (1955). Some principles of exchange and investment among the Tiv. American Anthropologist, 57, 60-70. Bohannan, P. (1968). Tiv economy. Evanston, IL: Northwestern University Press. Bolton, H. C. (1888). The counting-out rhymes of children: Their antiquity, origin, and wide distribution. New York: D. Appleton. (Re-issued 1969 by Singing Tree Press, Book Tower). Bourdieu, P. (1990/1980). The logic of practice. Stanford, CA: Stanford University Press. Brain, R. (1976). Friends and lovers. New York: Basic Books. Bravo, G. L. (2001). What does MDMA feel like? In J. Holland (Ed.), Ecstasy: The complete guide (pp. 21-38). Rochester, VT: Park Street Press. Brown, L. K., & Mussell, K. (Eds.). (1984). Ethnic and regional foodways in the United States. Knoxville: University of Tennessee Press.

3. CONSTITUTING RELATIONSHIPS

137

Brown, P. (1981). The cult of the saints: Its rise and function in Latin Christianity. Chicago: University of Chicago Press. Brown, P. (1988). The body and society: Men, women, and sexual renunciation in early Christianity. New York: Columbia University Press. Brown, P., & Levinson, S. (1979). Universals in language usage: Politeness phenomena. In E. Goody (Ed.), Questions and politeness (pp. 56-289). Cambridge, England: Cambridge University Press. Brown, R., & Gilman, A. (1960). The pronouns of power and solidarity. In T. A. Sebeok (Ed.), Style in language (pp. 253-276). Cambridge, MA: MIT Press. Bruner, J. S. (1964). The course of cognitive growth. American Psychologist, 19, 1-15. Reprinted in J. S. Bruner (1973), Beyond the information given: Studies in the psychology of knowing (pp. 325-351). New York: W. W. Norton. Bruner, J. S. (1966). On cognitive growth. In J. S. Bruner, R. R. Oliver, & P. M. Greenfield (Eds.), Studies in cognitive growth (pp. 1-29). New York: Wiley. Burton, R. V., & Whiting, J. W. M. (1961). The absent father and cross-sex identity. Merrill-Palmer Quarterly, 7, 85-95. Carr, L, lacoboni, M., Dubeau, M., Mazziotta, J. C., & Lenzi, G. L (2003). Neural mechanisms of empathy in humans: A relay from neural systems for imitation to limbic areas. Proceedings of the National Academy of Sciences, USA, 100, 5497-5502. Carter, C. S., DeVries, A. C., Taymans, S. E., Roberts, R. L., Williams, J. R., & Getz, L. L. (1999). Peptides, steroids, and pair bonding. In C. Carter, I. Sue, I. Lederhendler, & B. Kirkpatrick (Eds.), The integrative neurobiology of affiliation (pp. 169-181). Cambridge, MA: MIT Press. Chase, J. H. (1976). Street games of New York City. In B. Sutton Smith (Ed.), A children's games anthology. New York: Arno Press. Codere, H. (1968). Money-exchange systems and a theory of money. Man, 3, 557-577. Connerton, P. (1989). How societies remember. Cambridge, England: Cambridge University Press. Crawley, A. E. (1902/1965). The mystic rose: A study of primitive marriage and of primitive thought in its bearings on marriage. Revised and enlarged edition, 1927, reprinted in 1965. London: Spring Books. Dalton, G. (1965). Primitive money. American Anthropologist, 67, 44-65. Dampierre, E. de (1967). Un Ancien Royaume Bandia du Haut-Oubangui. Recherches en Sciences Humains #24. Paris: Librarie Plon. D'Andrade, R. G. (1984). Cultural meaning systems. In R. A. Shweder & R. A. LeVine (Eds.), Culture theory: Essays on mind, self, and emotion. Cambridge, England: Cambridge University Press. Dannenmaier, W. D., & Thumin, F. J. (1964). Authority status as a factor in perceptual distortion of size. Journal of Social Psychology, 63, 361-365. Darwin, C. (1872 /1965). The expression of the emotions in man and animals. Chicago: University of Chicago Press. Davis, J. H. R. (1977). People of the mediterranean: An essay in comparative social anthropology. London and Boston: Routledge & Kegan Paul. Deacon, T. W. (1997). The symbolic species: The convolution of language and the brain. New York: W. W. Norton. de Bono, M. (2003). Molecular approaches to aggregation behavior and social attachment. Journal of Neurobiology, 54, 78-92. Dettwyler, K. A. (1988). More than nutrition: Breastfeeding in urban Mali. Medical Anthropology Quarterly, 2, 172-183. de Waal, F. B. M. (1996). Good natured: The origins of right and wrong in humans and other animals. Cambridge, MA: Harvard University Press. de Waal, F. B. M. (2001). Apes from Venus; Bonobos and human sexual evolution. In F. B. M. de Waal (Ed.), Tree of origin: What primate behavior can tell us about human social evolution. Cambridge, MA: Harvard University Press.

138

FISKE

Donald, M. (1991). Origins of the modem mind: Three stages in the evolution of culture and cognition. Cambridge, MA: Harvard University Press. Donald, M. (1993). Precis of Origins of the modern mind: Three stages in the evolution of culture and cognition (with commentaries and reply). Behavioral and Brain Sciences, 16, 737-791. Douglas, M. (1963). TheLele oftheKasai. London: Published for the International African Institute by Oxford University Press. Douglas, M. (1966 /1985). Purity and danger: An analysis of the concepts of pollution and taboo. London: ARK (Routledge & Kegan Paul). Douglas, M. (1975). Deciphering a meal. In M. Douglas, Implicit meanings: Essays in anthropology (pp. 249-272). London: Routledge & Kegan Paul. Dugatin, L. A. (1997). Cooperation among animals: An evolutionary perspective. New York & Oxford: Oxford University Press. Dulaney, S., & Fiske, A. P. (1994). Cultural rituals and obsessive-compulsive disorder: Is there a common psychological mechanism? Ethos, 22, 243-283. Dumont, L. (1980). Homo hierarchicus: The caste system and its implications (M. Sainsbury, L. Dumont, & B. Gulati, Trans.). Chicago: University of Chicago Press. Dunbar, R. (1996). Grooming, gossip, and the evolution of language. Cambridge, MA: Harvard University Press. Duranti, A. (1994). From grammar to politics: Linguistic anthropology in a western Samoan village. Berkeley: University of California Press. Durkheim, E. (1893/1933). 7736 division of labour in society (G. Simpson, Trans.). New York: Free Press. Durkheim, E. (1901). Philosophic des Geldes [Review of Simmel 1900], L'Annee Sociologique, 5, 140-145. Durkheim, E. (1912/1965). The elementary forms of the religious life (J. W. Swain, Trans.). New York: Free Press. Eckerman, C. 0., Davis, C. C., & Didow, S. M. (1989). Toddlers' emerging ways of achieving social coordinations with a peer. Child Development, 60, 440-453. Eckerman, C. 0., & Didow, S. M. (1996). Nonverbal imitation and toddlers' mastery of verbal means of achieving coordinated action. Developmental Psychology, 32, 141-152. Eibl-Eibesfeldt, I. (1989). Human ethology. New York: Aldine de Gruyter. Einzig, P. (1966). Primitive money in its ethnological, historical and economic aspects (2nd ed.). Oxford, England: Pergamon. Eisenstadt, S. N. (1956). Ritualized personal relations—Blood brotherhood, best friends, compadre, etc.: Some comparative hypotheses and suggestions. Man, 56, 90-95 (Item #96). Empson, S. B. (1999). Equal sharing and shared meaning: The development of fraction concepts in a first-grade classroom. Cognition & Instruction, 17, 283-342. Evans-Pritchard, E. E. (1933). Zande blood-brotherhood. Africa, 6, 369-401. Evans-Pritchard, E. E. (1940/1967). The NuerA description of the modes of livelihood and political institutions of a Nilotic people. Oxford, England: Clarendon Press. Evans-Pritchard, E. E. (1951). Kinship and marriage among the Nuer. Oxford, England: Clarendon Press. Fajens, J. (1985). The person in social context: The social character of Baining "psychology." In G. M. White & J. Kirkpatrick (Eds.), Person, self, and experience: Exploring Pacific ethnopsychologies (pp. 367-397). Berkeley: University of California Press. Fildes, V. (1988). Wetnursing:A history from antiquity to the present. New York: Basil Blackwell. Firth, R. (1938/1965). Primitive Polynesian economy (2nd ed.). New York: Norton. Firth, R. (1970). Postures and gestures of respect. In J. Pouillon & P. Maranda (Eds.), Echanges et communications: Melanges offerts a Claude Levi-Strauss a {'occasion de son 60eme anniversaire (pp. 188-209). The Hague: Mouton. Reprinted in T. Polhemus (Ed.), 1978, in association with the Institute of Contemporary Arts, London, The body reader: Social aspects of the human body. New York: Pantheon.

3. CONSTITUTING RELATIONSHIPS

139

Firth, R. W., & Yamey, B. S. (1964). Capital, saving and credit in peasant societies: Studies from Asia, Oceania, the Caribbean and Middle America. Chicago: Aldine. Fiske, A. P. (1991). Structures of social life: The four elementary forms of human relations. New York: Free Press. Fiske, A. P. (2000). Complementarity theory: Why human social capacities evolved to require cultural complements. Personality and Social Psychology Review, 4, 76-94. Fiske, A. P. (2002). Socio-moral emotions motivate action to sustain social relationships. Self and Identity, 1, 169-175. Fiske, A. P. (2004a). Modes of constituting social relationships among the Moose of Burkina Faso. Manuscript in preparation. Fiske, A. P. ( 2004b). Learning culture the way informants do: Observing, imitating, and gradually participating. Manuscript in progress. Fiske, A. P., & Haslam, N. (1997). Is obsessive-compulsive disorder a pathology of the human disposition to perform socially meaningful rituals? Evidence of similar content. Journal of Nervous and Mental Disease, 185, 211-222. Fortes, M. (1945). The dynamics of clanship among the Tallensi: Being the first part of an analysis of the social structure of a trans-Volta tribe. London: Oxford University Press. Fortes, M. (1966). Introduction. In J. Goody (Ed.), The developmental cycle in domestic groups (pp. 1-14). Cambridge Papers in Social Anthropology, No. 1. Cambridge, England: Cambridge University Press. Fortes, M. (1969). Kinship and the social order. Lewis Henry Morgan Lectures. Chicago: Aldine. Frazer, J. G. (1890/1987). The golden bough: A study in magic and religion. Abridged ed. New York: Collier Books. Freud, S. (1913). Totem and taboo. In J. Strachey (Ed.), The standard edition of the complete works of Sigmund Freud (Vol. 13). London: Hogarth Press. Freud, S. (1939). Moses and monotheism (K. Jones, Trans.). New York: Vantage. Gallese, V. (2003). The manifold nature of interpersonal relations: The quest for a common mechanism. Philosophical Transactions of the Royal Society of London: Biological Sciences, 358, 517-528. Gallistel, C. R. (1990). The organization of learning. Cambridge, MA: MIT Press. GanfSmann, H. (1988). Money—A symbolically generalized means of communication? On the concept of money in recent sociology. Economy and Society, 17, 285-316. Garvin, P. L., & Riesenberg, S. H. (1952). Respect behavior on Ponape: An ethnolinguistic study. American Anthropologist, 54, 201-220. Gibson, J. J. (1979). The ecological approach to visual perception. Boston: Houghton Mifflin. Giesey, R. E. (1985). Models of rulership in French royal ceremonial. In S. Wilentz (Ed.), Rites of power: Symbolism, ritual, and politics since the middle ages (pp. 41-64). Philadelphia: University of Pennsylvania Press. Gillison, G. (1983). Cannibalism among women in the eastern highlands of Papua New Guinea. In P. Brown &D. Tuzin (Eds.), The ethnography of cannibalism (pp. 33-50). Washington, DC: Society for Psychological Anthropology. Girard, R. (1977). Violence and the sacred. Baltimore: Johns Hopkins University Press. Goffman, E. (1971). Relations in public: Microstudies of the public order. New York: Basic Books. Goffman, E. (1976). Gender advertisements. New York: Harper Colophon. Goodenough, W. (1944). A technique for scale analysis. Educational and Psychological Measurement, 4, 179-190. Goodenough, W. (1965). Rethinking 'status' and 'role': Toward a general model of the cultural organization of social relationships. In M. Banton (Ed.), The relevance of models to social anthropology. Association of Social Anthropologists Monograph 1. London: Tavistock. Greer, G., & Tolbert, R. (1986). Subjective reports of the effects of MDMA in a clinical setting. Journal of Psychoactive Drugs, 18, 319-327. Grierson, P. (1978). The origins of money. Research in Economic Anthropology, I, 1-35.

140

FISKE

Gutierrez, C. P. (1984). The social and symbolic uses of ethnic/regional foodways: Cajuns and crawfish in south Louisiana. In L. K. Brown & K. Mussell (Eds.), Ethnic and regional foodways in the United States (pp. 169-182). Knoxville: University of Tennessee Press. Hanley, S. (1985). Legend, ritual, and discourse in the Lit de Justice assembly: French constitutional ideology, 1527-1641. In S. Wilentz (Ed.), Rites of power: Symbolism, ritual, and politics since the middle ages (pp. 65-106). Philadelphia: University of Pennsylvania Press. Harlow, H. F. (1958). The nature of love. American Psychologist, 13, 673-685. Harlow, H. F., Harlow, M. K., & Hansen, E. W. (1963). The maternal affectional system of rhesus monkeys. In H. L. Rheingold (Ed.), Maternal behavior in mammals (pp. 254-281). New York: Wiley. Harris, 0. (1982). Labor and produce in an ethnic community. In D. Lehmann (Ed.), Ecology and exchange in the Andes (pp. 77-96). Cambridge, England: Cambridge University Press. Hartland, E. S. (1895/1972). The legend of Perseus; A study of tradition in story, custom and belief, Vol. 2. Originally published in London by David Nutt; reissued in New York by AMS Press. Hawkes, K., O'Connell, J. F., & Blurton Jones, N. G. (2001). Hunting and nuclear families: Some lessons from the Hadza about men's work (with commentaries & reply). Current Anthropology, 42, 681-709. Heiser, M., lacoboni, M., Maeda, F., Marcus, J., & Maziotta, J. C. (2003). The essential role of Broca's area in imitation. European Journal of Neuroscience, 17, 1123-1128. Henry, J. A., & Rella, J. G. (2001). Medical risks associated with MDMA use. In J. Holland (Ed.), Ecstasy: The complete guide (pp. 71-86). Rochester, VT: Park Street Press. Henry, S. (1988). Rules, rulers, and ruled in egalitarian collectives: Deviance and social control in cooperatives. In J. G. Flanagan & S. Rayner (Eds.), Rules, decisions and inequality in egalitarian societies. Avebury: Aldershot, UK. Hensley, W. E., & Cooper, R. (1987). Height and occupational success—A review and critique. Psychological Reports, 60, 843-849. Hepper, P. G. (Ed.). (1991). Kin recognition. New York: Cambridge University Press. Herdt, G. H. (Ed.). (1982). Rituals of manhood: Male initiation in Papua New Guinea. Berkeley: University of California Press. Hewes, G. W. (1955). World distribution of certain postural habits. American Anthropologist, 57, 231-244. Higham, P. A., & Garment, D. W. (1992). The rise and fall of poIiticians-The judged heights of Broadbent, Mulroney and Turner before and after the 1988 Canadian federal election. Canadian Journal of Behavioural Science—Revue Canadienne des Sciences du Comportement, 24, 404-409. Hirschfeld, L. A. (in press). Children's understanding of racial groups. In M. Barrett & E. Buchanan-Barrow (Eds.), Children's understanding of society. Philadelphia: Psychology Press. Hoben, A. (1963). The role ofambilineal descent groups in Gojjam Amhara social organization. Doctoral thesis, University of California, Berkeley. Hohmann, G., & Fruth, B. (2000). Use and function of genital contacts among female bonobos. Animal Behaviour, 60, 107-120. Holland, J. (200la). The godparents of MDMA: An interview with Ann and Sasha Shulgin. In J. Holland (Ed.), Ecstasy: The complete guide (pp. 58-65). Rochester, VT: Park Street Press. Holland, J. (2001b). MDMA and society: Introduction. In J. Holland (Ed.), Ecstasy: The complete guide (pp. 346-349). Rochester, VT: Park Street Press. Hollis, A. C. (1905). The Masai: Their language and folklore. Oxford, England: Clarendon Press. Hunting, R. P., &Sharpley, C. F. (1988). Fraction knowledge in preschool children. Journal for Research in Mathematics Education, 19, 175-180. Insel, T. R., & Young, L. J. (2001). The neurobiology of attachment. Nature Reviews Neuroscience, 2, 129-136.

3. CONSTITUTING RELATIONSHIPS

141

Insel, T. R., Young, L, & Wang, Z. (1999). Molecular aspects of monogamy. In C. S. Carter, I. I. Lederendler, & B. Kirkpatrick (Eds.), The integrative neurobiology of affiliation (pp. 183-197). Cambridge, MA: MIT Press. James, E. 0. (1917). Primitive ritual and belief: An anthropological essay. London: Methuen. Jaymous, R. (1981). Honneur et Baraka: Les structures sociales traditionelles dans le Rif. Cambridge, England: Cambridge University Press; Paris: Editions de la Maison des Sciences de I'Homme. Kakridis, H. J. (1963). La notion de I'amitie et de I'hospitalite chez Homere. H Bibliothke tou Philologou, ar 9. Thessaloniki. Keating, E. (2000). Moments of hierarchy: Constituting social stratification by means of language, food, space and the body in Pohnpei, Micronesia. American Anthropologist, 102, 303-320. Kendrick, K. M. (2000). Oxytocin, motherhood and bonding. Experimental Physiology, 85, 111S-124S. Khatib-Chahidi, J. (1992). Milk kinship in Shi'ite Islamic Iran. In V. Maher (Ed.), The anthropology of breast feeding: Natural law or social construct. Cross-Cultural Perspectives on Women. Oxford, England, and Providence, RI: Berg (distributed by St. Martin's Press, New York). Klam, M. (2001). Experiencing ecstasy. New York Times Magazine, January 21, 38-43, 64, 68, 72, 78-79. Lakoff, G., & Johnson, M. (1980). Metaphors we live by. Chicago & London: University of Chicago Press. Laponce, J. A. (1981). Left and right: The topography of political perceptions. Toronto: University of Toronto Press. Laye, C. (1966). L 'enfant noir. J. A. Hutchinson (Ed.). Cambridge, England: Cambridge University Press. Leach, E. (1972). The influence of cultural context on non-verbal communication in man. In R. A. Hinde (Ed.), Non-verbal communication (pp. 315-344). Cambridge, England: Cambridge University Press. Lee, S. S. (2000). Dys-appearing tongues and bodily memories: The aging of first-generation resident Koreans in Japan. Ethos, 28, 198-223. Lester, R. J. (1995). Embodied voices: Women's food asceticism and the negotiation of identity. Ethos, 23, 187-222. Levi-Strauss, C. (1949/1961). The elementary structures of kinship (Rev. ed.). (J. H. Bell, J. R. von Sturmer, & R. Needham, Trans.). Boston, MA: Beacon. Lindblom, G. (1920). The Akamba in British East Africa: An ethnological monograph (2nd ed.). Upsala: Archives d'Etudes Orientales, Vol. 17. Lindenbaum, S. (1979). Kuru sorcery: Disease and danger in the New Guinea highlands. Palo Alto, CA: Mayfield. Lucy, J. A. (1992). Language diversity and thought: A reformulation of the linguistic relativity hypothesis. Cambridge, England: Cambridge University Press. Lutz, C. (1988). Unnatural emotions: Everyday sentiments on a Micronesian atoll and their challenge to western theory. Chicago: University of Chicago Press. Malinowski, B. (1935/1978). Coral gardens and their magic: A study of the methods of tilling the soil and of agricultural rites in the Trobriand Islands. New York: Dover. Manson, J. H., Perry, S., & Parish, A. R. (1997). Nonconceptive sexual behavior in bonobos and capuchins. International Journal of Primatology, 18, 767-786. Marett, R. R. (1929/1909). The threshold of religion. London: Methuen. Matthiesen, A. S., Ransjo-Arvidson, A. B., Nissen, E., & Uvnas-Moberg, K. (2001). Postpartum maternal oxytocin release by newborns: Effects of infant hand massage and sucking. Birth-Issues in Perinatal Care, 28, 13-19. Mauss, M. (1936/1973). Les techniques du corps. In M. Mauss, Sociologie et anthropologie (pp. 363-386). Paris: Presses Universitaires de France. Translated as The notion of body techniques (1979), Sociology and psychology: Essays (B. Brewster, Trans.), (pp. 97-123). London: Routledge & Kegan Paul.

142

FISKE

McDougal, W. (1908). An introduction to social psychology. London: Methuen. McNeill, W. H. (1995). Keeping together in time: Dance and drill in human history. Cambridge, MA: Harvard University Press. Mehrabian, A. (1981). Silent messages: Implicit communication of emotions and attitudes (2nd ed.). Belmont, CA: Wadsworth. Meillassoux, C. (1975/1981). Maidens, meal and money: Capitalism and the domestic economy. Cambridge, England, and New York: Cambridge University Press. Melamed, T., & Bozionelos, N. (1992). Managerial promotion and height. Psychological Reports, 71, 587-593. Meltzoff, A. N., & Decety, J. (2003). What imitation tells us about social cognition: A rapprochement between developmental psychology and cognitive neuroscience. Philosophical Transactions of the Royal Society of London: Biological Sciences, 358, 491-500. Merleau-Ponty, M. (1962/1945). Phenomenology of perception (C. Smith, Trans.). New York: Humanities Press. Miller, L. F., Rozin, P., & Fiske, A. P. (1998). Food sharing and feeding another person suggest intimacy: A study of American college students. European Journal for Social Psychology, 28, 423-436. Mitchell, W. E. (1988). The defeat of hierarchy: Gambling as exchange in a Sepik society. American Ethnologist, 15, 638-657. Mithen, S. J. (1996). The prehistory of the mind: A search for the origins of art, religion, and science. London: Thames & Hudson. Mix, K. S., Levine, S. C., & Huttenlocher, J. (1999). Early fraction calculation ability. Developmental Psychology, 35, 164-174. Montagu, A. (1979). The skin, touch, and human development. In S. Weitz (Ed.), Nonverbal communication: Readings with commentary (2nd ed., pp. 328-337). New York: Oxford University Press. Morris, D. (1977). Man watching: A field guide to human behavior. New York: Harry Abrams. Nesse, R. M. (Ed.). (2001). Evolution and the capacity for commitment. New York: Russell Sage Foundation. Nisbett, R. E. (2003). The geography of thought: How Asians and westerners think differently. . . and why. New York: Free Press. Olausson, H., Lamarre, Y., Backlund, H., Morin, C., Wallin, B. G., Starck, G., Ekholm, S., Strigo, I., Worsley, K., Vallbo, A. B., & Bushnell, M. C. (2002). Unmyelinated tactile afferents signal touch and project to insular cortex. Nature Neuroscience, 5, 900-904. Olaveson, T. (in press). 'Connectedness' and the rave experience: Rave as new religious movement? In G. St. John (Ed.), Rave culture and religion. London: Routledge. Ottenberg, S. (1994). Initiations. In P. K. Bock (Ed.), Psychological anthropology (pp. 351-377). Westport, CT: Praeger. Panksepp, J. (1998). Affective neuroscience: The foundations of human and animal emotions. New York: Oxford University Press. Parsons, T. (1967). On the concept of influence. In T. Parsons, Sociological theory and modern society (pp. 355-382). New York: Free Press; London: Collier-Macmillan. Parsons, T. (1971). Levels of organization and the mediation of social interaction. In H. Turk & R. L. Simpson (Eds.), Institutions and social exchange: The sociologies of Talcott Parsons and George C. Homans (pp. 23-35). Indianapolis: Bobbs-Merrill. Peirce, C. S. (1931-1958). The collected papers of C. S. Peirce. C. Hartshorne, P. Weiss, & A. W. Burks (Eds.). Cambridge, MA: Harvard University Press. Peirce, C. S. (1985). Logic as semiotic: The theory of signs. In R. Innis (Ed.), Semiotics: An introductory anthology (pp. 1-23). Bloomington: Indiana University Press. Perry, S., Baker, M., Fedigan, L., Gros-Louis, J., Jack, K., MacKinnon, K. C., Manson, J. H., Panger, M., Pyle, K., & Rose, L. (2003). Social conventions in wild white-faced capuchin monkeys: Evidence for traditions in a neotropical primate. Current Anthropology, 44, 241-268.

3. CONSTITUTING RELATIONSHIPS

I 43

Piaget, J. (1932). The language and thought of the child (2nd ed.). New York: Harcourt, Brace. Piaget, J. (1941-1951/1965). Etudes sociologiques. Geneva: Librarie Droz. Piaget, J. (1945). La formation du symbole chez I'enfant: Imitation, jeu et reve, image et representation. Neuchatel, Switzerland: Delachaux & Niestle. Piaget, J. (1952). The child's conception of number. New York: Humanities Press. Pitt-Rivers, J. (1968). Pseudo-kinship. In D. L. Sills (Ed.), International encyclopedia of the social sciences (Vol. 7, pp. 408-413). New York: Macmillan and Free Press; London: Collier-Macmillan. Pitt-Rivers, J. (1973). The kith and the kin. In J. Goody (Ed.), The character of kinship (pp. 89-105). Cambridge, England: Cambridge University Press. Polanyi, K. (1957). The semantics of money uses. Explorations, 8, 19-29. Reprinted in G. Dalton (Ed.). (1968). Primitive, archaic and modern economies: Essays of Karl Polanyi (pp. 175-190). Garden City, NY: Doubleday (Anchor Books). Poole, F. J. P. (1985). Coming into social being: Cultural images of infants in Bimin-Kuskusmin folk psychology. In G. M. White & J. Kirkpatrick (Eds.), Person, self, and experience: Exploring Pacific ethnopsychologies (pp. 183-242). Berkeley: University of California Press. Rabain, J. (1979). L 'enfant du lineage: Du sevrage a la class d'dge chez les Wolof du Senegal. Paris: Payot. Rawls, J. (1955). Two concepts of rules. The Philosophical Review, 64, 3-32. Reece, S. (1993). The stranger's welcome: Oral theory and the aesthetics of the Homeric hospitality scene. Ann Arbor: University of Michigan Press. Reynolds, S. (1998). Energy flash: A journey through rave music and dance. London: Picador (Macmillan). Richards, A. I. (1932/1948). Hunger and work in a savage tribe: A functional study of nutrition among the southern Bantu. London: G. Routledge. Quotations from 1948 edition, Glencoe, IL: Free Press. Richerson, P., & Boyd, R. (2004). The nature of cultures. Chicago: University of Chicago Press. Rizzolatti, G., Fogassi, L., & Gallese, V. (2001). Neurophysiological mechanisms underlying the understanding and imitation of action. Nature Reviews Neuroscience, 2, 661-670. Rosaldo, R. I. (1968). Metaphors of hierarchy in a Mayan ritual. American Anthropologist, 70, 524-536. Rosenblatt, J. S. (1994). Psychobiology of maternal-behavior—Contribution to the clinical understanding of maternal-behavior among humans. Acta Paediatrica, 83, 3-8. Rothschild-Whitt, J. (1979). The collectivist organization: An alternative to rational-bureaucratic models. American Sociological Review, 44, 509-527. Rozin, P. (1999). Food is fundamental, fun, frightening, and far-reaching. Social Research, 66, 9-30. Rozin, P., & Nemeroff, C. (1994). The contagion concept in adult thinking in the United States: Transmission of germs and of interpersonal influence. Ethos, 22, 158-186. Rozin, P., Nemeroff, C., Wane, M., &Sherrod, A. (1989). Operation of the sympathetic magical law of contagion in interpersonal attitudes among Americans. Bulletin of the Psychonomic Society, 27, 367-370. Rushkoff, D. (2001). Ecstasy: Prescription for cultural renaissance. In J. Holland (Ed.), Ecstasy: The complete guide (pp. 350-357). Rochester, VT: Park Street Press. Russell, B. (1921). The analysis of mind. London: G. Allen & Unwin; New York: Macmillan. Ryle, G. (1949). The concept of mind. New York: Barnes & Noble. Sahlins, M. (1965). On the sociology of primitive exchange. In M. Banton (Ed.), The relevance of models for social anthropology. Association of Social Anthropologists, Monograph 1. London: Tavistock. Reprinted in Sahlins, M. (1972), Stone age economics. New York: Aldine. Sapir, E. (1949/1927). The unconscious patterning of behavior in society. In D. G. Mandelbaum (Ed.), Selected writings of Edward Sapir in language, culture, and personality (pp. 544-559). Berkeley: University of California Press.

144

FISKE

Schacter, D. L, Wagner, A. D., & Buckner, R. L. (2000). Memory systems of 1999. In M. Tooling & F. I. M. Clark (Eds.), The Oxford dictionary of memory (pp. 627-643). Oxford, England: Oxford University Press. Schieffelin, E. L. (1976). The sorrow of the lonely and the burning of the dancers. New York: St. Martin's Press. Schneider, D. M. (1980). American kinship: A cultural account (2nd ed.). Chicago: University of Chicago Press. Schutz, A. (1977/1951). Making music together: A study in social relationship. Social Research, 18, 76-97. Reprinted in J. L. Dolgin, D. S. Kemnitzer, & D. M. Schneider (Eds.), Symbolic anthropology: A reader in the study of symbols and meanings. New York: Columbia University Press. Schwartz, B. (1981). Vertical classification: A study in structuralism and the sociology of knowledge. Chicago: University of Chicago Press. Schwartz, B., Tesser, A., & Powell, E. (1982). Dominance cues in nonverbal behavior. Social Psychology Quarterly, 45, 114-120. Shweder, R. A., Jensen, L. A., & Goldstein, W. M. (1995). Who sleeps by whom revisited: A method for extracting the moral goods implicit in practice. In J. J. Goodenow, P. Miller, & F. Kessel (Eds.), Cultural practices as contexts for development (pp. 21-40). San Francisco: Jossey-Bass. Sieroszewski, W. (1896/1993). The Yakut: An experiment in ethnographic research. Moskva: Assotsiatsiia 'Rossiiskaia polit. entsiklopediia'. Human Relations Area Files translation, HRAF Computer File Ms.: 1044. New Haven, CT: HRAF. Simmel, G. (1900/1990). The philosophy of money (2nd enl. ed.). D. Frisby (Ed.), T. Bottomore &D. Frisby, Trans. London: Routledge. Simoons, F. J. (1994). Eat not this flesh: Food avoidances from prehistory to the present (2nd ed.). Madison: University of Wisconsin Press. Smith, W. R. (1885/1903). Kinship and marriage in early Arabia. London: Adam & Charles Black. Smith, W. R. (1889/1927). Lectures on the religion of the Semites: The fundamental institutions. Original ed. Edinburgh: Adam and Charles Black. 3rd ed. New York: KTAV. Sobal, J. (2000). Sociability and meals: Facilitation, commensality, and interaction. In H. L. Leiselman (Ed.), Dimensions of the meal: The science, culture, business and art of eating (pp. 119-133). Gaithersburg, MD: Aspen. Spencer, H. (1896). The principles of sociology (2nd ed., Vol. 2, Part 1). New York: D. Appleton. Spencer, P. (1985). The Maasai ofMatapato: A study of rituals of rebellion. International African Library. Bloomington: Indiana University Press. Sperber, D. (1994). The modularity of thought and the epidemiology of representations. In L. A. Hirschfeld & S. A. Gelman (Eds.), Mapping the mind: Domain specificity in cognition and culture (pp. 39-68). New York: Cambridge University Press. Sperber, D. (1996). Anthropology and psychology: Towards an epidemiology of representations. In D. Sperber, Explaining culture: A naturalistic approach (pp. 56-76). Oxford, England: Blackwell. Spiegel, J., & Machotka, P. (1974). Messages of the body. New York: Free Press; London: Collier Macmillan. Squire, L. R., & Knowlton, B. J. (1995). Memory, hippocampus, and brain systems. In M. S. Gazzaniga (Ed.), The cognitive neurosciences (pp. 825-838). Cambridge, MA: MIT Press. Strathern, A. (1973). Kinship, descent and locality: Some New Guinea examples. In J. Goody (Ed.), The character of kinship (pp. 21-33). Cambridge, England: Cambridge University Press. Strong, R. (1973). Splendor at court: Renaissance spectacle and the theater of power. Boston: Houghton Mifflin. Tajfel, H. (Ed.). (1978). Differentiation between social groups: Studies in the social psychology of intergroup relations. London: Academic Press. Takahashi, M., & Olaveson, T. (2003). Music, dance, and raving bodies: Raving as spirituality in the central Canadian rave scene. Journal of Ritual Studies, 17, 72-96. Tarde, G. (1900/1890). Les lois de limitation; Etude sociologique (3rd ed.). Paris: Felix Alcan.

3. CONSTITUTING RELATIONSHIPS

145

Tegnaeus, H. (1952). Blood brothers: An ethno-sociological study of the institutions of bloodbrotherhood with special reference to Africa. New York: Philosophical Library. Testart, A. (1987). Game sharing systems and kinship systems among hunter-gatherers. Man (new series), 22, 287-304. Toren, C. (1990). Making sense of hierarchy: Cognition as social process in Fiji. London School of Economics Monographs on Social Anthropology No. 61. London: Athlone Press. Trumbell, H. C. (1885). The blood covenant: A primitive rite and its bearing on scripture. New York: Scribner's. Tulving, E. (1995). Organization of memory: Quo vadis? In M. S. Gazzaniga (Ed.), The cognitive neurosciences (pp. 839-847). Cambridge, MA: MIT Press. Turner, T. S. (1980). The social skin. In J. Cherfas & R. Lewin (Eds.), Not work alone: A cross-cultural view of activities superfluous to survival (pp. 112-140). Beverly Hills, CA: Sage. Turner, V. (1967). The forest of symbols: Aspects ofNdembu ritual. Ithaca, NY: Cornell University Press. Turner, V. (1969). The ritual process: Structure and anti-structure. London: Routledge & Kegan Paul. Turner, V. (1973). The center out there: Pilgrim's goal. History of Religions, 12, 191-230. Turner, V. (1975). Revelation and divination in Ndembu ritual. Ithaca and London: Cornell University Press. Tyler, T. R., & Lind, E. A. (2001). Procedural justice. In J. Sanders & V. L. Hamilton (Eds.), Handbook of justice research in law (pp. 65-92). Dordrecht, Netherlands: Kluwer Academic Publishers. Tylor, E. B. (1871). Primitive culture: Researches into the development of mythology, philosophy, religion, art, and custom. London: J. Murray. Uvnas-Moberg, K. (1998). Oxytocin may mediate the benefits of positive social interaction and emotions. Psychoneuroendocrinology, 23, 819-835. van Gennep, A. (1908/1960). The rites of passage (M. B. Vizedom & G. L. Gaffe, Trans.). Chicago: University of Chicago Press. Vygotsky, L. S. (1934/1962). Thought and language. In E. Hanfmann & G. Vaker (Eds. and Trans.). Cambridge, MA: MIT Press; New York: Wiley. Watson, J. L. (1987). From the common pot: Feasting with equals in Chinese society. Anthropos, 82, 389-401. Watson, J. M., Campbell, K. J., & Collis, K. F. (1999). The structural development of the concept of fraction by young children. Journal of Structural Learning & Intelligent Systems, 13, 171-193. Wells, R. (1997). Divine secrets of the ya-ya sisterhood: A novel. New York: Harper Collins. Wetzel, P. J. (1993). The language of vertical relationships and linguistic analysis. Multilingua, 12, 387-406. White, G. M. (1985). Premises and purposes in a Solomon Islands ethnopsychology. In G. M. White & J. Kirkpatrick (Eds.), Person, self and experience: Exploring Pacific ethnopsychologies. Los Angeles: University of California Press. Whitehead, H. (2000). Food rules: Hunting, sharing, and tabooing game in Papua New Guinea. Ann Arbor: University of Michigan Press. Whitehouse, H. (2000). Arguments and icons: Divergent modes of religiosity. Oxford, England: Oxford University Press. Whiting, J. W. M. (1941). Becoming a Kwoma: Teaching and learning in a New Guinea tribe. New Haven, CT: Published for the Institute of Human Relations by Yale University Press. Whiting, J. W. M. (1961). Socialization process and personality. In F. L. K. Hsu (Ed.), Psychological anthropology. Homewood, IL: Dorsey Press. Whiting, J. W. M., Kluckhohn, R., & Anthony, A. (1958). The function of male initiation ceremonies at puberty. In E. E. Macoby, T. M. Newcomb, & E. L. Hartley (Eds.), Readings in social psychology (pp. 359-370). New York: Henry Holt.

146

FISKE

Wilson, M. L., Hauser, M. D., & Wrangham, R. W. (2001). Does participation in intergroup conflict depend on numerical assessment, range location, or rank for wild chimpanzees? Animal Behavior, 61, 1203-1216. Woodburn, J. (1982). Egalitarian societies. Man, New Series, 17, 431-451. Yamagata, N. (1994). Homeric morality. Leiden and New York: E. J. Brill. Young, F. W. (1965). Initiation ceremonies: A cross-cultural study of status dramatization. Indianapolis: Bobbs-Merrill. Young, L. J., Lim, M. M., Gingrich, B., & Insel, T. R. (2001). Cellular mechanisms of social attachment. Hormones and Behavior, 40, 133-138. Young, L. J., Nilsen, R., Waymire, K. G., MacGregor, G. R., & Insel, T. R. (1999). Increased affiliative response to vasopressin in mice expressing the via receptor from a monogamous vole. Nature, 400, 766-768. Zelizer, V. A. (1994). The social meaning of money: Pin money, paychecks, poor relief, and other currencies. New York: Basic Books. Zemplini-Rabain, J. (1973). Food and the strategy involved in learning fraternal exchange among Wolof children. In P. Alexandre (Ed.), French perspectives in African studies (pp. 221-233). London: Oxford University Press.

C H A P T E R

4 Social Expertise: Theory of Mind or Theory of Relationships? Nick Haslam Alan Page Fiske

In Alain de Botton's novel On Love (1993), the narrator and the woman he desires wander around a gallery and settle in front of a painting of Venus and Cupid: She turned to face the painting again, her hand once more brushing against mine.... Her action could have meant more or less anything; it was an empty space onto which one could decide to attach almost any intention from desire to innocence. Was this a piece of subtle symbolism ... that would one day allow me [like Cupid on the wall] to reach over and kiss her, or an innocent, unconscious spasm by a tired arm muscle? (p. 25)

The narrator's predicament is a familiar one. Only in comic strips are people furnished with thought balloons, and those of us who live in three dimensions are left to guess what others think, feel, and want, imagining our way into their subjective worlds with the aid of behavioral clues and folk psychology. Mercifully, other people's minds are not entirely inaccessible. Indeed, humans appear to be equipped with specialized abilities to read other minds, imputing mental states such as beliefs, wishes, intentions, and feelings and using these inferences to shape interpersonal behavior. These abilities have emerged as an active domain of psychological inquiry. Under the rubric of "theory of mind" (Premack & Woodruff, 1978), this work examines the capacity of humans and other primates to understand their peers mentalistically, as individuals with potentially knowable 147

I 48

HASLAM AND FISKE

and manipulable mental states. Developmental psychologists have sought to understand the steps through which children pass toward a fully mentalistic understanding of others, charting a progression from joint attention, to pretend play, understanding others' desires and then beliefs, and ultimately deceit and irony (e.g., Wellman, 1990). Comparative psychologists try to discover how far along this path different species progress in their social understanding (e.g., Whiten, 1990). Finally, theory of mind is the focus of efforts to comprehend some forms of psychopathology, with research suggesting that autism may involve a specific deficit in imputing mental states (e.g., Baron-Cohen, 1995). The theory of mind approach to social cognition—the view that our social expertise is fundamentally based on mindreading—has yet to penetrate the mainstream study of human relationships. However, it is a perspective that aspires to encompass the cognitive basis of human sociality, and its proponents make strong claims on its behalf. "Theory of mind," argues O'Connell (1998, p. 206), "allows us to establish and handle highly complex social relationships." "Understanding minds permits sophisticated social co-ordination," claims Mitchell (1997, p. 54), "which in turn gives rise to society," and furthermore "human culture could be the product of our passion for contemplating psychological matters" (p. 12). In this chapter we argue that the theory of mind approach has important limitations as an account of the cognitive underpinnings of sociality and relationships. Although it represents an exciting research frontier, we claim that it presents a partial and in some ways misleading view of social cognition. It also offers an extremely instrumental vision of human relations that is at times disturbing, proposing that, in essence, "we are primates who are experts in deceit, double-dealing, lying, cheating, conniving and concealing" (O'Connell, 1998, p. 155). Our critique focuses on four lines of argument. First, we argue that the theory of mind perspective, by attending to the mental content of individuals, neglects the relational dimension of social cognition which concerns the interpersonal linkages and networks in which individuals are embedded. Our research shows that relational cognition underlies a wide variety of social-cognitive activities and is irreducible to cognition about individuals, their attributes, or their mental states. Second, we review anthropological evidence that, in contrast to the industrialized and psychology-saturated West, mindreading and intersubjective attribution is less central to social life in many other cultures. We argue that the model of sociality that the theory of mind perspective presents is similarly culture-bound in its individualistic representation of the person as a selfish, Machiavellian maximizer of personal opportunities. Third, we review evidence about the social capabilities of other species that lack a well-developed theory of mind; we show that a great deal of complex social organization is mediated by so-

4. SOCIAL EXPERTISE

I 49

phisticated social-relational cognition that apparently does not involve mindreading. Fourth, we discuss the theory of mind analysis of autism, arguing that it has limitations both as a theory of the condition and as an illustration of what social relations would be like in the absence of mindreading. We acknowledge the place of mindreading in social cognition, and do not question the crucial role of some of its cognitive building blocks and precursors, such as joint attention and the understanding of others as intentional, goal-directed agents (Tomasello, 1999). Nevertheless, we conclude that the role of mindreading is more limited than is often thought and that it is not the fundamental cognitive basis for social life. Theory of mind is not required for the enactment of complex social relationships, its role in human social interaction is culturally circumscribed, and its proponents assume questionably that people are fundamentally guileful, selfish, and extrinsically motivated. Being a social creature involves knowing what is in others' heads less than it involves knowing the nature of our relationships with them.

RELATIONAL COGNITION

A theory of mind is a representation of the mental worlds of other individuals. Mindreading, in short, is a matter of imagining the contents of other minds. This approach to social cognition therefore belongs to an essentially individualistic view of social relations, in which the attributes that matter for social interaction are those that belong to individual persons. In theory of mind these individual attributes are mental states, whereas in other research traditions within social cognition, such as social categorization, stereotyping, and impression formation, they are demographic characteristics or personality traits. Individual characteristics such as these are clearly important ways of classifying people and making social inferences that influence how we think about and interact with others. However, people must also represent, attend to, and think about the properties of the relationships between individuals. Social cognition is individual and relational, concerned with both the nodes and the links in the social network. Imagine, for example, that I am a colleague of yours. In my everyday dealings with you, I will probably take into account (whether explicitly or implicitly) your gender, your age, your personality, and your mental states such as your moods, your beliefs about my competence, and your desire to succeed on a particular project. I will also be concerned with the nature of my relationship with you: its warmth, any differences in rank, our work roles vis-a-vis one another, any favors I owe or am owed, insults and offenses committed or redressed, and so on. I will also be concerned with the nature of your relationships with other co-workers: your distressing close-

I 50

HASLAM AND FISKE

ness to our boss, the friendlier reception you receive from the secretaries. In order to conduct our social lives, we must be keenly attuned to the state of our relationships and those of others, and must have sophisticated ways of representing, evaluating, and thinking about these relationships. This point should be obvious, but it bears repeating in view of the individualist assumptions of mainstream social cognition research (Fiske & Haslam, 1996). These assumptions have delayed the growing recognition that there is a distinctly relational layer of social cognition that is vital for social interaction and social competence (e.g., Baldwin, 1992; Jackendoff, 1992). Along with our abilities to guess what others are thinking and feeling and our beliefs about the kinds of people we may encounter (i.e., demographic classifications and personality traits), we come to our interactions with schemas and models for making sense of, conducting, coordinating, and evaluating relationships. Relational models (RMs) theory is one account of these cognitive underpinnings of relationships, proposing a set of abstract schemas, frames or "grammars" for social cognition and cultural learning (for reviews of the theory and its research base see chaps. 1 and 2). By this account, a fundamental part of each person's social-cognitive competence consists of the four RMs together with detailed knowledge of their culture's implementation of those models. All of these aspects of social or relational expertise are expressible in terms of social rules, principles, and conventions that concern how relationships should be conducted, and therefore do not demand any kind of guesswork about the mental states of particular interactants. Importantly, amidst the extensive support for the RMs' capacity to account for social cognition and behavior, several studies (Fiske, 1995; Fiske & Haslam, 1996, 1998; Fiske, Haslam, & Fiske, 1991) have found that the models organize everyday social cognition more strongly than, and independently of, individual attributes. For example, when people make natural errors of person memory, they tend to confuse people of the same gender and race, of similar age and personality, and with whom they have the same kind of relationship; but the last, relational effect is commonly the most powerful. What do all of these findings imply about the role of theory of mind in social cognition? First, they show that mindreading, the imputation of mental states to individuals, is only one part of the social-cognitive apparatus that underlies social expertise, and is not obviously pre-eminent among the different parts. Shared understandings of relationships, represented in terms of the RMs and their implementation parameters and paradigms, are distinct and demonstrably important in everyday social cognition. Second, theory of mind represents a form of the individualist approach to social cognition—a way of attending to the characteristics of persons rather than to the relationships between them—which has not fared well in comparisons of so-

4. SOCIAL EXPERTISE

151

cial-cognitive influence with the relational approach. Third, the findings indicate that the relational approach to social cognition, as represented by the RMs, is readily translatable across cultures, a question that we now raise with respect to the theory of mind approach.

CULTURE AND THEORY OF MIND

The evidence reviewed earlier indicates that theory of mind is only one kind of cognitive competence that underpins our social relations, and that making inferences about the contents of other minds may be less fundamental than having shared, generative models for constructing and making sense of relationships. However, in addition to questioning the general role of theory of mind in social relationships, we can also ask whether this role is equally prominent throughout the world. Is theory of mind a fundamental, universal part of social cognition in relationships, or is it circumscribed by culture, implying a more delimited view of its importance? Lillard (1997, 1998) has carried out a searching review of the anthropological evidence pertinent to this question, and concludes that theory of mind is highly variant in its cultural manifestations and importance. She shows that folk psychologies—implicit understandings of the structure and workings of the mind—differ widely across cultures, and influence the nature and reach of everyday mindreading. Moreover, she argues that it is primarily in the industrialized West that people hold elaborately mentalistic understandings of social activity. For instance, Westerners are especially prone to attribute behavior to internal, psychological causes and dispositions, prompting them to employ theory of mind in making sense of it. Modern Westerners take mental states to be mediators of the link between the outside world and behavioral responses, and explain emotions as products of mental, rather than physical or social, processes. They have richly differentiated understandings of mental states, indicating the cultural importance ascribed to psychological matters. They care deeply about whether behavior is intentional or unintentional, and use this distinction in weighing responsibility, punishment, and praise. People in non-Western cultures provide an often stark contrast to this understanding of the mind (see Lillard, 1998, for a rich assortment of ethnographic illustrations). They are generally more likely to attribute behavior to situational determinants, social context and influence, cultural prescriptions, and causes that in the view of Western modernity are occult (cf. Lee, Hallahan, & Herzog, 1996; Miller, 1984; Morris, Nisbett, & Peng, 1995; Shweder & Bourne, 1984; see also the review in Fiske, Kitayama, Markus, & Nisbett, 1998). They may not sharply distinguish mental from bodily processes, or may do so in ways that violate Western folk psychology. They are

I 52

HASLAM AND FISKE

more likely to view the minds of others as essentially private and unknowable, to find speculation on their contents improper, and to have relatively restricted vocabularies for psychological states. They tend to judge culpability on the basis of the consequences of the person's actions, rather than on their intentions or motivations. Responsibility is less likely to be individualized, and more likely to be allocated to social groups and causes. In most cultures, people frame moral judgments about each other primarily in terms of relational obligations and social responsibilities (see Fiske, 1990; Fiske et al., 1998). A quote from Lillard (1998) concerning the Illongots of the Philippines nicely captures the gist of this collection of differences between Western and non-Western models of mind and behavior: "What is important for Illongots is not what goes on in the rinawa [a term that corresponds loosely to mind but is located in the heart and implies general life force] but rather what happens between people. The focus is not on a world of discrete selves containing mental worlds but on relationships" (p. 12). In short, the theory of mind approach to social cognition assumes a kind of concern with inferring and differentiating the psychological states of individual minds that is de-emphasized in a large proportion of humanity. Although mindreading capacities are surely at some level universal (Avis & Harris, 1991), the role that we allocate to it in human social life must be correspondingly limited if we are to avoid mistaking a "hypercognized" aspect of Western culture for a universal truth about social cognition (Fiske et al., 1998). We should at the same time increase our appreciation of the underrecognized role of relational considerations in social cognition. The theory of mind approach overemphasizes individualistic attribution processes that are prominent in few cultures and may not be the primary mechanisms of social cognition in any culture. Furthermore, this approach arguably misrepresents the nature of social expertise and the motives that drive it. Theory of mind researchers present social interaction as a frequently antagonistic process that is riven with scheming and deceit. Emphasis is placed on the benefits that theory of mind yields for political intrigue, with the primary goal of mindreading seen as the manipulation and deception of others in the service of enhancing the manipulator's social dominance. Interactants are often pictured as Machiavellians (e.g., Byrne & Whiten, 1988) who are constantly alert to the possibility of being cheated and who strive for personal or genetic (i.e., inclusive fitness) advantage. Although mindreading is also occasionally said to enable and promote more positive ends, such as empathy, cooperation, communication, and moral sensitivity, it is primarily theorized as a cognitive adaptation for securing benefits through clever tactical competition with others. Although it is clear that social intelligence is often employed in deceptive and self-seeking ways, it is questionable whether this hard-Darwinian

4. SOCIAL EXPERTISE

153

view of social relations provides an accurate representation of the cognitive and motivational processes that underlie social cognition. The theory of mind position, and the broader evolutionary framework in which it is embedded, makes four basic assumptions about these processes: They are based on calculations of benefit and cost; the calculations aim to maximize the benefits and minimize the costs; the relevant benefits are to the individual or to close genetic relatives; and the benefits are generally to be understood as extrinsic to the interaction and essentially nonsocial (e.g., reproductive opportunities, food, adaptive fitness). Several observations must be made about this view of social interaction as driven by selfish maximization of benefits. Even if it were granted that it correctly describes at an evolutionary level of analysis the process that ultimately drives social relations, it is quite implausible that it describes the cognitive processes that govern the ongoing social relations of the creatures involved. It is very unlikely that organisms engaged in social interaction are performing cost-benefit analyses of the opportunities available to them and selecting social tactics to maximize the relevant ratios. Instead, they follow simpler rules of thumb—heuristic decision processes—which may have been selected over evolutionary time to yield more or less optimal results on average. No nonhuman animal seems capable of the kind of reasoning with proportions and the distributive law (combining addition with multiplication) that is required for cost-benefit calculations. On the other hand, there is copious evidence of relational forms of social cognition in all social animals (Haslam, 1997). Hence it is invalid to attribute nonhuman social interaction to online maximizing calculation. Even in the case of humans, utility calculation (let alone selfish maximization) is only one rather restricted form of social cognition in relationships. Indeed, it is best seen as one implementation of the Market Pricing (MP) model, a model based on reasoning in terms of rates and proportions that appears to be distinctively human (Haslam, 1997). Most human relationships are governed by models that are do not involve any explicit cost-benefit calculation or any other online analysis of the expected net utility of tactical alternatives. For instance, as RMs theory suggests, people cognize some relationships in terms of balanced egalitarian exchange, some in terms of asymmetries of rank and precedence, and some in terms of undifferentiated in-groups. Moreover, people generally seek social relationships not as means to extrinsic, nonsocial ends, but for their own sake, as engagements that are intrinsically motivated and meaningful (Fiske, 1991). These intrinsically relational, motivated social proclivities presumably evolved because of their selective advantages for inclusive fitness. But human fitness is not typically maximized by a short-term exploitive selfishness. If social cognition in human relationships is not based exclusively or even typically on the cost-benefit calculating and selfish advantage-seeking

I 54

HASLAM AND FISKE

model that is implied by many proponents of the theory of mind perspective on social cognition, then what accounts for this view? Fiske et al. (1998) suggest that MP (especially implemented as self-interested cost-benefit calculation and instrumental, economic rationality) is a highly developed and favored cultural ideology in the West—one that is commonly reflected in Western understandings of decision making and choice. By this account, a cultural orientation that is unusually prominent in Western capitalism, with its emphases on economic rationality and individualism, is reflected in many Western folk and academic theories of human psychology and social life. In particular, the theory of mind perspective represents social interaction as the sophisticated bargaining of individual agents striving for the best deal. Although this may indeed characterize some interactive strategies, especially when viewed in terms of adaptive advantages of relational structures, it does not characterize everyday, real-time cognition or motivation in most social relationships. In most of the world, where the MP model is neither culturally salient nor positively valued, this kind of instrumental, exploitive tactical manipulation is not the norm, and is often maladaptive.

SOCIAL COGNITION IN OTHER SPECIES

Humans take part in social arrangements of unparalleled complexity and are capable of subtle mindreading. Does this correlation, which is clearly no accident, mean that sophisticated mindreading abilities are necessary for complex social relations? If complex social arrangements operate in the absence of theory of mind skills, then the cognitive foundations of social relationships must be sought elsewhere. This would imply that mindreading abilities are not essential to complex social cognition. This question brings us into the field of comparative psychology and ethology, where the social intelligence, social organization, and theory of mind capabilities of many species have been extensively studied in recent years (e.g., Povinelli & Preuss, 1995; Tomasello & Call, 1997). The evidence with respect to theory of mind is fairly clear at this point. Chimpanzees and some other great apes appear to have a very limited degree of mentalistic understanding of their peers. Some chimps have demonstrated capacities for simple deception under special conditions, such as deliberately concealing information from the view of competitors or actively distracting them. However, often apes only exhibit deceptive capacities after prolonged laboratory training or language training. Spontaneous deception is rather rare under natural conditions. Chimps and some other apes such as gorillas and orangutans generally show the capacity for self-recognition in mirrors, a low-level theory of mind skill involving basic self-awareness and hence, it is argued, the ability to represent one's own mental states. However, any level

4. SOCIAL EXPERTISE

I 55

of a more sophisticated capacity such as empathy has been claimed for chimpanzees only, and the evidence on this matter is controversial. In sum, theory of mind is quite limited in nonhuman species; it is restricted to a small subset of higher primates who display it inconsistently in very limited ways. The mindreading abilities of the most capable nonhuman species is, at most, probably equivalent to that of an average human 3-year-old. Do the severe limitations of other species' theory of mind capabilities indicate that the social-cognitive skills that underpin their social relationships are similarly restricted? We would argue forcefully that they do not. Many species that show little or no evidence of even the most simple theory of mind nonetheless lead sophisticated social lives of considerable organizational complexity, complete with dominance hierarchies, shifting alliances, enduring coalitions, incipient moralities, and involved kinship structures (e.g., Harcourt & de Waal, 1992; Quiatt & Reynolds, 1993). These arrangements are particularly developed in some species of monkey (e.g., Cheney & Seyfarth, 1990), which typically do very poorly on theory of mind tests. It is important to recognize that these social arrangements do not simply self-assemble or exist free of cognitive supports, but manifest an often astute political intelligence. All of them involve patterned relationships among individuals who must have a mentally represented repertoire of relationship types and interaction rules, a capacity to recognize other individuals and attribute to them distinct relational roles (e.g., subordinate, kingroup member, alliance partner), and a capacity to keep track of the state of particular relationships (e.g., one's position in a hierarchy being at risk, the need to retaliate against a member of a rival kin-group). None of these social-cognitive capabilities require any inference about states of mind, but together they yield complex social organization and sophisticated interaction patterns. In a recent review of the literature on primate social relations and the cognitive capabilities that support them, Haslam (1997) argued that versions of three of Fiske's RMs are in evidence among many primate species, including those whose theories of mind are primitive or nonexistent. Monkeys commonly show not only the exclusive favoritism of Communal Sharing relationships in their relationships with kin and the transitive figuring of Authority Ranking in their dominance hierarchies, but also, to some degree, the capacity to conduct proto-egalitarian relationships (rudimentary Equality Matching), tracking favors and obligations and carrying out tit-for-tat retaliation. Only the proportion-based calculus of MP appears to be beyond the cognitive grasp of nonhuman primates. In short, many nonhuman primates lack a theory of mind, but do bring to the conduct of their social relationships a set of social-cognitive capabilities that appear homologous to human relational cognition. These relational capacities yield a high level of social expertise, social problem solving, and organizational complexity.

I 56

HASLAM AND FISKE

Of course, this does not mean that theory of mind is irrelevant to social cognition in human relationships, but it does demonstrate that sophisticated relational capabilities exist beyond and before those that theory of mind bestows. It should also make us pause and consider whether mindreading is truly fundamental to the conduct of our everyday relationships. It seems more likely that mindreading is superimposed on a more basic substructure of social grammars, principles, and shared understandings. Having said this, we believe that the relational approach can build on, extend, and be integrated with the mindreading approach. This integration of the two approaches should be based on revising the theory of mind position to incorporate the following considerations: 1. A Hobbesian war of all against all is not to anyone's advantage; neither is social isolation. Short-term selfish exploitation of others generally leads to ostracism, exile, or violent retaliation. Furthermore, humans cannot live on their own; there is no human society in which people do not consistently participate in groups and dyadic relationships to which they feel strongly bound. From an evolutionary point of view, people's expected inclusive fitness is not maximized by blatantly selfish individualism, but by a high degree of sociality, and it is generally advantageous for them to participate "sincerely" and "honestly" in certain fundamental types of social relationships. Individual fitness is enhanced by strong motivations to participate in, sustain, and make moral commitments to these relationships. 2. It is advantageous for individuals to be adept at operating in a world of morally loaded social relationships: People need to be able to anticipate, perceive, and influence each other's actions, feelings, and thoughts. To do so, however, people need to think in terms of the RMs that they are all using to plan, generate, coordinate, interpret, contest, judge, and sanction their interaction. People may manipulate their relationships, but the relationships themselves are essential to survival and reproduction, not to mention psychological well-being. So, although people may pursue advantageous strategies within the framework of any social relationship, it is very rare that it is worth sacrificing crucial relationships for any extrinsic end. In a small, closed community, good will, trust, and affection in key relationships are essential for virtually all other needs and goals. 3. Hence the mindreading that people must do to function in society is primarily the reading of each other's social intentions, desires, and moral judgments. People have to anticipate (and acknowledge) when someone will think it is their turn, reciprocate a favor, or expect an equal share—and when they will revenge an offense an-eye-for-an-eye, a-tooth-for-a-tooth. People have to understand who is a "superior" expecting deference, respect, and precedence—and whether this superior will offer generous largesse, protect, and stand up for their followers. In addition, people have to know when a su-

4. SOCIAL EXPERTISE

157

perior will punish a disobedient or disrespectful subordinate. People have to know when people will expect and require specific kinds of sharing, pooling of labor, or altruistic defense—and they have to recognize what transgressions will cause a person to be cast out. 4. Much of this thinking is "automatic" and unreflective. Understanding of the four fundamental RMs is innate, as are the motives and fears that induce people to join and conduct adaptive social relationships. People generate speech and understand each other without any articulate understanding of the syntactical principles they are using; likewise, people construct, coordinate, and contest social relationships without an articulate understanding of the RMs they are using. This means that the basic substrate of everyday, ongoing relational thinking is not self-conscious. People can articulate problematic issues and dispute with each other, but everyday life is based on takenfor-granted principles and processes that operate without much awareness most of the time. In this sense, social cognition is not really an explicit "theory" of mind: It is more like an implicit "grammar" of relations. 5. All social relationships are culturally implemented. Indeed, all social action is culturally mediated. Consequently, any theory or grammar of mind, including social mindreading, is a cultural grammar. Again, the language analogy is apt: People have innate proclivities that enable them to learn to speak, but the language each person speaks depends entirely on the community in which they learn to speak and must converse. Universal grammar potentiates the acquisition of a particular language: Speech is possible only in a particular language. Similarly, although human sociality is built out of universal models, what people actually need to know (implicitly) about each other's sociality involves cultural-specific parameters, prototypes, and paradigms. Is there an honor complex here, which means that her boyfriend will kill me if I flirt with her? What do I have to do to obtain a bride here? Now they are offering me their daughter in marriage, but what are they going to expect in return? I've killed a kudu; who expects a share? If I want to obtain land to farm here, what do I have to do? If I admire their baby will they be flattered or will they accuse me of the evil eye? What does this smile convey: intense humiliation? What do those epithets, sexual insults, and attacks with lighted brands imply: a ritualized joking relationship? To understand and get along with each other, we need to know the culturally mediated forms of universal relationships. A generic, culture-free psychological "theory" of mind is insufficient. AUTISM, PSYCHOSOCIAL DISORDERS, AND THEORY OF MIND-LESSNESS One of the primary contributions of the theory of mind perspective has been its account of autism, a serious disorder emerging in early childhood that is marked by often severe impairment of nonverbal and verbal commu-

I 58

HASLAM AND FISKE

nication, social and emotional reciprocity, and imaginative or pretend play. In an extensive body of experimental work (see Baron-Cohen, 1995, for a review), researchers have shown that autistic children have deficits on an assortment of theory of mind tasks, even taking into account their usual generalized intellectual impairment. Consequently, it has been argued that autism, and milder related conditions such as Asperger's syndrome, reflect neurological dysfunctions in brain structures that subserve mental state attribution. Autistic people are, by this popular account, "mindblind." A variation of this intriguing account has recently been applied to the case of certain symptoms of schizophrenia (Frith & Corcoran, 1996). This view of autism as mindblindness is important not just as a theory of the disorder, but also, some propose, as a demonstration of the centrality of mindreading to ordinary social expertise. To Bogdan (1997), for example, "autistic people play an important role in the evolutionary story of interpretation [i.e., social cognition] because, being interpretively abnormal, they provide a good control group against which to test . . . hypotheses about normal interpretation" (p. 165). Autism has rhetorical value for the theory of mind perspective because its deep social debilitation supposedly reveals the crucial role that mindreading plays in the normal case. There is no question that autistic people typically suffer in part from an inability to understand other people's thoughts and feelings, which makes ordinary social interaction difficult, threatening, and often mystifying for them. However, we must be careful in the inferences we draw from this fact. It is especially important to avoid inferring that without theory of mind capabilities social relationships are inevitably deformed and robotic, or that autism is what results when theory of mind is subtracted from normal sociality. There are three reasons why we should not make these superficially sensible inferences, which would imply that theory of mind is absolutely fundamental to human social cognition. First, autistic people have many additional impairments besides their mindblindness, often including chromosomal abnormalities causing mental retardation, and abnormalities in neurotransmitter and hormonal levels, autonomic responses, and brainwave patterns. They manifest brain abnormalities that extend far beyond those structures normally associated with social cognition and theory of mind, including the cerebellum, corpus callosum, hippocampus, basal ganglia, and even cranial nerves (Schulkin, 2000). Importantly, they also show psychological abnormalities including deficits in the perception of facial and vocal emotion (Hobson, Ouston, & Lee, 1989), elementary social skills that do not require mindreading. Moreover, the socioemotional deviancy of autistic children emerges well before the age of four, when mindreading abilities normally make their first appearance in normal children (Klin, Volkmar, & Sparrow, 1992), and there is even evidence of defective sensori-

4. SOCIAL EXPERTISE

159

motor (and basal ganglia) function in autistic neonates (Teitelbaum et al., 1998). Second, a whole class of autistic symptoms involving repetitive and stereotyped behavior patterns, such as odd preoccupations and physical mannerisms, are difficult to comprehend in terms of a theory of mind deficit. This has prompted some to challenge the mindblindness theory and substitute one that invokes impairment in the brain's more general executive functions (Russell, 1998), a theory consistent with the frontal lobe aberrations obtained in autistic people (Ozonoff, Pennington, & Rogers, 1991). Third, the disturbances in interpersonal relations that are characteristic of autistic adults do not provide a good model of what ordinary social interaction would be like if people could not mindread. After all, autistic adults are not simply people who lack certain innate capacities; they are people who have suffered persistent and pervasive deficits in social learning and development, resulting from their own and others' social avoidance and segregation. They not only lack mindreading abilities, but they also lack a competent command of essential social rules, scripts and conventions. Consequently, their interpersonal difficulties do not directly or clearly indicate the importance of theory of mind in everyday relationships, but may result in part from other social-cognitive deficits. To summarize our observations on autism, there are many reasons besides mindreading deficits why autistic people might show impaired social competence, and why we should not take their impairment as naturalistic evidence of how social relationships would be in the absence of mindreading. It is interesting to contemplate what it would be like to interact with someone of normal intelligence, executive functioning, and social development whose theory of mind had been surgically obliterated (if that were anatomically possible), but it is quite unlikely that it would resemble interacting with an autistic person. We would argue that such a person, despite having some interpersonal peculiarity, would retain considerable social expertise, because many of the social-cognitive capacities that underpin relationships would be left intact. On the other hand, a person who lacked any implicit understanding of the RMs would be unable to function in society. Such a person would be unable to perceive or experience social emotions such as love, belonging, pride, embarrassment, shame, or guilt; form a social identity or recognize social roles; understand sanctions; assert rights; make redress or retaliate in a socially meaningful manner; work with others; exchange; buy or sell or calculate value; make a social distribution; promise; take turns, or conduct any other basic social intercourse. Whether this corresponds to autism or any other actual psychological disorder is still unclear, but it makes evident the problems of taking theory of mind as the obligatory centerpiece of social expertise.

I 60

HASLAM AND FISKE

CONCLUSIONS

We have argued that theory of mind has a variety of serious limitations as a way of accounting for social cognition in human social relations. Mindreading is only one part of an assemblage of social-cognitive capabilities, and a part whose universality and indispensability for everyday social life have been overstated. However, we do not mean to deny the importance of the theory of mind repertoire or the new order of complexity that it introduces into human social interaction and social perception. Without question, in the absence of mindreading skills and the forms of intersubjectivity that they allow, our interpersonal lives would be distinctly different. What is required is not a demolition of the theory of mind view of social expertise, but an assignment of it to its proper place and a reining in of some of its ambitions to be a comprehensive account of social cognition. One way to restore theory of mind to its proper place in social cognition is to recognize its role alongside other capabilities, including RMs and schemas, moral systems, emotion recognition skills, interaction scripts, social categorization, and so on. In this view, theory of mind is one of several distinctly social components of our cognitive apparatus. Social cognition should therefore not be equated with mindreading, as it often has been, for example, in the burgeoning neuroimaging literature, which has almost exclusively focused on theory of mind tasks (Fletcher et al., 1995; Gallagher et al., 2000). Like the ingredients of a good recipe, all of the components of social cognition must be functionally present for the combination to integrate properly, so that our lives with others proceed smoothly. A less ecumenical way to give theory of mind its proper due is to expand its meaning to include some of these additional capabilities. In this spirit, Bogdan (1997) equates theory of mind with folk psychology and dubs it "interpretation," a capacity that includes the attribution of dispositions, traits and emotions and the figuring out of one's position in dominance hierarchies, none of which strictly require mentalistic understanding. Neither of these ways of reformulating the place of theory of mind in social cognition appeals to us. The former is unhelpfully even-handed, refusing any judgment of what capabilities are primary. The latter expands the meaning of theory of mind beyond recognition, although it is in some ways congenial to the relational account that we favor, proposing that interpretation is "a pragmatic policy of tracking the subject's [i.e., other's] relations to the world to the extent that they affect the interpreter" (Bogdan, 1997, p. 7) rather than a reader of minds. Instead, we propose that the cognitive models that organize social relationships are fundamental to social expertise, that the other social-cognitive capabilities essentially serve and enable them, and that relational cognition is substantially distinct from theory of mind rather than both composing a single interpretive capacity. What peo-

4. SOCIAL EXPERTISE

161

pie need, as social beings, is to have coordinated social relationships with others, whether these involve shared work activities, reproduction and child-rearing, defense of a territory, education of the young, group identities, and all of the myriad forms of social engagement that we seek intrinsically. For people to implement these relationships in everyday interaction they must recognize emotional expressions, respect and care about shared moral and conventional prohibitions, infer others' mental states, and so on. However, these capabilities are all in the service of the basic requirement of coordinated relatedness. Moreover, our mental state inferences are focused on other people's social beliefs and desires, and they are usually motivated not by the urge to seek personal advantage but by the urge to coordinate with others to conduct meaningful relationships. Let us return to the perplexed narrator who appeared at the beginning of this chapter. Why is he engaged in intense speculation about his beloved's feelings and intentions toward him? Why does he attend closely to her face and her vocal inflections for evidence of her emotional state? Why is he careful not to violate social conventions governing physical contact and self-disclosure with nonintimate friends? The answer, surely, is that he wants to find out what kind of relationship he is to have with the woman, desires a particular kind of close communal relationship with her, and is anxious to realize this possibility. His mindreading, emotion perception, and mindfulness of social conventions are all in the service of his relational motives and models, his goal of establishing a particular kind of relationship using the rich assortment of social-cognitive skills that are at the disposal of modern humans. In short, we would argue that theory of mind is a recently evolved auxiliary to the more fundamental task of relational cognition. Theory of mind offers a fresh and revealing perspective on the cognitive capacities that underlie our lives with others. Behind many of our everyday transactions, thoughts and desires there is a rustle of mindreading, an active process of mulling over other people's beliefs, wishes and feelings. Social life is intersubjective as well as interactive, so that figuring out what is inside the heads and hearts of others is often important. However, there is much more to social cognition than imputing mental states to other people; in our interpersonal lives we must not only read minds but read relationships. To fail to recognize this is to make several common errors of academic psychology: to focus on individuals rather than relationships between them, to neglect the points of view of the broader social sciences, and to mistake cultural phenomena that are particular to Western modernity—the cultural matrix responsible for novels and for organized psychology itself—for universal truths of social life. Certainly we humans are creatures who are deeply concerned about and sensitive to one another's mental states, and who sometimes use our mindreading capabilities to manipulate and take advan-

I 62

HASLAM AND FISKE

tage of others. We are also creatures who seek relationships for their own sake, and who are cognitively equipped with shared relational grammars to take part in complex webs of cooperative social life with or without psychological speculation.

REFERENCES Avis, J., & Harris, P. (1991). Belief-desire reasoning among Baka children: Evidence for a universal conception of mind. Child Development, 62, 460-467. Baldwin, M. W. (1992). Relational schemas and the processing of social information. Psychological Bulletin, 112, 461-484. Baron-Cohen, S. (1995). Mindblindness: An essay on autism and theory of mind. London: MIT Press. Bogdan, R. J. (1997). Interpreting minds: The evolution of a practice. Cambridge, MA: MIT Press. Byrne, R. W., & Whiten, A. (Eds.). (1988). Machiavellian intelligence: Social expertise and the evolution of intellect in monkeys, apes and humans. Oxford, England: Oxford University Press. Cheney, D., & Seyfarth, R. (1990). How monkeys see the world: Inside the mind of another species. Chicago: University of Chicago Press. de Botton, A. (1993). On love. New York: Atlantic Monthly Press. Fiske, A. P. (1990). Relativity within Moose ("Mossi") culture: Four incommensurable models for social relationships. Ethos, 18, 180-204. Fiske, A. P. (1991). The cultural relativity of selfish individualism: Anthropological evidence that humans are intrinsically sociable. In M. Clark (Ed.), Review of personality and social psychology: Vol. 12. Altruism and prosocial behavior (pp. 176-214). Beverly Hills, CA: Sage. Fiske, A. P. (1995). Social schemata for remembering people: Relationships and person attributes that affect clustering in free recall of acquaintances. Journal of Quantitative Anthropology, 5, 305-324. Fiske, A. P., & Haslam, N. (1996). Social cognition is thinking about relationships. Current Directions in Psychological Science, 5, 143-148. Fiske, A. P., & Haslam, N. (1998). The structure of social substitutions: A test of relational models theory. European Journal of Social Psychology, 25, 725-729. Fiske, A. P., Kitayama, S., Markus, H. R., & Nisbett, R. E. (1998). The cultural matrix of social psychology. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (pp. 915-981). Boston: McGraw-Hill. Fletcher, P. C., Happe, F., Frith, U., Baker, S. C., Dolan, R. J., Frackowiak, R. S., & Frith, C. D. (1995). Other minds in the brain: A functional imaging study of "theory of mind" in story comprehension. Cognition, 57, 109-128. Frith, C., & Corcoran, R. (1996). Exploring 'theory of mind' in people with schizophrenia. Psychological Medicine, 26, 521-530. Gallagher, H. L, Happe, F., Brunswick, N., Fletcher, P. C., Frith, U., & Frith, C. D. (2000). Reading the mind in cartoons and stories: An fMRI study of "theory of mind" in verbal and nonverbal tasks. Neuropsychologia, 38, 11-21. Harcourt, A. H., &de Waal, F. B. M. (Eds.). (1992). Coalitions and alliances in humans and other animals. Oxford, England: Oxford University Press. Haslam, N. (1997). Four grammars for primate social relations. In J. Simpson & D. Kenrick (Eds.), Evolutionary social psychology (pp. 297-316). Mahwah, NJ: Lawrence Erlbaum Associates. Hobson, R. P., Ouston, J., & Lee, A. (1989). Naming emotion in faces and voices: Abilities and disabilities in autism and mental retardation. British Journal of Developmental Psychology, 7, 237-250.

4. SOCIAL EXPERTISE

I 63

Jackendoff, R. (1992). Is there a faculty of social cognition? In Languages of the mind: Essays on mental representation (pp. 69-81). Cambridge, MA: MIT Press. Klin, A., Volkmar, F. R., & Sparrow, S. S. (1992). Autistic social dysfunction: Some limitations of the theory of mind hypothesis. Journal of Child Psychology and Psychiatry, 33, 861-876. Lee, F., Hallahan, M., & Herzog, T. (1996). Explaining real life events: How culture and domain shape attributions. Personality and Social Psychology Bulletin, 22, 732-741. Lillard, A. S. (1997). Other folks' theories of mind and behavior. Psychological Science, 8, 268-274. Lillard, A. S. (1998). Ethnopsychologies: Cultural variations in theory of mind. Psychological Bulletin, 123, 3-32. Miller, J. G. (1984). Culture and the development of everyday social explanation. Journal of Personality and Social Psychology, 46, 961-978. Mitchell, P. (1997). Introduction to theory of mind: Children, autism and apes. London: Arnold. Morris, M., Nisbett, R. E., & Peng, K. (1995). Causal understanding across domains and cultures. In D. Sperber, D. Premack, & A. J. Premack (Eds.), Causal cognition: A multidisciplinary debate. Oxford, England: Oxford University Press. O'Connell, S. (1998). Mindreading: An investigation into how we learn to love and lie. New York: Doubleday. Ozonoff, S., Pennington, B. F., & Rogers, S. J. (1991). Executive function deficits in highfunctioning autistic individuals: Relationship to theory of mind. Journal of Child Psychology and Psychiatry, 32, 1081-1105. Povinelli, D. J., & Preuss, T. M. (1995). Theory of mind: Evolutionary history of cognitive specialization. Trends in Neurosciences, 18, 418-424. Premack, D., & Woodruff, G. (1978). Does the chimpanzee have a theory of mind? Behavioral and Brain Sciences, 1, 515-526. Quiatt, D., & Reynolds, V. (1993). Primate behavior: Information, social knowledge, and the evolution of culture. Cambridge, England: Cambridge University Press. Russell, J. (1998). Autism as an executive disorder. Oxford, England: Oxford University Press. Schulkin, J. (2000). Roots of social sensibility and neural function. Cambridge, MA: MIT Press. Shweder, R. A., & Bourne, L. (1984). Does the concept of the person vary cross-culturally? In R. A. Shweder & R. A. Levine (Eds.), Culture theory: Essays on mind, self, and emotion (pp. 158-199). New York: Cambridge University Press. Teitelbaum, P., Teitelbaum, O., Nye, J., Fryman, J., & Mauer, R. G. (1998). Movement analysis in infancy may be useful for early diagnosis of autism. Proceedings of the National Academy of Sciences, 23, 13982-13988. Tomasello, M. (1999). The cultural origins of human cognition. Cambridge, MA: Harvard University Press. Tomasello, M., & Call, J. (1997). Primate cognition. New York: Cambridge University Press. Wellman, H. M. (1990). The child's theory of mind. Cambridge, MA: MIT Press. Whiten, A. (Ed.). (1990). Natural theories of mind: Evolution, development and simulation of everyday mindreading. Oxford: Blackwell.

This page intentionally left blank

P A R T

III JUSTICE AND FAIRNESS

This page intentionally left blank

C H A P T E R

5 The Domain of Work in Households: A Relational Models Approach* Jacqueline J. Goodnow

Fiske (1991) proposed that one of the directions in which to develop his original model consists of asking "who uses which models, in what domains, and when" (p. 137). This chapter takes up that question, using as a database the views, actions, and feelings of children and adults with regard to household work contributions. For this area of activity—this "domain," to use Fiske's term—the chapter considers the "manifestations" of each of Fiske's four orientations to social life: Communal Sharing, Authority Ranking, Equality Matching, and Market Pricing. In the process, it brings out the value of Fiske's approach as a way of conceptualizing this area of activity, argues for the richness of this content area as a way of pursuing Fiske's questions, and raises some general questions and suggestions with regard to the orientations he proposes. The chapter opens with a background section built around the question: Why this domain? (The general argument is for two-way benefits: benefits to the analysis of both this area and of Fiske's proposals.) The sections that follow work through the several types of relationship that Fiske has proposed (Communal Sharing, Authority Ranking, Equality Matching, and Mar*This chapter has a single author. 1 am, however, clearly indebted to several colleagues for their contributions to the data and the ideas expressed here. Alphabetically, they are Jennifer Bowes, Susan Delaney (who brought me also "The Saga of Paul"), Jeanette Lawrence, Jacqui Smith, and Pamela Warton. Thanks are also due to the Australian Research Council for support in past and current studies on the organization and the perceptions of work within families.

167

168

GOODNOW

ket Pricing). Each section considers the forms that these orientations take, the circumstances that appear to be related to these particular forms, the ways in which distinctions and forms of implementation come to be understood and observed, and the significance attached to errors or confusions between one type of relationship and another. The final section takes up some general questions about the identifying features of various relationships, occasions of satisfaction and dissatisfaction, sources of tension, and areas of cultural difference.

INTRODUCTION: WHY THIS DOMAIN?

One of the domains that Fiske (1991) singles out as amenable to analysis by the application of his model is the organization of work (p. 42). The reference is to paid work or to work that produces tradable or tangible goods. Household work contributions are not a direct part of this analysis. That area, however, warrants attention for three reasons. Reason one is that relationship issues are central. The second reason is that the database for this area is now rich enough to warrant linking to Fiske's proposals about relationships and work. The third reason is that this is an area particularly in need of the kinds of integrative proposals that Fiske's proposals offer. The Centrality of Relationships The area of household work is not usually considered as a prime area for exploring theories about the nature of interpersonal orientations. Relationship issues, however, have emerged as central within a series of studies my colleagues and I have carried out on work in household settings. Some of these studies have dealt with children's contributions. Others have dealt with adults' "divisions of labor" in the course of couple relationships. Some current research is moving to a later part of the family cycle, exploring the distribution among adult children of the several tasks involved in giving personal and household assistance to an elderly parent in need of help. Studies concerned with the first two topics are reviewed in Goodnow (1996) and Goodnow and Lawrence (2001). Studies related to the third topic are covered in Goodnow, Lawrence, Ryan, Karantzas, and King (2002) and Lawrence, Goodnow, Woods, and Karantzas (2002). The opening study in this series provides an illustration of the way in which relationship issues emerged as central. It started from an interest in the way mothers construed the place of children within the work of a household (Goodnow & Delaney, 1989). (The mothers in this group and in most of the later studies were Australians, predominantly "Anglo" in background.) Expected were the kinds of comments reported in earlier studies

5. THE DOMAIN OF WORK IN HOUSEHOLDS

169

carried out by others in Australia, England, and North America: comments, for example, about the value of task contributions as a way to promote, in children, a sense of responsibility and perhaps a level of competence that would reduce a mother's load or stand the child in good stead later in life. The expected comments about the development of responsibility and competence did appear. The surprising result was the frequency and strength of comments made about distinctions among relationships, and the degree of affect attached to whether work contributions were being made in a way that met what was expected to occur "in a family." When asked about "good moments" and "low moments" with regard to children's contributions, for example, mothers described themselves as frequently using phrases such as "this is a home, not a restaurant, boarding house, delicatessen, hotel, cafeteria, laundromat, etc." They pointed out the need for children to learn the difference between "mothers" and "maids" and the importance of not behaving like "a king" or a "queen," expecting everyone to pick up after them. At the same time, these mothers worried that the task of "getting the message across" was undermining the kind of relationship they wanted to establish. They felt, many said, often forced to "behave like a nag" or "a Simon Legree" when what they hoped for was a warm and easy relationship in which their children's voluntary helpfulness showed that they "really do sometimes think about you and recognize what you do for them." Mothers, it turned out, were not unique in their concern with such relationship distinctions. Fathers were likely to say to children that they were "not running a taxi service" or were not "automatic banking machines." Children would use, toward each other, phrases such as "I'm not your slave," "you're not paying me," or "who do you think you are?" In short, a concern with relationships, distinctions among relationships, and what appeared to be relationship errors or confusions seemed pervasive. My colleagues and I began to point out the several forms that this concern took (e.g., Goodnow, 1996; Goodnow & Warton, 1991). We began also to give closer attention to models that distinguished among various types of relationships, with particular attention to Alan Fiske's description of four types of relationships. Here was a model with several attractive features. It was integrative. Work now appeared not as a segregated area but as one of several activities to which an overarching model applied. These activities ranged from "the circulation of things in all kinds of transactions" to "the organization of work . . . marriage systems . . . group structure and norms" and "responses to transgression and misfortune" (Fiske, 1991, p. 407). At the same time, the model also proposed several forms of domain specificity. To each domain, the same models were expected to apply. The "many manifestations of the four basic models," however, were expected to vary "as a function of the

170

GOODNOW

domains . . . , of the options that people a d o p t . . . , and of the flexibility people have in implementing a model in concrete action" (Fiske, 1991, p. 24). Fiske's model also offered some specific proposals about the forms that domain specificity might take. For each domain, he outlined the ways in which each of the basic four orientations might be expected to appear and the "cultural implementation rules" (p. 46) expected to accompany them. For the area of work, for example, this outline covered the ways in which work was organized, how orientations to work could become sources of "discord" and "distress" (p. 130), and some developmental aspects to relationship distinctions, primarily in the form of suggesting ages at which the first signs of distinctions appeared. How well do these proposals fit or extend the data on work in household settings? And what does that data suggest with regard to the proposals and the model they stem from? Those questions, I felt, called for a reasonable database in a particular domain. That database has been accumulating. We now have a reasonable amount of data on the way people act, feel, and think about household work contributions. We know a fair amount, for example, about how people distinguish among tasks, how they divide work in practice, what the meanings are of terms such as fairness or sharing, where the points of tension or negotiability often are, and what particular arrangements or procedures give rise to the sense that all is not turning out as one hopes it would. We even have the beginnings of knowing how some of these facets of work vary from one cultural group to another, although the bulk of our data comes from one group: urban Australians, predominantly middleclass and predominantly Anglo in background. This volume now provides the spur to bringing together that database and Fiske's proposals. The Need for an Integrative Approach to Work and Relationships Already noted is one of the particular strengths of Fiske's proposals: their integrative power. That strength offers a special benefit when it comes to exploring work in households and its ties to relationships. That exploration quickly encounters two challenges: diversity in the conceptual frameworks usually brought to bear on each of three lines of research—children's work contributions, adults' work contributions, and types of relationships—and the separateness of these lines of research from one another. Children's Work Contributions. Traditionally, these have been considered under conceptual rubrics that do not highlight relationship issues (Goodnow, 1988, provides a review). In some analyses, the framing is in economic terms. Children's task contributions are treated as forms of labor, a

5. THE DOMAIN OF WORK IN HOUSEHOLDS

I7 I

framing that tends to prompt questions about the productive value of their work, the structural conditions that make child labor needed, and the economic consequences for children and their families of this work being undertaken or discontinued. In other analyses—usually for countries where the economic value of the work is not a family's concern—children's task contributions are most often regarded by researchers as aspects of "socialization," directed primarily toward the development of competence in a task or a sense of responsibility. That framing has meant that children's contributions are seldom regarded as an introduction into the understanding of relationships. It has also meant the making of few connections with analyses of work contributions by adults in couple relationships: an area where an interest in relationship issues has been stronger although still of a limited kind. Adults' Work Contributions. Among scholars who analyze the differential contributions of men and women to household work, the recognition of relationships is especially marked within feminist analyses (Goodnow & Bowes, 1994, provide one review). These analyses are major sources for insights into the way people distinguish, both in research and in daily life, between "labors of love" and "scut work," between work done "for others" and "at the behest of others," between work that is "a gift" and work that stems from "oppression" or from a distinctly junior status. By and large, however, these analyses of adults' work concentrate on one particular aspect of relationships. This is the presence of "equality" or "inequality," indicated by differences between people in terms of who does what percentage or what type of work, who decides what work will be done by each, or who has the ability to change what is done or to move it to someone else. The main distinction among relationships then is based on differences in power, and the main questions have to do with the conditions that give rise to this difference, sustain it, or change it. Not surprisingly, there has been little interest within this line of research in considering other kinds of distinctions, even though these appear likely to be very relevant: distinctions, for example, between "communal" and "exchange" relationships (e.g., Clark & Chrisman, 1994; Clark & Mills, 1979), or among the four types of relationship described by Fiske (1991, 1992). General Accounts of Relationships. There are by now several accounts of the ways in which relationships or orientations toward social life may be distinguished from one another. To abstract from Fiske's (1991) summary (pp. 25-39), these range from a distinction between associations based on organic as against mechanical solidarity (Durkheim) to distinctions between gesellschaft and gemeinshaft (Tonnies), orientations toward commu-

172

GOODNOW

nion and agency (Bakan), transactions based on householding, redistribution by a central authority, reciprocity, and marketing (Polanyi), familial, custodial, reciprocal, contractual and voluntary modes of organizing work (Udy), and—from social psychology rather than from anthropology or sociology—between communal and exchange relationships (Clark). Why have these analyses not reached out to the analysis of household task contributions by children or adults? In large part, the answer lies in the nature of what are regarded as key problems. Within anthropology and sociology, for example, one driving concern has been the exploration of how tradable goods come to be produced and exchanged: the exploration of how economic and social life are interwoven. That emphasis on tradable goods has helped exclude household task contributions from the world of "work." It has also meant that studies of "economic" work have provided the readier database for Fiske to use when asking how his basic four orientations apply in the domain of work. Within social psychology, a different basis applies. The driving concern has been the explanation of close or intimate relationships among adults. Some concern with the organization of work is present. It is present, for example, in studies of the extent to which people restrict score-keeping when, in the course of a joint task, they hope to establish a friendship with a person newly met (Clark, 1984). In general, however, an emphasis on friendships and on the early stages of relationships has reduced the appeal of research on the organization of work in households. It is then predominantly within studies of established couple relationships (and, in the work of Mikula and his colleagues, of people sharing space with or without a couple relationship; e.g., Mikula, 1998) that there is a recognition of the significance of household work arrangements for the sense of satisfaction with a relationship. In short, here is an area in need of an integrative set of proposals. Here also is an area with promise as a way of adding to proposals that emphasize the significance of relationship issues in the way work is organized and regarded. To bring out that aspect, we turn to the ways in which the material brings out features and "manifestations" for each of the four models or orientations that Fiske has proposed.

COMMUNAL SHARING At this point, the chapter turns to each of Fiske's four kinds of relationships. The amount of space given to each is uneven. Communal Sharing, for example, is given more space than the other orientations, in part because it appears first and in part because much of the available database has a specific relevance to the circumstances under which this orientation appears.

5. THE DOMAIN OF WORK IN HOUSEHOLDS

I 73

For each type of relationship, however, the same questions are raised: questions about the identifying features proposed by Fiske, the forms in which this orientation appears, the links to emotion, and the signs of any developmental sequence. Identifying Features Fiske's (1991) description first notes a quality that cuts across all areas of activity. The "essence of Communal Sharing" is that "it is a relationship based on duties and sentiments generating kindness and generosity among people considered to be of the same kind, especially kin" (p. 14). The ideology, and the basis for moral judgment, is one of "caring, kindness, altruism, selfless generosity" (p. 46), together with "protecting intimate relationships" (p. 46). Displays of this kind of orientation (signs of its being "externalized") are suggested as emerging early, probably in infancy (pp. 46, 402). Fiske's description contains also some features proposed as specific to the organization of work. When the orientation is one of Communal Sharing, "everyone pitches in and does what he or she can, without keeping track of inputs" (p. 42). "Tasks are regarded as (the) collective responsibility of the group, without dividing the job or assigning specific individual duties" (p. 42). Forms of Manifestation What does our data on household work contributions add to these descriptions of Communal Sharing? Several features stand out: • The appearance of this orientation depends on the criteria used. If, for example, we stay with the general marks of a communal orientation—an emphasis on kindness, generosity, sensitivity, respect for intimate relationships—examples of this orientation are easy to find. Mothers, for instance, describe themselves as thinking twice about asking children to do tasks that the children especially dislike. They encourage their children to help each other. They point out that "we're a family and in a family people help each other out." They are delighted when their children make a "gift" of some particular task contribution: a contribution that is unasked-for and— ideally—truly helpful. Communal orientations also appear to be present when adolescents describe themselves as cautious about the way they give reminders or check that a delegated job has been done. In their terms, "you don't want to hurt their feelings," "you don't want to look as if you don't trust them," "especially if it's someone in the family" ("you should be able to trust your family") or "especially when they're doing you a favor" (Goodnow & Warton, 1992).

174

GOODNOW

Adults in couple relationships express similar sentiments. In the words of many, "it's the relationship that matters." Dividing work is then not simply a matter of efficiency or competence but a means of expressing and protecting the relationship. To quote one man's explanation of why he shares household work, "you can't have a good relationship if one person is full of resentments" (Goodnow & Bowes, 1994). A more complex picture emerges when we look at some of the specifics of implementation, at some of the particular ways in which the orientation of Communal Sharing may be expressed. To repeat Fiske's description, the prototypical sign of Communal Sharing in the domain of work organization is that "everyone pitches in," with tasks "treated as the collective responsibility of the group, without dividing the jobs or assigning specific individual duties" (Fiske, 1991, p. 42). In the particular cultural group that my colleagues and I have worked with, this way of expressing "collective responsibility" turned out to be limited in its appearance. At all ages, for example, people used the phrases "my job" or "your job" far more than they used the phrase "our job," suggesting that the bias was toward perceiving tasks as individually owned rather than collectively owned. In the majority of families and couples also, tasks were assigned to, associated with, or done by particular individuals rather than by the group or by any member of a group. In one sample of Australian families, for example, only 11/6 took a collective orientation in the form of frequently "pitching in." In that subset, jobs did belong to anyone. The prevailing mode was one of allowing jobs to pile up until one person's limit was reached and then "everyone pitches in, regardless" (Goodnow & Delaney, 1989). In effect, this way of organizing household work did occur. As a regular way of proceeding, however, it was clearly a minority mode. Does this kind of result mean that Communal Sharing rarely appears as an orientation among these Anglo-Australians? What one needs to do, as Fiske suggests, is to look at the specifics of implementation and at the cultural rules that indicate how an on orientation may be expressed. In any cultural group, for example, it seems unlikely that a collective responsibility approach will be taken toward all forms of work. Even in nonindustrial societies, for example, some areas of work are likely to show signs of individual rather than collective ownership: to be, for example, "women's work," "chiefs' work," or "priests' work." The interesting question is then not whether a group is to be regarded as more versus less communal in its orientation but what the particular types of work or the particular conditions are that make a communal approach to ownership more versus less likely to emerge on specific occasions. The interesting question is also one of how we can help these specifics to emerge.

5. THE DOMAIN OF WORK IN HOUSEHOLDS

I 75

• To gain a finer picture of "ownership" or "collective responsibility," it is useful to add a further feature: the extent to which a task is movable from one person to another. Two pieces of work, for example, may normally be done by one person within a group. On that basis, both may appear to be an individual responsibility. One task, however, may be movable to other members in the group, while the other is not. The move may be made easily or with difficulty. In one way or another, however, the task ultimately becomes the responsibility of one or more people in the group. In that respect, it may be regarded as having more of a collective quality than does the less movable task. The phenomenon to explore is then the nature of movability. That can be done in several ways (see Goodnow, 1996; Goodnow & Lawrence, 2001). Over a series of studies, for example, my colleagues and I have asked which tasks—from a given set—one family member could ask another family member to do on a day when the asker is especially busy. We have also explored which other family member could be asked, whether the request would be made easily or after some second thoughts, what kind of justification might be offered for asking that a task move from one person to another, and whether ownership ends when the person who has been asked agrees to take on a task (e.g., do you still have the responsibility to check that the job has been done?). Under what conditions does a move cease to be within a family network, involving instead people or agencies outside the family? In short, there are several ways of exploring the nature and the limits of "ownership," providing a base for seeing how and when individual or collective ownership applies and is enacted. The next proposals stem from this kind of exploration. • Individual or collective ownership varies from one type of task to another. Take the results that occur when we ask mothers whether they would ask one child to do a task that is usually done by another. Our first inquiries along these lines quickly netted the reply, "depends on the job." Household tasks, it turns out, tend to be of two kinds when movability is the issue. Some of these are "self-care" tasks. They involve looking after one's "own stuff" or one's "own space." They also involve being responsible for what one has caused: putting away what one has used, cleaning up any mess created, making one's bed, doing something about a problem one has given rise to. Except under special circumstances, and "not regularly," the mothers in this group felt reluctant to ask one child to take over another's "self-care" tasks. In contrast are jobs that involve "other-care." These are jobs that do not have the adjective "own" attached to them and that involve shared space or benefit to others: cleaning a living room compared with cleaning "one's

176

GOODNOW

own room," for example, setting the table, preparing a meal or helping prepare a meal that others eat. Mothers felt far more comfortable about moving these jobs from one child to another and children offered fewer objections. Similar task differentiations are also made by adults. Even among couples who moved most tasks easily from one to another, for example, individual ownership surfaced when questions were asked about arrangements for washing a car or getting it serviced (Goodnow & Bowes, 1994). "Whose car?" was the frequent answer, or "depends on who's driving it most." The same kind of differentiation was often made with regard to ironing clothes. Where the task survived at all among these couples, it tended not to be collectively owned or shared but to be most often done by the person who wore the clothes that needed to be ironed: the person who "cared about whether things were ironed" or who "bought stuff that needed to be ironed." Among both children and adults, it appears, the ownership of a task often goes with the perception of causation, the perception of who gave rise to the need for the work to be done. The perception of causation is certainly not the only basis for tasks being seen as more versus less movable. Some tasks, for example, are more versus less movable across gender lines. To take an example from the distribution of work in the course of assisting a mother recovering from surgery, tasks such as shopping or doing the laundry are far more movable from daughters to sons than are tasks such as making her bed or giving her a shower (Lawrence et al., 2002). The specific tasks that are more versus less movable may vary from one situation or one social group to another. The challenge in all cases, however, is one of determining the specific tasks that are seen as more versus less movable in various directions—more versus less "owned" by particular individuals, particular types of people, or the entire group—and the bases for their being perceived in these ways. • Collective ownership emerges as an overriding appeal. For this point, I return to comments from mothers with regard to children's work contributions. The mothers in our "Anglo" groups clearly had some sympathy with children who, when asked to take on a task, responded with ownershiprelated objections: "That's not mine," " I didn't use that stuff," "didn't take it out," "didn't leave it there," etc. The mothers' response took the form of "Yes, but... ." The objection was acknowledged as legitimate but the child was expected to respect in turn the justification that "I need your help," "we're a family and we all help each other," "we all have to do more than just look after our own things," "I don't just clean up or cook for myself," or—less communally —"I'll ask the others next time." The overall implication is that family members have a particular responsibility for their "own stuff" or "own mess." They are also expected, how-

5. THE DOMAIN OF WORK IN HOUSEHOLDS

177

ever, to rise above that limited definition when something more is needed or asked for: an expectation that may not be the easiest for parents to teach or for children to understand (Goodnow, 2000). Sources of Tension and Emotion One of the striking features to household work organization in the groups we have considered, at all ages, is the amount of affect it generates: both positive and negative. Striking also is the amount left unexplained by accounts that concentrate on perceptions of fairness based mainly on the overall amounts of work that each person contributes (Goodnow, 1998). I have accordingly looked, within Fiske's model, for proposals that would help account for such feelings. One of Fiske's proposals is that difficulties can occur when there are "unstated—or even explicit—discrepancies between the models that people use in structuring their relationships. . . . For example, in a marriage or any other relationship, actions concordant with Authority Ranking offend a partner who expects Equality Matching, just as adherence to Market Pricing principles violates a partner's Communal Sharing assumptions" (Fiske, 1991, p. 134). The area of household work contributions provides further examples of the impact of such discrepancies, both between people and, within an individual's experience, between what a person hopes for and feels compelled to settle for. The mother who feels distressed because she is "forced to nag," for example, is essentially hoping to find a relationship of Communal Sharing, with tasks willingly or generously done. She finds herself, however, feeling that the only way to proceed is by becoming a harsh authority: not simply asserting authority but doing so in a way that takes much of the "family" quality out of the interaction. More subtly, a sense of dissatisfaction can also occur when people do not follow what Fiske describes as the expected "cultural implementation rules." For an orientation of Communal Sharing, for example, the cultural implementation rules specify such aspects as "what kinds of restraint people must exercise in taking from others, and what excuses them from giving" (p. 46). The notion of such implementation rules fits nicely with some otherwise puzzling adult interactions about work. These are interactions around the use of "incompetence" as a basis for not making a particular contribution. The data in this case come from a sample of couples who did not follow gender-stereotyped patterns of household work (Goodnow & Bowes, 1994). Most of this group avoided expressions of unwillingness. Incompetence ("I don't know how," "I just don't see what needs to be done," "I forgot") was a more acceptable plea to most than unwillingness, even when the excuse of-

178

GOODNOW

fered was felt to be spurious. An open unwillingness to take on a piece of work, leaving it for someone else to do and without volunteering to make some other contribution: These approaches—once a person is old enough for the attribution "doesn't understand" to no longer hold—imply that there is no closeness and no respect within the relationship at all. Unwillingness accompanied by a readiness for someone else to take on a scorned task appears to be a particularly unhappy combination. The usual "cultural rule" in this group of couples may then be seen as one that specifies pleading incompetence rather than unwillingness. The nature of the rule is interesting in itself. So also is the question of when it is not followed. Within this group, for example, the people who openly declined particular tasks on the grounds of their dislike for these tasks were people who had already established a record of general willingness and who were sensitive to the issue of what might need to happen to tasks that both partners disliked (e.g., divide equally, or hire someone else who finds the job reasonable to do, but not try to pass it to one's partner). There is still a great deal to be learned about when people can set aside the usual implementation rules and when they cannot do so. The phenomenon, however, is clearly worth more careful attention, both in this area of activity and in others. Developmental Aspects Fiske (1991) suggests that signs of a general readiness to engage in Communal Sharing appear in infancy, in the form of sharing food and being responsive to the feelings of others (p. 402). Where work is concerned, we know at least that, within the United States, children in the toddler-age group are interested in the work that mothers do. In fact, they often try to take part or to take over, to the point where mothers may schedule their work for the children's nap-time (Rheingold, 1982). Around age 2 is also the age for which there is one report of a mother (English in this case) using causation as a reason for not accepting a proposal about work. The mother had asked the child to put away something she had played with. The child replied "Annie," implying that an older sister could put the toy away. The mother's response was: "Annie didn't take it out, did she?" (Dunn, 1988, p. 54). By a slightly older age, the developmental picture seems to be one of young children coming to adopt ownership based on causation, but initially in a relatively black-and-white fashion. Within a U.S. study, for example, 4-year-olds judged as less likable a child who did not share in the cleaning up after a joint game. "They both played, they both should clean up" was the reason given (Shure, 1968). It was not until after age 8 that children in Shure's study also began to see virtue in a

5. THE DOMAIN OF WORK IN HOUSEHOLDS

I 79

child who helped with the cleaning up but had not played. Altruism in this form, apparently, is slow to be appreciated. In related fashion, it was only among 11- to 14-year-olds in an Australian study that children began to regard it as "fair" that one child in a family might be asked to clean up after two had played a game (Warton & Goodnow, 1991). This group was still a minority, and the judgment of fair was on the grounds that "things would even out in the long run." "Evening out in the long run" is perhaps a mixture of Equality Matching and Communal Sharing. It does suggest, however, that one of the essential bases to being willing to engage in Communal Sharing, and to tolerate some uneven distributions in the process, may be the developing confidence that there will be "a long run" to the relationship and that some account will be taken of what was contributed in the past.

AUTHORITY RANKING

The analysis of this orientation begins again by noting the identifying features proposed. It then considers, for the domain of household work contributions and within "Anglo" groups, forms of manifestation and areas of tension and negotiation, with particular attention to the latter. Identifying Features

In general, this orientation is seen as "a relationship of inequality" in which people "construe each other as differing in social importance or status" and possibly in "knowledge and mastery over events" (Fiske, 1991, p. 14). The orientation differs from "purely coercive power in which people dominate others by force or threat of harm . . . typically, subordinates believe that their subordination is legitimate" (p. 14). The cultural implementation rules have primarily to do with "the criteria for according rank" and "in what domains may authority be exercised" (p. 46). Within the domain of work, the pattern is one in which "superiors direct and control the work of subordinates, while often doing less of the arduous or menial labour" (Fiske, 1991, p. 42). In effect, some people assign tasks to others, or make the final decisions as to how work will be done. The form that actions take and the ways in which these are understood, however, are expected to vary, not only among cultural groups but also among situations: "Authority Ranking in family relations (where it is linked to Communal Sharing)," for example, "may be experienced differently from Authority Ranking in the military" (p. 24).

180

GOODNOW

Forms of Manifestation Authority Ranking is clearly relevant to the way children's work contributions to the household are organized. The most frequent arrangement in our samples of Australian families is one in which parents assign tasks and check that tasks have been done (Goodnow & Delaney, 1989). Parents may present a task as a request, but the request still has overtones of being an order. Certainly the expectation is that the request will not be declined without a good reason being offered. Parents may also avoid some arrangements on the grounds that they appear "too managerial": the grounds offered by some mothers, for example, for not using rosters. Parents may also soften their supervision by asking such questions as, "Have you had a chance to do X yet?" That type of approach is reported with approval by adolescents ("it makes us more like equals"; Goodnow & Warton, 1992). The legitimacy of checking on the completion of an assigned and accepted job is not the issue. It is the procedure, the way of doing so, that needs to be modified to fit the family setting. Between adults in couple relationships, issues of authority are more problematic (Goodnow & Bowes, 1994). Neither partner is expected to have the right to tell the other what should be done. What often occurs, however, is negotiation over who has the major say when it comes to what needs to be done ("is this necessary?"), the standards that should apply ("How well or how often does this need to be done?"), and the possibility of outsourcing—of moving jobs to people or agencies outside the family ("Do I have the right to move some of my jobs to people outside the family? If so, who should pay for it? Should it come out of 'my' money or 'our' money?"). Problematic also for these adults is the sense that, despite an ideology of sharing or of equality, they may still end up with tasks that, in the world at large, are indicative of a junior status. Tasks that involve "picking up someone else's dirty socks" or their equivalent are of this type. The sense of having somehow acquired the low position in a hierarchy seems often to sit uneasily with an expected partnership of equals.

Areas of Tension and Negotiation The occasions that call for particular attention appear to be those where both Authority Ranking and Communal Sharing may be evoked, or where a relationship previously marked by one of these orientations changes into the other. For a close account of such an interplay, we turn to an account offered by an Australian mother.

5. THE DOMAIN OF WORK IN HOUSEHOLDS

I8 I

The Saga of Paul We have a kitchen tidy (a kitchen trash can) which needs emptying every couple of days. At one stage, Paul—13—took this job on himself without my asking him—generally doing it while I was in the kitchen, and so gaining my notice, my thanks, and my comments about his helpfulness. I was happy to hand the task over to him and before long would automatically call on him to do the job if it needed doing. (Included in the job was putting out the rubbish bin out on the street twice a week). Before long, Paul began to rebel. He insisted that it was not his job and why didn't I ask one of the other two children to do it. I suspect he may also have been blamed for not putting the rubbish bin out, so missing the collection. Automatically assuming he would remember, I too readily handed that responsibility over to him, not even thinking to remind him. The consequence seemed to be that poor Paul had, through his own goodwill, become saddled with a responsibility he had not intended to take on. When the garbage was my job, he got thanks and praise for helping me. When it became his job, the thanks and praise diminished, he felt taken for granted and even worse, growled at if he slipped up. We resolved the situation by going back to the beginning. Emptying the kitchen tidy became my job again. I would do it, or pointedly (in Paul's hearing) ask one of the other two to oblige. Occasionally I would ask Paul and he would willingly do it with the comment "even though it's not my job." Now when I ask him to do it, I will sometimes preface the request with "I know it's not your job, but ..." and he is happy to do it. Even better, he will voluntarily take the rubbish out for me again—sometimes forgetting to replace the plastic bag—a little hint to me, I think, that I must notice and thank him. As for putting the garbage bin out on the street, he's doing that too. On rubbish night, I might ask everyone to remind me to put the bin out or say out loud that I must go and do it. Very often, Paul—in a grown-up voice—will tell me not to worry, he's already done it—or will readily go and do it for me. The lines between gifts and expected or assigned contributions, it appears, can be easily eroded. They are also lines that people may have a vested interest in maintaining, in the direction of either ensuring that a contribution retains its status as a gift or, in the reverse direction, avoiding the translation of an assigned contribution into one that is regarded as optional. Developmental Aspects Fiske (1991) suggests that children's sensitivity to "size, age, and priority" begins around the age of 3, indicated by children's frequent use of "me first" and "lots of talk about . . . who was bigger than whom . . . power and rank clearly appeared as a new, distinct, very salient issue" (p. 402).

182

GOODNOW

Our database again says less about the first signs and more about a continuing sensitivity to occasions when one is treated as a subordinate. The degree of sensitivity may vary with age, perhaps becoming more marked at adolescence when the division lines between child and adult become matters of great importance. There is probably also an increasing sophistication with age in the ways by which children signal that they feel an adult's way of proposing a task has been inappropriate. (E.g., an 8-year-old I have observed greets parental requests that he regards as inappropriately expressed with a fake salute and a crisp "Yes Sir!") The data of particular interest have to do with signs of cultural differences. These come from a study with parents in Beijing by Bowes, Chen, Li, and Li (1999) and studies with Anglo and African-American parents in the United States by Smetana (Smetana, 2000; Smetana & Asquith, 1994). The study by Bowes et al. (1999) is especially relevant to cultural implementation rules that have to do with when authority can be questioned. This study used the format noted earlier of a parent asking one child to do all the cleaning up after two children have played. Given the high proportion of one-child families in Beijing, the child in this case was a friend. Mothers were asked to rate the frequency and the acceptability to them of various responses that the child might make. One of these responses was "I didn't use all of those things." Beijing parents rated that response as neither reasonable nor frequent. In contrast, "Later, after I do my homework" was regarded as more acceptable and more frequent. Both mothers and fathers made the point that children must learn to clean up after themselves, to do "their" jobs. The principle of being especially responsible for what is one's "own" appeared then to be alive and well. Using this principle to query a parental request, however, was a behavior not expected to occur. The research by Smetana is especially relevant to cultural implementation rules that have to do with "the domains in which authority may be exercised" (Fiske, 1991, p. 46). The results of most interest for our present purposes are those that stem from the comparison of African-American with European-American families. Smetana (2000) notes: • The appearance of similar age changes: In both, moving toward adolescence brought an increase in the extent to which children expressed reservations about parents making the rules. • The significance of domains in both groups: As in the earlier work, parental authority was seen as more legitimate ("Is it OK for parents to make rules about this?") in areas that have to do with a child's safety or with clearly moral issues (e.g., keeping the promises made to parents) than in areas having to do with one's friends or with the state of one's room (these are regarded by Smetana as "personal" or mixed domains).

5. THE DOMAIN OF WORK IN HOUSEHOLDS

183

• A difference in the range of areas where it was seen as legitimate for adolescents to set the rules (narrower in the African-American group). • A difference in the view taken of two aspects of authority: the perception of it being legitimate for parents to make the rules in an area and the perception of an adolescent's having the duty or obligation to comply with the rule. The latter perception is tapped by the question: "Do children have a duty or an obligation to follow the rule if parents make a rule about an issue and the child doesn't agree with it?" Among African-American adolescents, these two perceptions were negatively correlated, suggesting that they draw more of a distinction between obedience and "rights" than the European-Americans do. As Smetana (2000) commented, such variations make it inappropriate to regard African-American families as broadly more authoritarian than Anglo families are. Fiske's model fits nicely with the more differentiated picture that Smetana offers. The critical comparisons have to do with the domains in which authority is seen as legitimate or not, and in the cultural implementation rules that apply to the actions that may follow from those perceptions. A parental rule may be seen as questionable, but the cultural mores may make that questioning private rather than explicit.

EQUALITY MATCHING AND MARKET PRICING

These two orientations or types of relationship are by no means identical. I place them in one section partly because they are, as Fiske (1991) points out, often treated together under the same label of "exchange" relationships (p. 16). Both are also less frequent in our database than are the two previous orientations. Both, however, do appear and the forms of their appearance add considerably to our understanding of what people regard as ideal or as tolerable. Identifying Features Equality Matching is marked by "an egalitarian relationship among peers. . . . Equality Matching may be manifested in turn-taking, (or) in strict in-kind reciprocity, such that people receive and give back what they construe as the 'same thing'... regardless of the objective differences in the entities involved. . . . It does not matter who gets (or gives) what share or who takes which turn: in short, everyone is equal and things come out even" (Fiske, 1991, pp. 14-15). The cultural implementation rules have to do with "Who and what counts as equal. What procedures people use for matching and

184

GOODNOW

balancing. How people initiate turn-taking. What are the appropriate delays before reciprocating" (p. 47). For the specific domain of work, the expected forms of organization take the form of each person doing "the same thing in each phase of the work, either by working in synchrony, by aligning allotted tasks so that they match, or by taking turns" (p. 43). There are then several ways by which Equality Matching can be achieved: a variety that opens up questions about when one route comes to be preferred over others or avoided. The main feature of a Market Pricing relationship is that "people typically value other people's actions, services and products according to the rates at which they can be exchanged for other commodities . . . people structure their interactions with regard to a common standard in the domain (money, time, or utility)" (Fiske, 1991, p. 16). The cultural implementation rules have to do with "What entities may be bought and sold (sex? drugs? votes? people?). What are the ratios of exchange.... What counts as a cost or a benefit (either in money or utility terms)" (p. 47). For the domain of work, the identifying feature proposed is the presence of work "for a wage calculated as a rate per unit of time or output" (p. 43). Equality Matching: Manifestations In a rare and provocative study of score-keeping, Clark (1984) found that people who expected that a friendship might develop between them avoided strict and overt score-keeping in the course of a joint task. Clark (1984) then raised the possibility that people might regard score-keeping as incompatible with a Communal as against an Exchange relationship. Raised also, in relation to Fiske's later model of relationships, is the question of whether people can blend Communal Sharing and Equality Matching orientations and how they might do so. Within our database on household work contributions, the most relevant material comes from couples whose general ethos is one of sharing work and of being "fair" (Goodnow & Bowes, 1994). What stands out is the tendency of these couples to avoid strict equality, except for jobs that they both dislike. Dividing each job in half, or taking strict turns, tended to emerge only in the early stages of working out a pattern. These patterns were usually modified in the face of differences in skill (making unequal the time or effort required), differences in liking for a task (making unequal the emotional cost of doing the work), and the increasing confidence that each could be trusted to do a "fairly equal" share. It is not that score-keeping disappeared or was regarded as always incompatible with a close and caring relationship. Partners who were comfortable with one another clearly felt able to say that contributions were getting out of balance. These comments, however, were usually made with humor and the score was not closely

5. THE DOMAIN OF WORK IN HOUSEHOLDS

I 85

kept. Equality Matching, it appears, can be combined with a Communal orientation, but there are limits to when and how a concern with strict equality is expressed. Equality Matching: Sources of Tension Three stand out. One has to do with ways of achieving a balance. For some of the people in the Goodnow and Bowes (1994) group, a favored method was the one described by Fiske (1991) as "working in synchrony," often described by the people in this group as "one up, all up." For the initiators, this way of proceeding seemed to be both a way of making work balance and a way of underlining a group identity (a relationship certainly different from "one of us working in the kitchen and the other reading the newspaper"). The difficulty in several partnerships, however, was that the noninitiating partner had a different sense of appropriate times for "all up" or had a preference for balancing in less immediate fashion, with the goal of equality seen as met if things worked out fairly evenly over the course of a week or longer. Under such circumstances, frequent proposals of "one up, all up" can readily appear coercive. A second source of tension appeared in partnerships where the relationship was not new and one partner sought to change the pattern toward one of more equal shares. The tension in these situations stemmed from the need to avoid appearing to criticize one's partner, to imply that they have been unfair, exploitive, insensitive, or sexist: all threats to the appearance of either an Equality or a Communal orientation to the relationship. For the third source of tension, I turn to a group outside our own database. In a study of Scottish families, where the wives had given up paid work in order to be full-time mothers and homemakers, Backett (1982) found that both partners were aware that the shares of work were uneven. Both emphasized, however, that this was a temporary state of affairs. It would change once the youngest child started school and the mothers were able to return to paid work. The past inequality would not be redressed by an inequality that reversed the old pattern. For the moment, however, comfort needed to be taken in the thought that there would come a time when the distribution of work would be more even. Faith then is needed that this time will come. Equality Matching: Developmental Aspects Fiske (1991) describes the children he has observed as displaying, around age 4, a phase "in which concern with equality and justice dominated their thinking about social relations" (p. 402). We have no data to add on the beginnings of this concern. The data do point, however, to this concern con-

186

GOODNOW

tinuing throughout childhood and adulthood. What appears to vary with age, with cultural group, and with the length of a relationship is the understanding of the implementation rules (e.g., what counts as "even" or as "the same thing," and the view taken of what would be a short-term or a longterm period for balancing the books). Market Pricing: Manifestations At first glance, market pricing might not appear relevant to household work contributions. Except when discussions arise about what it might cost to replace a full-time mother, whether it makes economic sense for mothers to be in paid work, or whether family work should not be included in estimates of the gross national product, people's contributions within families are not usually worked out on a wage or market price basis. Within our database, however, some aspects of market pricing appear at the family level in relation to paying children or adolescents for the work they do. Moreover, these aspects bring out some particular interplays among the orientations people may bring to household work contributions (Goodnow & Warton, 1992; Warton & Goodnow, 1995). Among the parents we have interviewed, for example, there was agreement about rejecting one particular link between money and jobs: paying separately for each job that is done. This kind of arrangement, parents said, could easily lead to endless negotiation and to work being undertaken only when a child was short of money (in effect, to Market Pricing of what Fiske, 1991, calls a "take-it-or-leave-it" kind; p. 16). These parents were also more accepting of payment for "special jobs" (e.g., washing windows) than they were of "regular jobs" ("their jobs," "the jobs they have to do"). At the same time, many of these parents were willing to pay for jobs "if this is the only way to get them to do something." Making no contribution is apparently worse than paid-for contributions. The latter at least allow the hope that payment can cease later or that some intrinsic motivation will develop. Many parents also occupied an ambivalent position. They described themselves, for example, as considering that pocket money and household jobs are completely separate. They also, however, described themselves as often asking "do you deserve your pocket money this week?" if the expected jobs were not done or not done well. Overall, the impression is of two kinds of ethos operating. On the one hand, there is the ideal that household contributions should be made out of goodwill or a sense of being one of the family. Money has no place in this expectation of Communal Sharing. On the other hand, the fact that money is changing hands seems to evoke for many of these parents an ethos of Market Pricing, in the form of feeling that people should avoid positions where the other person "is getting something for nothing." As long as jobs

5. THE DOMAIN OF WORK IN HOUSEHOLDS

187

are being done willingly and well, that sense of a poor market arrangement can be avoided. Once they are not, the feeling that one is paying more than one should seems to rise readily to the surface. Market Pricing: Sources of Tension Within the Western world, Fiske (1991) points out, people may readily think of market pricing as necessarily selfish, materialist, or competitive (p. 396). These associations may well be sources of discord if one party to an interaction sees any mention of money as having these connotations while the other does not. Fiske also draws attention, however, to another source of tension: one that helps account for the interesting concern with whether one's children are "getting something for nothing." His proposal is that we all have some investment in paying what one should or receiving what one should (p. 394). The aim is for competence in the market: for being, in the words of a famous poster, "neither a patsy nor a scrooge." To allow one's children "to get something for nothing" may then be experienced not only as bad for children but as a failing in one's own competence as a consumer or manager. Small wonder then that one subgroup of parents in our sample (about 5%) avoided ambiguities by ruling out any possible association between household jobs and money (Warton & Goodnow, 1995). In these families, jobs needed to be done because children were members of the household. Pocket money or other money was allocated only on the basis of age or particular needs. How their children felt about this way of avoiding any possible ambiguity, we do not know. For parents, however, the boundaries between one type of relationship and another were sharply drawn and the possibilities of confusion or the need for negotiation were considerably reduced. Market Pricing: Developmental Aspects In a child Fiske (1991) observed, the first clear signs of thinking in terms of Market Pricing appeared around age 8, manifested in a new interest in "money, in buying things, and in prices" (p. 404). Our data suggest that age also influences the understanding of a cultural implementation rule: a rule specifying what it is reasonable or moral to be paid for. At age 8, for example, the Australian children we interviewed regarded it as "fair" to be paid for making their bed, on the grounds that this was "a big job." At age 11, they saw the expectation of payment for making one's bed as absurd, as "like expecting to be paid for washing one's face" (Warton & Goodnow, 1991). For those areas where money can be considered, age and experience emerge also as influencing the understanding of what represents a reasonable amount and the timing of proposing that a contribution might be "a

188

GOODNOW

money job." In the words of one mother commenting on the suggestion of payment after an apparently voluntary job, "you have to tell them—ask before, not after it's done; asking afterwards spoils the favor."

CONCLUDING COMMENTS

Noted throughout have been the ways in which Fiske's (1991) proposals fit the data and help account for some otherwise puzzling phenomena. Noted also have been the several ways in which the database on household work contributions takes up and adds to a general research direction suggested by Fiske: namely, asking "who uses which models, in what domains, and when" (p. 137). The core of the chapter has dealt with each of the four orientations in turn. In this final section, we draw together some points that cut across all four. They have to do with identifying features, sources of tension, developmental aspects, and cultural differences. Identifying Features My suggestions and questions have to do with (a) the need to distinguish among settings or situations as well as between orientations, (b) the nature of the occasions people choose as contrast cases (e.g., as a contrast to "family" relationships), and (c) the significance of procedures as a feature distinguishing among orientations. Distinguishing Among Orientations and Among Situations. As a starting point, I return to Fiske's (1991) comment that Authority Ranking in the family is different from Authority Ranking in the military. The family setting, he suggests, combines Authority Ranking with Communal Sharing. The latter does not (p. 16). That kind of comment prompts first of all the need to distinguish between two uses of the term relationships: relationships in the form of connections that are socially recognized or conventionally labeled (e.g., marriage, family, business partners, employees) and relationships in the form of abstracted orientations or models. A conventionally labeled social relationship may involve more than one orientation: a point brought out especially by Haslam (1994). To the social relationship of marriage, for example, some people may bring only one orientation. Others may bring more than one, shifting from one orientation to another as situations arise. To the social relationship thought of as "the Army," some people may bring only the orientation of Authority Ranking. Others may expect and act on some perceptions of the Army as communal, as similar to "family."

5. THE DOMAIN OF WORK IN HOUSEHOLDS

I 89

To distinctions among orientations, we may also need to add distinctions among situations. Some seem likely to evoke a greater or a lesser mix of orientations. The mix in some, for example, may be relatively even. In others, the mix may be one orientation dominant, with a small dash of the other. Some situations may also evoke some combinations rather than others. The probability of tension—or at least negotiation—seems likely to be greater in situations that allow a greater mixture, unless the cultural implementation rules are explicit as to when each will prevail. In related fashion, the probability of tension and negotiation seems likely to be greater in situations where there is the expectation of a change in orientation at some point in time (e.g., parent-child relationships), again unless the points of changeover are institutionalized and allow few options. The Nature of Contrast Cases. Fiske uses "family" and "military" as contrast cases. So also do adolescents when commenting on close supervision as an inappropriate procedure when one person takes on a task for another family member ("it's not the army, you know"; Goodnow & Warton, 1992). That comparison prompts the question: What do people come to use as contrast cases? How do people come to use these contrast cases? How do children come to know, for example, that the "military" style is inappropriate in a family setting? The latter two questions are prompted especially by having recently watched a group of children aged 4 to 9 viewing the video The Sound of Music. Notable, especially among those of school age, was the pleasure felt in the absurdity of a father summoning his children and giving orders by way of a whistle: tables turned by the new governess who wonders what whistle would mean that his presence is needed. The script suggests that "the military" is a culturally available contrast case that children will understand, even though they may have had no direct experience of it. In the same way, mothers say "this is not a hotel," or "I am not a maid" to children who will seldom have had direct experience with either hotels or maids. The contrast labels are nonetheless being supplied before the direct experience and are presumably understood in some fashion. The Significance of Procedures. This is the most general point to be made about identifying features. Among the several domains that Fiske proposes, many have to do with what happens to goods or contributions of various kinds. This is the case, for example, for the domains labeled Reciprocal Exchange, Distributive Justice, Contribution, Work, The Meaning of Things, and Orientation to Land. In one other domain, however (Decision Making) what matters is not an outcome but a procedure. The difference between orientations in this case has to do with whether decision-making is

190

GOODNOW

based on seeking consensus, on authoritative fiat, on arrangements for an equal say, or on letting the market decide. The question that arises has to do with whether a procedure such as the manner of Decision Making is a separate domain or an ingredient in every domain. I raise the question in part because of adolescents' sensitivity to whether parents express their wishes as "a request" or "an order." I raise it also because of increasing evidence to the effect that the ways in which things are done permeates the sense of satisfaction or dissatisfaction people feel about arrangements in several domains. In courts, for example, a major aspect of satisfaction and dissatisfaction on occasions of distributive justice is not simply the outcome but the extent to which people feel there has been an opportunity to state their opinion or to have someone else present their position adequately (Lind & Tyler, 1988). In inheritance situations also (part of Fiske's domain labeled "The Meaning of Things"), Jeanette Lawrence and I are now learning that the sense of "really being a family" depends not only on what has been left but on the sense of there having been consultation or the opportunity for "voice" (Goodnow & Lawrence, 2001). "Decision Making" then may call for particular attention not only as a separate domain of particular importance but as a feature or ingredient of every domain, influencing in each case people's sense of the orientation that is in play. Sources of Tension

Fiske (1991) proposed that negative emotions (discord, tension, moral disapproval) may arise when people bring different orientations to the same event, on occasions when each feels that his or her own orientation is the appropriate one. Fiske and Tetlock (1997) have added the proposals that (a) negative affect goes with "relationship errors" (mistakes in the assignment or expectation of particular orientations), and (b) the degree of affect varies with the distance between two orientations when the four are thought of as stretched along a line running from Communal Sharing to Authority Ranking, Equality Matching, and Market Pricing. Negative affect also seems likely also to arise when the cultural implementation rules are breached, in circumstances where it is not easy to make some charitable attribution for the breach (attributions such as "still young," "still new," or "a relative stranger"). To these proposals our database suggests adding also questions about when orientation errors are most likely to occur and who is most likely to be wrongly placed (i.e., treated in the fashion of one orientation when another is expected). When are orientation errors most likely to occur? One possibility is that errors are more likely to occur when situations allow for a mix of orientations

5. THE DOMAIN OF WORK IN HOUSEHOLDS

I9 I

and when the markers for the appropriate mix or the appropriate occasions are not clear. "The Army," for example, usually contains multiple markers for differences in rank: markers of clothing, manner, and space (e.g., the physical distance between "officers" and "other ranks"). "The Army" also makes it clear very early in the course of initiation that errors in orientation—treating a superior as an equal, for example—are serious breaches of proper behavior. "The family," in contrast, is usually a greater mix of orientations and, in many circumstances, one with less clear markers of differences in rank and of the seriousness of an error. A second possibility suggested by our data is that relationship errors are especially likely to occur in situations where a change in orientation is expected to occur over the course of the relationship. Within most work situations, for example (households or paid work) a change is expected from contributions being carefully assigned and checked to work contributions being more loosely defined and needing little supervision (in effect, a move out of Authority Ranking into Communal Sharing or self-regulated Equality Matching). When the times and occasions of transition are uncertain and have to be negotiated, the probability of one person perceiving the other as "in error" (presuming too much on the one hand, holding on to privilege or authority for too long on the other) seems likely to increase. Who is most likely to experience a sense of relationship error? In many situations, there may be consensus as to what constitutes a relationship error. The only occasions of difference are then likely to be between people who are new and people who already know the customary rules. In other situations, differences between people may arise because the rules are interpreted differently or because some are more alert than others to what they perceive as a relationship error, as an orientation that they regard as inappropriate and unwanted. One result prompting this kind of base to differences among people is the sensitivity of many adolescents to whether a task is presented to them as "a request or an order." Their sensitivity seems linked to their sense of themselves as being at a possible changeover point from "child" to "adult." Relevant also is the frequent reference by Australian mothers to not being "maids," "not being paid," "not running a hotel, restaurant, etc.," not being "a child's brain" (remembering for the child what needs to be done). (I am told that the same phenomenon also occurs within families in some parts of Europe.) The sensitivity in this case seems to arise from the particular combination of mothers' actions and expectations. The mothers we interviewed hope that the contributions they make will be seen as a voluntary act of love. They do not wish their work to be seen as undertaken because they have nothing better to do, because they have no choice, or because they are building up a debt on the child's part, to be repaid later in life. At the

192

GOODNOW

same time, the relatively routine nature of what they do invites the attribution that their work contributions are "just what mothers do." In addition, many of the tasks these mothers undertake are also often performed by people who are paid to do so. The mothers we interviewed do cook for other people. They make other people's beds. They clean up after others. They may reorganize their own timetables and priorities for the benefit of others. They may then be easily assigned to the same relationship category as paid others are. Small wonder that children need to be reminded of distinctions between "mothers and maids," between "homes and hotels." Small wonder also that, in such ambiguous situations, mothers can become alert to signs that their labors of love are becoming taken for granted, seen as signs of lesser status, or interpreted as undertaken for some personal reward. One last source of sensitivity to relationship errors as a function of the position one occupies comes from data that covers not only suggestions from others about ways of managing the organization of household work but also ways of managing money, time, dress, exercise, or remembering appointments. The original expectation in this study was that the group most likely to be sensitive to unasked-for advice would be those in the oldest age group (70 plus, compared with a group in their 20s and a group in their 40s). In fact, the group reporting the highest frequency and the least pleasure in association with such advice was the young adult group: a group that apparently is still in the process of establishing its independence and having its competence recognized (Smith & Goodnow, 1999). Developmental Aspects Fiske (1991) concentrates on proposals with regard to the first signs of children acting on the basis of each of the four basic orientations. My colleagues and I have been more concerned with the circumstances and processes involved in the elaboration of that first understanding. Two comments seem especially warranted. One is the need to be specific about the phenomenon being tracked. The course of first appearance and of elaboration is likely to be different for a general alertness to the presence of an orientation (e.g., an alertness to pecking orders or to equal shares) and for some particular features of various orientations (e.g., an understanding of the particular implementation rules that are expected to apply, the extent to which they may be negotiated, the preferred, tolerated, and unacceptable styles of negotiation, and the seriousness of various rule violations or relationship errors). The second comment has more to do with process. Our results point to the need to consider two routes to change (spontaneous discovery and di-

5. THE DOMAIN OF WORK IN HOUSEHOLDS

193

rect teaching) and two possible correlates (changes in age and changes in social position). Those two correlates are often thought of as going together. Adolescence, for example, usually brings both a change in age and a change in social status (e.g., in privileges, in the type of school attended, or in obligations). The two bases of change need not, however, coincide: a point brought out especially by studies where children have the same age but are not in the same circumstances. Eleven-year-olds in a U.S. county that moves them into a middle school at the end of year 5, for example, may be compared with 11-year-olds who remain in a 6-year primary school (Alfieri, Ruble, & Higgins, 1996). Despite their similarity in age, they differ in the extent to which they question authority and gender stereotypes. Closer to the area of work contributions is a comparison of 5-year-olds (in Australia) who have started a kindergarten year in school and 5-yearolds who have not (Leonard, 1993). The comparison has to do with the sophistication of arguments offered for declining a task request. The former show more sophistication than the latter, with the difference apparently stemming from the increased demand to present a better case than one needs to present within the well-trodden interactions with mothers. Cultural Similarities and Differences Fiske (1991) proposed that all cultures share the basic four orientations. The differences lie in the way these orientations are played out in various domains. From one group to another, for example, there may be differences in the extent to which a particular area of activity attracts one orientation rather than another. There may also be differences in the extent to which the cultural implementation rules are set by tradition or are open to options and negotiation. To these proposals, and to the examples that Fiske (1991) provides, I would now add two points. One of these has to do with the need for some particular kinds of data in order to pin down the nature of a cultural difference. An example is the phenomenon already noted: the lesser acceptability in Beijing families of "Not all mine" as a response to a parental request to put away the stuff that two people have used (Bowes et al., 1999). That difference could reflect a very specific expectation: Children should not use this kind of rationale in response to a parental request. It could also reflect some larger expectations that cut across a variety of situations. Between Chinese parents and children, for example, less may be seen as open to negotiation than is the case in Australian families. Or—more broadly still—the culture in general may place more emphasis on Authority Ranking than on the autonomy and rights of individuals. Sorting out those several possible bases needs to wait for the investigation of other situations and other domains.

194

GOODNOW

The same kind of point applies to some further data on a difference among adolescents in several western countries. Given a choice of several reasons for contributions of household work, adolescents in Australia and the United States chose "the development of a sense of responsibility" more often than they chose "benefits to the family." The reverse pattern was true for adolescents in Czechoslovakia, Hungary, and Sweden (Bowes, Flanagan, & Taylor, 2001). Again, the basis may be a difference in the general level of importance attached to outcomes for the individual rather than outcomes for the family. To be sure of that, however, we would need a variety of situations that cover more than household work contributions. Is there then no particular indication of a difference in general ethos, as against a difference in some particular situations? In the course of the series of studies outlined, for example, I have become sensitive to the extent, within this particular cultural group and in this generation, of an emphasis on "gifts" and on the "volunteered" nature of contributions. I end this section, and the chapter, then with a speculative comment on the relative importance of "duties" and "volunteered actions." "The Saga of Paul," for example, is a saga of contributions construed as duties or as gifts. The concern of mothers with a particular error—having their work thought of as similar to what happens in commercial relationships—is a concern with the recognition of their work as a voluntary gift rather than as simply "their job." The same mothers, when asked to describe "good moments" in relation to their children's contributions, produce accounts of children who unexpectedly take on a piece of work without it being their job or their obligation. In their ideal form, volunteered contributions show the signs of being a well-chosen gift: That is, they show some appreciation of what mothers would actually like to have done for them. Other cultural groups, I suggest, may have less investment in work contributions being made in gift-like fashion. They are instead more likely to think and talk in terms of "duties." More speculatively, the value Australian mothers place on gifts of work may be related to a Western position often found in the domain of moral action. In that domain, a high value is often placed on actions based not on the following of custom or some clearly outlined duty but on having faced temptation and chosen to resist it, chosen to follow the less easy path as a result of an individual decision to do what is "good." Whether such values cut across more than one domain and differentiate among cultural groups, we cannot say at this point. It is, however, one of the provocative possibilities that arises when we take up Fiske's (1991) challenge to ask about the particular ways and circumstances in which people in various groups manifest and implement the four basic orientations to social life that he so richly outlines.

5. THE DOMAIN OF WORK IN HOUSEHOLDS

195

REFERENCES Alfieri, T. J., Ruble, D. N., & Higgins, E. T. (1996). Gender stereotypes during adolescence: Developmental changes and the transition to high school. Developmental Psychology, 32, 11291137. Backett, C. (1982). Mothers and fathers. London: MacMillan. Bowes, J. M., Chen, M. J., Li, Q. S., & Li, Y. (1999). Asking children to help: Culturally acceptable justifications. Journal for Australian Research in Early Childhood, 6, 9-17. Bowes, J. M., Flanagan, C., & Taylor, A. J. (2001). Adolescents' ideas about individual and social responsibility in relation to children's household work: Some international comparisons. International Journal of Behavioral Development, 25, 60-68. Clark, M. S. (1984). Record keeping in two types of relationships. Journal of Personality and Social Psychology, 47, 549-557. Clark, M. S., & Chrisman, K. (1994). Resource allocation in intimate relationships: Trying to make sense of a confusing literature. In M. J. Lerner & G. Mikula (Eds.), Entitlement and the affectional bond (pp. 65-88). New York: Plenum Press. Clark, M. S., & Mills, J. (1979). Interpersonal attraction in communal and exchange relationships. Journal of Personality and Social Psychology, 37, 12-24. Dunn, J. (1988). The beginnings of social understanding. Cambridge, MA: Harvard University Press. Fiske, A. P. (1991). Structures of social life: The four elementary forms of human relations. New York: Free Press. Fiske, A. P. (1992).The four elementary forms of social life: Framework for a unified theory of social relations. Psychological Review, 99, 689-723. Fiske, A. P., & Tetlock, P. (1997). Taboo tradeoffs: Reactions to transactions that transgress spheres of exchange. Journal of Political Psychology, 18, 255-297. Goodnow, J. J. (1988). Children's household work: Its nature and functions. Psychological Bulletin, 103, 5-26. Goodnow, J. J. (1996). From household practices to parents' ideas about work and interpersonal relationships. In S. Harkness & C. Super (Eds.), Parents' cultural belief systems (pp. 313-344). New York: Guilford. Goodnow, J. J. (1998). Beyond the overall balance: The significance of particular tasks and procedures for perceptions of fairness in household work distributions. Social Justice Research, 11, 359-376. Goodnow, J. J. (2000). On being responsible for more than you have directly caused. In W. van Haaften & A. Stellings (Eds.), Moral sensibilities and education (Vol. 2, pp. 35-60). Bettels, The Netherlands: Concorde Press. Goodnow, J. J., & Bowes, J. A. (1994). Men, women, and household work. Sydney/New York: Oxford University Press. Goodnow, J. J., & Delaney, S. (1989). Children's household work: Task differences, styles of assignment, and links to family relationships. Journal of Applied Developmental Psychology, 10, 209-226. Goodnow, J. J., & Lawrence, J. A. (2001). Work contributions to the household: Toward a distributive and social framework. In A. Fuligni (Ed.), Family assistance and obligation during adolescence: Contextual variations and developmental implications (pp. 5-22). San Francisco: JosseyBass. Goodnow, J. J. & Lawrence, J. A. (2003). The relational significance of inheritance arrangements. Unpublished paper, Macquarie University and University of Melbourne. Goodnow, J. J., Lawrence, J. A., Ryan, J., Karantzas, G., & King, K. (2002). Extending studies of collaborative cognition by way of caregiving situations. International Journal of Behavioral Development, 26, 16-26.

196

GOODNOW

Goodnow, J. J., & Warton, P. M. (1991). The social basis of social cognition: Interactions about work and lessons about relationships. Merrill-Palmer Quarterly, 37, 27-58. Goodnow, J. J., & Warton, P. M. (1992). Understanding responsibility: Adolescents' concepts of delegation and follow-through within the family. Social Development, 1, 89-106. Haslam, N. (1994). Categories of social relationship. Cognition, 53, 59-90. Lawrence, J. A., Goodnow, J. J., Woods, K., & Karantzas, G. (2002). Distributions of caregiving tasks among family members: The place of gender and availability. Journal of Family Psychology, 16, 493-509. Leonard, R. (1993). Mother-child disputes as arenas for fostering negotiation skills. Early Development and Parenting, 2, 157-167. Lind, E. A., & Tyler, T. R. (1988). The social psychology of procedural justice. New York: Plenum. Mikula, G. (1998). Divisions of household labor and perceived justice: A growing field of research. SocialJustice Research, 11, 215-241. Rheingold, H. (1982). Nascent prosocial behavior. Child Development, 53, 114-125. Shure, M. B. (1968). Fairness, generosity, and selfishness: The naive psychology of children and adults. Child Development, 39, 875-886. Smetana, J. G. (2000). Middle-class African-American adolescents' and their parents' conceptions of parental authority and parenting practices: A longitudinal investigation. Child Development, 71, 1672-1686. Smetana, J. G., & Asquith, P. (1994). Adolescents' and parents' perceptions of parental authority and adolescent autonomy. Child Development, 65, 1147-1162. Smith, J., & Goodnow, J. J. (1999). Unsolicited support, unasked-for advice: Age differences in interpretation and affective response. Psychology and Aging, 14, 108-121. Warton, P. M., & Goodnow, J. J. (1991). The nature of responsibility: Children's understanding of "your job." Child Development, 62, 156-165. Warton, P. M., & Goodnow, J. J. (1995). Money and children's household jobs: Parents' views of their interconnections. International Journal of Behavioral Development, 18, 235-350.

C H A P T E R

6 Hidden Bias: The Impact of Relational Models on Perceptions of Fairness in Human Resource Systems Debra Louis Connelley Robert Folger

In recent years, the topic of managing a diverse workforce has garnered much attention in management journals and trade publications. Ranging from "how to" articles on recruitment and retention, to academically oriented empirical and theoretical papers, the management literature has striven to understand and document the special and unique characteristics of managing heterogeneity. Most of these efforts have focused either on cataloging differences between racioethnic groups and/or genders, studying the implications of relational demography, or diagnosing the causes and effects of intergroup conflict (see Nkomo & Cox, 1996 for a review). To date, however, no one has looked at how diversity affects the social construction of relationships between organizational members and the structures that govern organizational operations. Relational models theory provides a valuable tool for probing the relationship between diversity and organizational relationships, policies, and procedures. In this chapter, we examine how the gender and race of organizational members affects the way they experience and interpret the organization around them. We begin by proposing a general model for how diversity can shape variations on organizational styles, behaviors, practices, norms, and traditions, in the light of Fiske's four relational models. In the second part of the chapter, we consider how we, as scientific observers, interpret the effects of these models and diversity on the design and practice of human resource management in one large U.S. firm. Finally, we offer fur197

I 98

CONNELLEY AND FOLGER

ther commentary on the implications of relational models in organizations of all types.

RELATIONAL MODELS

Classical sociological theory suggests several typologies of social order in society. Tonnies (1887/1988) suggest that social order was based on form of production, while Durkheim (1893/1933) argued that it was based on division of labor and Weber (1946) related social order to ideology. Modern sociologists and social psychologists have continued to explain organizations and human behavior by analyzing the bases and dynamics of social relationships. Fiske's (1991) theory of relational models attempts to integrate extant theories into a unified theory of social relations. Relational models theory (RMT) posits that there are just four elemental models for the construction of all social relationships. These models provide the scripts or schemata that allow individuals to anticipate and relate to the behavior of others. Although a more extensive discussion of relational models can be found in chapter 1 of this volume, we provide a brief overview here. The four models are: 1. Communal Sharing CCS)—This is a relationship based on the concept of equivalence. It is a kinship model, broadly defined. To individuals operating under a CS model, the common group identity is important, not the individual concept of self. According to Fiske (1992), "the members of a group or dyad treat each other as all the same, focusing on commonalities and disregarding distinct individual identities" (p. 690). Ingroup-outgroup boundaries are salient, and regular sharing of information, resources and support occurs within the group, but not across group boundaries. Ceremonies and rituals (such as taking meals or recreation together) are used to reinforce the common bonds. This model relies exclusively on a nominal metric whereby categorical relations are relevant. An example of this type of model would be fraternal organizations. 2. Authority Ranking (AR)—This model is based on a fundamental assumption of ranked differences between actors. Status is ascribed and powers as well as expertise are attributed to rank. Subordinates are expected to be deferential, obedient, and loyal while benevolent superiors are expected to adopt a policy of noblesse oblige. The relevant metric is ordinal as relationships are interpreted in terms of hierarchies and rank ordering. Most families and western organizations are largely structured according to an AR model. 3. Equality Matching (EM)—EM relationships begin with the assumption of egalitarianism. Differences are acknowledged, but individuals are consid-

6. HIDDEN BIAS

199

ered to be co-equal. Contributions and influence match one-to-one, and there is no concept of profit, privilege, or advantage. Turn-taking is one preferred way of organizing activity. EM utilizes an interval metric; the magnitude of imbalance between what is given and returned is easily assessed and it is clear how much needs to be added or subtracted to correct an imbalance (e.g., If your car is in the shop, you know how many turns in a car pool you've missed and how many you will have to make up to get "even"). An example of EM is the United States Senate, where each state is allocated two representatives, regardless of size. 4. Market Pricing (MP)-MP relationships are based on the notion of utility, price, or value. Status is achieved or earned by merit. Entry into MP systems is open to all competent participants. The MP model operates according to a ratio metric; value is assessed by cost-benefit or input-output analysis. The U.S. Congress, where the number of representatives is determined proportionally according to the size of the state's population, is an example of an MP model. Fiske argues that these models are combined in different ways in various societies and domains (i.e., work vs. familial relationships). In essence, they form the fundamental building blocks of human relationships and although they are rarely found in "pure types," often one model or another will be dominant in a given society and domain. Thus, for the most part, members of the same social system will agree on which model to implement in a given situation. Fiske and Tetlock (1997) propose that people make decisions about which model to use based on ideology, precedent, or tradition, and/or situational context (members of a co-op will use consensus as a decision rule). Although sometimes the employment of a particular relational model is straightforward, such as the use of seniority to determine holiday work assignments, raises, and promotions, there are two general cases where multiple models may serve to govern different aspects of the same relationship. In some cases, models may be horizontally "linked" (Fiske & Tetlock, 1997) so that relationships may be comprised of a combination of two or more models governing different types of interactions within a single domain. For example, among faculty at a university, resource allocation may be governed by several different models according to the specific resource in question. Paper clips and rubber bands may be treated as common property, with each faculty member taking as much as he or she requires without anyone keeping count (CS), while each member is given one and only one parking permit (EM), long distance or photocopying is charged according to actual use (MP), and office size/location is determined by rank (AR). The second case where multiple models are brought to bear on relationships within a single domain is where models are vertically "nested." In this

200

CONNELLEY AND FOLGER

situation, there is a primary model that sets the abstract rules for a system, but a different model determines the actual implementation rules. For example, in a law firm, one achieves partnership based on an attorney's ability to make rain for the firm. In this sense, the decision is an MP one. However, the ability to bring business to the firm is determined in large part by CS contacts who give the attorney greater access to potential clients. Thus, although partnership decisions appear on the surface to be a pure MP decision, they actually rest heavily on CS relationships in practice. Fiske and Tetlock (1997) argue that there is nothing implicit in the relational models to determine which model should be employed in any given situation. Thus, it is not uncommon for conflict and confusion to exist over which model to apply or how it should be implemented (Whitehead, 1993). If different parties prefer different models to govern allocation decisions, for example, such confusion is likely to result in perceptions of the rules as unfair. Indeed, Tornblom and Jonsson (1985) found that when implementation rules for one model are applied to a context where another model is preferred (e.g., AR rules in an MP domain), the perceived inappropriateness of the rules accounts for the largest portion of fairness judgments. The injustice that is experienced when disparate nested models are used is particularly subtle and hard to diagnose. In such cases, people may agree at the abstract level, so that ostensibly there is consensus, but have differences at the level of implementation rules. For example, a faculty may be in complete agreement that it is important for all members to have an equal say in assigning graduate assistants (EM). The implementation rule suggested as fairest may be one-person-one-vote. But is this really an EM rule if tenured faculty outnumber junior faculty by four to one? Might the senior professors band together to vote the most talented students to themselves (CS/AR)? If the junior faculty cannot exercise equal influence because of their smaller numbers, doesn't a one-person-one-vote rule really function as MP? Is it not likely that the faculty in the minority would feel that the procedure for assigning graduate assistants was patently unfair to them? Might they not think a "true" EM implementation model, say a lottery or turn taking, would be a fairer way to allocate graduate assistants because then everyone would have strict equality in terms of outcomes? This line of reasoning would suggest that when organizational members are met with a relational model (in the form of rules, policies, or procedures) that favors another party or puts them at a disadvantage, they would react with assessments that the system is unfair. In response, they will advocate an alternative model (employing different behaviors, procedures, and decision rules) that allows them to succeed and is therefore seen as being fair. In the next section of this chapter, we take a closer look at this issue by examining how differences in implementation rules for nested models in human resources systems create conflict and feelings of

6. HIDDEN BIAS

201

injustice between White males, White females, and African Americans in a Fortune 100 company. In particular, we demonstrate that these judgments of unfairness are accompanied by preferences for alternative systems that reflect different relational models.

THE CASE OF CORONA CORPORATION

Corona Corporation (our pseudonym) is a Fortune 100 manufacturing company headquartered in the United States but with operations worldwide. Historically, the company hired mostly White males, the majority of whom had engineering or other technical degrees. There were few "management specialists"; people were promoted into management based on their success in technical areas, receiving little, if any, management training. The company never stressed formal human resource management systems, instead many informal systems developed over time to fill the needs for coordination and control. These informal systems reflected a corporate culture that stressed interpersonal cooperation, a strong affiliation with the company, and conformity to norms governing behavior. Social networks within the company met much of the responsibility for mentoring, feedback, and career development. In the early 1990s, the company began a push to increase the diversity of its employee base, largely in response to published reports concerning the changing demography of the workforce. They quickly realized, however, that simply hiring greater numbers of women and minorities was not the answer. The turnover rate for these new hires was unacceptably high. The company knew from exit interviews that African-American and White female employees were leaving because of perceived prejudice and the belief that they were blocked from contributing their skills and expertise to the company. Many stated that they felt they had no future there. In addition, a climate survey revealed a shocking disparity in the way White males, White females, and African Americans viewed the company. White males saw the work environment as open, friendly, rewarding, supportive, and fair. African Americans, on the other hand, saw the company as closed, hostile, punishing, evaluative, and unfair. The responses of White women indicated that they saw the climate of the company less positively than White men, but not as negatively as Black employees (see Connelley, 1994). In an attempt to discover why White men, White women, and AfricanAmericans viewed the company so differently, Corona convened a series of focus groups. Content analysis of the transcripts from the focus groups revealed widespread tension and dissatisfaction within and between the three groups. Much of this conflict centered on human resource management systems, particularly those dealing with hiring practices, performance

202

CONNELLEY AND FOLGER

evaluation, and promotion decisions. Although White males seemed to be for the most part comfortable and satisfied with these systems, White women and African-American employees expressed frustration and anger with systems that they saw as biased and discriminatory. Hiring, evaluation, and promotion systems are essential to any organization in determining what new members get admitted, how their work is evaluated and rewarded, and how people are selected to move into positions of greater authority. In terms of relational models theory, they are fundamentally AR systems, but they can be implemented using any of the four models. In the section that follows, we show how the implementation rules for human resource practices at Corona flow from a CS model, how White female and Black employees view these systems as biased and discriminatory, and cite alternative systems they argue would be more fair. Their favored systems are shown to reflect MP and EM models. Communal Sharing at Corona Before we turn to the specific human resource practices, it is important to understand how the CS model works at Corona. One of the key issues for CS groups is the criteria by which people are judged as "us" or "other" (Fiske, 1991). At Corona, the prevailing definition of "us" is quite narrow: White, male, with a technical background. In the words of one employee: • We tend to work better with people similar to ourselves. This is still a White male company and [we] are still calling the shots. (White male)

This means that White women, African Americans, ethnic minorities, as well as White males with a nontechnical background are, by definition, assigned outsider status. The point is critical to understanding why White women and African-American employees view the human resource practices as unfair: In the CS model, information, resources, and support are freely shared within the ingroup, but are not extended beyond that boundary. Being seen as an outsider excludes one from full membership in the system: • Black employees feel like outsiders. They are reminded by coworkers, sometimes, that they are outsiders. (African American) • Even after 18'/2 years, I still don't feel accepted. When a person is perceived as not fitting in, they will try to run you off. (White female) Hiring Systems

Historically, Corona has hired from the local community as well as a small number of exclusive technical colleges that form the company's traditional recruiting pool. Because the rural areas where the company locates

6. HIDDEN BIAS

203

its operations are overwhelmingly White, and the colleges do not attract large numbers of women or minority students (especially in engineering and technical areas), this ensures that most of the new hires will continue to be similar to the existing workforce at the company. According to Tolbert (1988) one way that organizations acquire workers who share common values is through educational homogeneity (defined as the number of organizational members who attended the same school). This practice produces employees who "share interpretive frameworks created by common backgrounds and experiences" (Tolbert, 1988, p. 105). Therefore an exclusive, limited hiring pool reinforces the sense of unity and solidarity typical of a CS model. Over time, a strong ingroup has been developed and maintained at Corona. Hiring outsiders is perceived as a threat to the security of the CS ingroup. For example, in an effort to increase the demographic diversity of the workforce, the company began to hire people from outside the local area. This was taken by many employees, particularly White males, as a violation of their CS norms: • We have to import minorities, because there are not enough around the valley. This makes it difficult for people from the area to try to get a job with [the company], especially for people out on the floor. (White male) • A lot of men resent the diversity initiatives. Their sons are not being looked at because women are being looked at instead. They hold women personally responsible. (White female)

Fiske (1991) argued that when threatened, the CS relationship is capable of great violence toward outsiders. Although bloodshed or genocide is not acceptable in a business setting, aggression against outsiders is not unknown. Some "socially acceptable" forms of aggression include actions that exaggerate the extent of the threat to the ingroup, denigrate or disparage the competency and skills of the outsiders, and/or deny recognition of the value of the contributions outsiders make. At Corona, White males complain that White men cannot get hired, and are concerned that the company is lowering standards to hire women and minorities. This belief is held despite company statistics that show many more White males continue to be hired than women or minorities. The following quote from a White male shows how the threat has been exaggerated, and how skills and contributions of outsiders are being devalued. • The perception is that [the company] is creating jobs for women and minorities that don't contribute to production. Having targets has resulted in hiring people who are "qualified for the job " rather than being "the best of the best." [I've heard stories that] managers are asked to only hire Black employees, not to look at

204

CONNELLEY AND FOLGER

White resumes; they have to hire a poorly qualified Black and can't even look at White candidates. There is so much emphasis on hiring minorities and women that people asked me when I was hired, "How did you get a job [there]? (White male) Performance Evaluation

In the organizational field, we have identified two key components to the performance evaluation process. The first is informal feedback on job mastery, the second is the formal evaluation procedure. Both formal and informal practices reflect a CS model and are characterized by the free and open exchange of information within the ingroup—but not to outsiders. The formal procedure shows a particularly heavy dependence on the CS norm of solidarity. Informal Feedback. Most day-to-day feedback is received through informal networks. This is where the employer is reinforced for a job well done, and corrected for poor performance or inappropriate behavior. In the words of one White male: • [To be successful]. . . you need to have a network and build constituencies. The network needs to be able to coach, teach, and mentor you. Success depends on . .. how compatible you are with those who can offer that learning to you. (White male)

In keeping with the CS model, these networks are quite homogeneous and it is difficult for women and minorities to gain acceptance in these networks because they are viewed as outsiders. They are by definition not "compatible" with "the Club" or "old boy network" and thus do not receive the same feedback and coaching as do their White male coworkers. The following comments illustrate this dilemma. • The person who is a minority misses being in "the Club". . . made me aware of why feedback doesn 't come. lam one woman in the group; they didn 't want me to be a part of the group. (White female) « . . . even though [the company] talks "Diversity," they still want to have their exclusive Country-Club image and "old-boy" White network. Blacks are easily excluded. (African American)

The extent to which outsiders are isolated in CS relationships is succinctly and eloquently communicated in the following comment: • / know two Black guys are putting in their best effort, but nobody will help them, teach them. Everybody—supervisors, peers—looks down on them. (White female)

6. HIDDEN BIAS

205

This isolation is not only enforced through exclusion from the informal networks, women and minorities also report that supervisors are not comfortable with them because they are "different," and do not respond to them as they do to other White males. Specifically, they assert that they cannot get job feedback from their own supervisors: • [This] is a company where Black employees keep hearing stories indirectly about their unsuccessful performance. It is hard to get any negative feedback [from superiors]. You have to put words in their mouths, and even then it is hard to get them to agree with it. (African American) • Women and Blacks spend more time trying to figure out how to fit given our obvious differences. . .. It's hard for four managers] to give feedback. If the subordinate was more like them, it would be easier to relate to... them and give advice. (White female)

Formal Feedback. Formal performance reviews involve the use of an instrument we call the FPA (formal performance appraisal). A common practice in preparing the document is to circulate a form to colleagues whom the supervisor trusts, asking for their assessment of the individual being evaluated, how they get along with others, and so forth. Indeed, the ability to get along and conform to group expectations and norms is a very important part of the review. Interestingly, the colleague-reviewer does not even have to be familiar with the work of the employee. What matters is how the reviewer's impressions match the group consensus about how well the employee fits in. • Managers do assessment by group consensus and perception. People will often ask vague, open-ended questions about an individual to find out more about how they work, and what their style is, etc. It's amazing, but the people who have a great deal of influence in my FPA and my pay raise do not have a clue about what I am doing or how well I do it. (White male)

This method of gathering information on the employee demonstrates the CS model's preference for consensus, unity, and desire to be similar to others. Thus, even though the White males admit that the reviewer may not have much knowledge of the employee's performance, they still accept the FPA as providing accurate and important information. • The [FPA] process helps manage the feedback you get from others about how you are involved and get along with others. The feedback is honest and usually thorough. (White male)

However, if the people providing input on an outsider are themselves insiders, their perceptions are likely to be distorted or at least subject to se-

206

CONNELLEY AND FOLGER

lective attention biases. Contrast the views of outsiders with the above insider. • Managers measure success by group consensus. Manager's perceptions are not [fact-based]. There's too much reliance on hearsay. (African American) • In reviews you hear, "Somebody said this about you" and it's [assumed to be] true. They don't know. (White female) Promotion Policy

Many personnel matters are kept secret at Corona. Grade levels and salaries are not discussed. Nor are recipients of cash awards for performance made known. Reasons for promotions and firings are obscure. Yet this situation seems to concern White males less than the other groups, perhaps because the large numbers of White males in higher management positions allows White men to take promotions for granted. Although they aren't specific about it, they seem to have a "sixth sense" about the process of promotion, something that is not shared by the female and Black employees: • Promotions are promised all the time, and you can see them coming from so far away that you know how good or bad your chances are. People are selected early and then treated favorably. (White male) • People are selected for promotion through an inscrutable process. (African American) • We don't know how men get promoted, how women get promoted. (White female)

One thing that is clear is that interpersonal relationships play a major role in determining who gets promotions; and these decisions hinge on consensus and similarity. This involves relationships with co-workers, a sponsor, and the immediate supervisor; smoothing the way, increasing visibility, and opening doors. • You know that some people will go up the ladder regardless of how well or poorly they do here. They have a godfather somewhere in [the company]. Sometimes it's because they went to the right school. How you grow and how much you grow is dependent on how much your senior manager supports your growth. (White female)

This emphasis on interpersonal relationships, nurturance, and maintaining continuity is characteristic of a CS model. It is in keeping with this model that when promotions are made for different reasons, such as pro-

6. HIDDEN BIAS

207

moting underrepresented groups into management, the White males respond with the same aggression as with hiring practices: • Some perceive that Diversity is taking away opportunities for White men and that seems unfair. The wrong people have been promoted.. . because of Diversity on the managerial side. Some careers are moving simply for diversity reasons instead of being qualified. (White male)

Alternative Models So far, we have indicated that human resource systems at the company reflect a CS orientation. Each of the elements described earlier demonstrates that the overall effect of Corona's human resource system is to maintain and reward a workforce composed of people who are "just like us." Put succinctly by one employee: • 7776 problem with [the company] is that we really are not interested in having a diverse workforce, in terms of behaviors, styles, etc. [We] will have Blacks and women, but only if they act like White males and demonstrate conformity in their behavior. (White male)

The women and African Americans at Corona realize that they can never be members of an ingroup that is defined by gender and race. As a result, CS implementation rules for human resource systems will not work for them because they will always be seen as outsiders, and therefore unable to succeed. They see policies that reflect the CS model as being at best unfair and at worst deliberately discriminatory. As a result, they express a preference for different relational models in implementing Corona's human resource systems. White Female Employees

The next section demonstrates that the White women in this company are operating out of a Market Pricing (MP) model. They are continually asking that they be hired, evaluated, and promoted based on an objective assessment of their qualifications and competence. They are seeking judgment based on their "market value" or a measurement of their worth to the company. Hiring Systems. White female employees protest that hiring systems emphasize "fit" as a primary consideration before skills. According to one employee: • Following interviews, the first question is, "Will they fit?" Then you talk about technical skills. (White female)

208

CONNELLEY AND FOLGER

In the CS implementation model, everything hinges on consensus, cooperation, and conformity. In such a case, "fit" would be a primary consideration in hiring—homogeneity is essential and is maintained by the policies of hiring from the local area and recruiting from traditional pools. However, these women feel that hiring decisions should be based on the candidate's abilities and how much those proficiencies are needed by the company—an MP orientation. To this group of employees, similarity, or "fit" should be a secondary concern. We have argued that the hiring targets instituted by the company fly in the face of CS norms and values. As Whitehead (1993) and Fiske and Tetlock (1997) would predict, these "inappropriate" procedures generated great hostility and considerable backlash on the part of White males. As adherents to an MP model, White females don't necessarily support numerical targets either. Their position is somewhat different, however, from White males who view the targets as a zero-sum game (e.g., more qualified White males are losing out to incompetent women and minorities). White women may not like the targets either, but it tends to be because they feel the targets increase hostility and undermine the perception of their competence. They worry that others will see them as being hired only to fill a "quota": • It's obvious that certain jobs are targeted for women or Blacks and you go into meetings at the corporate level and it is degrading to be one of three women in the room and they are talking about the numbers.. .. The perception is that they are brought in because of the numbers, not competence. It reinforces the hatred of the White men. You can feel the tension. (White female)

Performance Evaluation. When discussing Corona's system of performance evaluation, we indicated that White females complain that it is hard to get feedback from supervisors and that they are evaluated on frequently incorrect perceptions. They seem to be asking for a list of objective criteria that they could strive to meet, and then they could be fairly evaluated on their success in attaining concrete goals. Following the MP model, they would prefer to be evaluated mainly on "hard data" such as results or output, but are dismayed that this is a very small part of their assessments: • They definitely have different expectations [for women and minoritiesJ. To the extent that they measure you on results, it is the same fas men]; but that is only 2% of how they evaluate you. (White female)

Promotion Policy. Finally, White female employees are more likely to advocate a system of advancement based on meritocracy. They get very anxious when discussing promotions, and once again seem to be looking for specific guidelines that they feel did not exist. The women are clearly

6. HIDDEN BIAS

209

frustrated by the vague criteria that seem to be understood by the White men, but no one else: • Men take promotions for granted. We don't know how they got there. What skills did they need? We don't know what the policies are. If there was more talk about this we might be able to break through. (White female)

In an MP implementation model, there would be a strong preference for clear, pragmatic, and public guidelines for achievement. The degree to which one was able to achieve these goals would then be used as the basis for promotion decisions. This seems to be what the women at Corona are asking for. African-American Employees

Like the White women, Black employees at the company reject the CS model, labeling it hypocritical and deceptive. However, in contrast to the MP policies preferred earlier, they draw consistently on an EM model. Hiring Systems. In hiring and placement decisions, the Black employees emphasize the importance of equal access and opportunity. Like the White women, Black employees know that they are viewed as being less competent than their White male peers. • It is commonly thought that Blacks got their jobs because of their race; that they are not really qualified, that they are going to screw up. (African American)

However, whereas the White women see this as a result of the hiring targets, the African Americans see this as a result of institutional racism. They literally feel that they are seen as being less qualified solely because they are Black, and, as such, will never be judged according to their true value. • Black employees are not treated the same as others . . . White managers have lower expectations . . . Blacks are often not given credit for their expertise. .. . They will often go outside to confirm your input. A graduate of a Black school has an even tougher time gaining credibility. (African American)

Indeed, whereas the White males see the company as lowering standards to hire minorities, the Black employees argue they are overqualified for the positions in which they are placed. • Minority candidates continue to be overqualified. [The company keeps] bringing in experienced midcareer minorities at the wrong level, in the wrong jobs, andexpect(s) them to stay. (African American)

2 I0

CONNELLEY AND FOLGER

Interestingly, although Black employees agree with White females that targets can result in backlash, they are much more likely to support hiring targets because they think that is the only way the company will hire minorities at all. In accordance with an EM implementation model, they feel that because their skills are underrated, institutionally mandated programs that enforce equal prospects for employment are necessary. Furthermore, targets are seen as a "starting point," a leveling of the field that must exist as a precondition to further opening of equal opportunities. There is a strong feeling that the company should be held accountable for meeting these targets and overcoming resistance. • Four Black female engineers were recently laid off. That shows no commitment to diversity. No, [the company] doesn 't walk its talk. [It] has repeatedly missed its [hiring] targets f o r . . . Black employees. [Diversity target] numbers should be the minimum, not the maximum. Diversity for all levels! Now! Not 10 years from now! (African American)

Performance Evaluation. We pointed out earlier that Black employees receive little informal feedback due to exclusion from networks and the unwillingness of superiors to be candid and direct. However, when they are given formal feedback, African-American employees state that they are not judged by equitable criteria. They protest there is a double standard, and that they are scrutinized more closely than other employees. • Managers use a different yardstick for Blacks in defining success. Black employees need to do twice as much as their coworkers. Blacks are always under the spotlight and scrutinized. There is a magnification of any small problem that involves a Black employee. [This] is a company where Black employees are not forgiven of their early mistakes, as are their White peers. (African American)

Essentially, they are saying that institutional bias and prejudice proscribe fair evaluation for Black employees. Unlike White females with their preference for the MP implementation model, they have little faith that individual effort will be fairly recognized and rewarded. Instead, an EM model would maintain that strict equality of treatment would be institutionally enforced. This would insure that all employees started out on an equal basis and were judged by identical standards, so that evaluation could be impartial and fair. The emphasis is on strict equality, not on a ratio calculated according to individual circumstance. Promotion Policy. Concerning promotion policy, the focus is again on the problem of parity (i.e., equal opportunity). Here too, there are frequent comparisons pointing to the imbalance between the experiences of White and Black employees.

6. HIDDEN BIAS

2 II

• Black employees are kept in the same jobs rather than being given an opportunity. Black males have a slow painful track for promotions, but this is even slower for Black females. This is, however, quite different for White men and women. (African American)

As was the case with hiring and performance evaluation, African-American employees at this company feel that they do not start from a level playing field because of assumptions of incompetence. Furthermore, because of institutional racism, they felt strongly that these assumptions would not be overcome no matter how good they were. Therefore, the preference for EMstyle institutional mandates offers an alternative implementation model that guarantees equal access to promotions. Discussion

In the beginning of this section, we noted that White males, White females, and African Americans hold very different views of Corona Corporation. The White males regard the company most positively, the White females have a less favorable image, and the African Americans hold the most negative view. We argue that part of the reason for these different views is that the groups experience the organization as more fair or unfair depending on how they see themselves affected by the dominant organizational norms and systems. The data from Corona demonstrates the ways in which the norms of a dominant relational model (CS) become institutionalized in organizational systems controlling hiring, performance evaluation, and promotions, with implementation procedures depending heavily on judgments of similarity to an ingroup. Interestingly, as a part of a diversity management program, the company tried to open the systems to diverse employees by creating procedures that did not depend on judgments of similarity. These new structures clashed with the dominant model and created conflict. IMPLICATIONS FOR THEORY AND RESEARCH

Using relational models theory to interpret events at Corona brings a new perspective to the field of organizational justice. Next, we describe these implications by focusing first on distributive norms, then broadening the scope of discussion to encompass a wider range of issues. Distributive Justice Norms Although some of the work on organizational justice can be traced to equity theory (Adams, 1965), scholars have often noted that equity is not the only norm of fairness for the distribution of outcomes such as employees' com-

2 I2

CONNELLEY AND FOLGER

pensation. In the title of a well-known article, for example, Deutsch (1975) referred to the norms of equity, equality, and need. Adams (1965) described the equity norm in terms of two ratios: one representing a target or focal outcomes-per-inputs return rate (e.g., target employee's rate of compensation such as in dollars-per-hour), and the second a relevant and comparable ratio of outcomes-inputs used as the standard or frame of reference for judging the target ratio's fairness (e.g., compensation of a similarly qualified employee doing related work, or other sources used to derive a standard indicating a "fair day's pay for a fair day's work"). Obviously the use of ratios fits the MP model. Equality, a second distributive norm, also has a clear parallel with a relational model, namely EM. According to Fiske (1991), EM distribution rules involve the division of resources into identical shares, or if the good is not divisible, equal rights for everyone to have access to the resource via turn taking or lottery. The norm of distribution according to need can be seen as an outgrowth of applying CS. For example, Fiske describes people in CS relationships as treating resources as a "shared commons" available to the group as a whole or to individual members freely as required. What does the perspective of relational models contribute to the study of distributive justice norms? First, note that if MP, EM, and CS imply the norms of equity, equality, and need, then viewing such norms as an outgrowth of relational models focuses attention on those models as the root source of distribution norms. Such an outlook may yield new insights regarding when each norm is likely to be invoked—an issue that past research and theory has left unresolved. We argue that by considering the circumstances antecedent to the use of a given relational model, scholars may discern new grounds for predicting when employees will invoke a given distribution norm as the fairness standard for compensation and other workplace outcomes. Our description of comments by employees at Corona illustrates this application of relational models theory to the thinking about distributive norms. Note, for instance, that the strong, family-like bond of a common identity shared by White males helps explain why they would view a diversity policy in hiring as a threat to their needs. Put another way, they viewed the possession of jobs as the outcome and viewed the potential loss of jobs (when women or Blacks are recipients) as inconsistent with the identity relation. Possession of jobs (obtaining them, keeping them, advancing in them) thus came to represent an issue conceptualized in inclusion-exclusion terms rather than as a calculation of outcome-input ratios. Such a viewpoint would explain why White males saw nothing amiss with informal evaluations based on whether someone "fit in" rather than the more formalized type of assessment consistent with the calculation of ratios required by an MP model.

6. HIDDEN BIAS

213

White women, on the other hand, found the MP model appealing as a way to circumvent the customary practice of awarding jobs, promotions, and pay increases on CS grounds. We speculate that calling for MP is a standard way of confronting CS, on grounds that Fiske's theorizing helps make salient. CS, according to Fiske, models social relations around sources of a common identity such as kinship. This means that CS encourages ingroup partiality and favoritism rather than the impartiality required by an MP calculation of ratios based on measurable attributes (e.g., merit-based contributions). When an entrenched ingroup fosters the use of CS, therefore, MP should figure prominently as an alternative source of distribution norms favored by some of those who consider themselves to be in the outgroup, in this case White women. We aver that these White women favored the MP model because they possessed a relatively strong sense of group efficacy (see Connelley & Goodrick, 2004), which refers to a collectively held belief regarding the ability of members of a given group to accomplish objectives and influence outcomes (Gibson, Randel, & Earley, 2000). There is evidence that gender, race, and ethnicity affect the perceived efficacy among members of different demographic groups, and that American Caucasians have a significantly stronger perception of personal control over outcomes than African Americans (Gurin, Gurin, & Morrison, 1978). Thus, White women at Corona called for human resource systems that reward individual skills and competencies because they felt that if they knew the standards, they would be able to achieve them and be rewarded on their merits. Black employees also opposed the dominant use of CS, but they looked to EM as a source of alternative distribution norms rather than MP. The exact reasons for this difference between women and Blacks obviously cannot be known with certainty, but Gurin et al.'s findings provide a basis for speculation that extends beyond considerations typically identified in the literature on organizational justice. In that literature, calling for equal outcomes as a distributive norm has been most often explained in functional terms. That is, distributing outcomes equally is said to function as a method for promoting harmony and solidarity (cf. Deutsch, 1975). But an equality norm is also a way of ensuring fair distribution of resources regardless of power to influence the outcomes. For example, the turn taking in EM makes moot any differential in ability to influence distribution. Conversely, MP and CS models are predicated on differential abilities to influence input-output ratios or gain access to people, resources, and information. Fiske's anthropological perspective supports our argument that groups with a lower sense of efficacy will support an EM model, as comments by Whitehead (1993) indicate. She notes that "vendettas are governed by the EM orientation as well," for example, and continues as follows: "It is common for this orientation to prevail between tribal groups, corporate kin

2 I4

CONNELLEY AND FOLGER

groups, or any assemblage of social peer units who must maintain a relation but something less than a trusting relation" (p. 323). Perhaps Black distrust of institutional racism, therefore, contributed to the call for EM at Corona. Black employees, believing that institutional racism would prevent them from being judged fairly no matter how well they performed, felt that policies guaranteeing equal treatment were most just. An EM model would meet this call for strict even-handedness, using across-the-board criteria that could not be biased by prejudices beyond the individual's control. Notable in this regard is that Fiske defines EM mathematically as the relation that allows for the operations of addition and subtraction. As Whitehead (1993) puts it, "Is the other party... ahead or behind oneself by one turn or two, three pigs or four ... ?" (pp. 325-326). This seems consistent with a view that "it's our turn now" and an emphasis on noting the gaps or discrepancies (cf. subtraction) between hiring targets and actual numbers of minorities hired. Thus, members of high versus low efficacy groups will judge different models to be fair, as each model engenders different assumptions concerning degrees of outcome control (see also Sondak, 1998). Depending on the degree of self-appraised group efficacy possessed by members of the different groups at Corona, we expect to see different models invoked, ranging from a more individualistic, self-reliant approach (MP in the case of White women at Corona, although MP need not always be individualistic—see Fiske, 1991) to one that focuses on institutionally mandated factors (EM in the case of Black employees at Corona). Finally, we note that Fiske's inclusion of AR also helps point out considerations that have tended to be overlooked by traditional treatments of distributive norms in the organizational justice literature. Indeed, it is notable that the equity-equality-need triad completely omits any distributive principle related to a linear "pecking order" of authority (AR) so robustly prevalent not only throughout the animal kingdom but also in all walks of life throughout the world. Fiske's analysis is a helpful antidote to this lacuna in several respects, as we discuss next. First, as mentioned, examples of distribution according to linear orderings such as rank are very infrequently even mentioned in the literature on organizational justice (one possible exception being a discussion of justice according to power, yet even it does not denote a specific distributive norm [Walster, Walster, & Berscheid, 1978]). In contrast, Fiske (1991) is quite explicit in describing an AR distributive-justice norm of allocations by rank: Persons with higher rank get more and those with lower rank get less (including sometimes what is inferior or the left-overs). Moreover, those at the upper levels of a hierarchy generally have more choices in many domains of social life. Second, Fiske extends the AR discussion of fairness norms to the domain of procedural justice by indicating the common practice of granting those

6. HIDDEN BIAS

2 I5

at the top the option of authoritative fiat (command by decree). Indeed, the scope of hierarchical authority over procedures can encompass a large range of decisions, as a leader's wishes, instructions, and the like get transmitted downward throughout an entire chain of command. These sorts of top-down (AR) institutionalized practices, common to the modern bureaucracy, often create an overall impression about the basic form of the organization (i.e., this single feature becomes the salient basis for the impression about how to characterize or classify it—for instance, as "an AR-type of organization"). We want to caution about one possible point of confusion regarding what might seem at first blush like a parallel between such an overall characterization ("it's AR") and a seemingly related term from the literature on organizational justice—"systemic justice" (Sheppard, Lewicki, & Minton, 1992). Systemic justice was described by Sheppard et al. as an overall fairness judgment regarding an institution such as an organization— that is, as a global assessment about the overall fairness of the organization as a whole and on the average, etc. (e.g., "the organization where I work is fair"). In the next section we use Fiske's perspective on one type of part-whole relationship, however, whereas systemic justice refers to another type of part-whole relationship. Systemic justice refers to an overall or global fairness assessment made by somehow combining (taking into account) all the specific fairness assessments associated with different aspects of one's situation. A given employee might decide that his or her company is "fair" overall simply because his or her own pay seems fair; more likely, however, he or she might take into account a wide variety of fair and unfair experiences with respect to a range of issues (e.g., from the fairness of decision making procedures about drug-screening policy to the fairness of being shown respect and dignity by supervisors). Forming an overall judgment about the "whole" (the company) on the basis of those "parts" (e.g., specific incidents) is analogous to the situation in which work satisfaction questionnaires might have both "facet" items (e.g., satisfaction with opportunities for promotions; with the degree of autonomy regarding how to perform one's job) and "global" items (how satisfied with your job). Note that forming a global opinion or evaluation based on taking into account various component experiences is not at all what we mean by making the judgment that an organization falls into one of the four relational categories in particular because of a single, salient feature (e.g., saying the organization is AR because hierarchy seems to be its most distinguishing characteristic). The "systemic justice" construct refers to a global impression made by using impressions about component parts, whereas what we emphasize (based on Fiske) is that the subcomponents—and, hence, impressions about them—can differ from the judgments about the whole. Hence, an employee can judge that "I'm in an AR organization" and yet realize that different activities within that organi-

2 I6

CONNELLEY AND FOLGER

zation are sometimes governed by one of the other relational models. Thus, Fiske's idea of AR as a sociality mode suggests it could apply at the systemic level as an overall norm, whereas specific implementation models might appear with respect to different areas of practice. We amplify this comment in the next section. Relational Models and New Conceptions of Justice

Earlier we restricted our attention to distributive norms. We now shift by using that focus as a bridge to wider implications. Specifically, we address a prominent tendency in the organizational justice literature to draw a one-toone correspondence between (a) the "type" of relationship (e.g., involving harmony and solidarity versus productivity and efficiency) and (b) a particular distributive norm. Deutsch (1975) seems to have been one of the first to propose such a correspondence, although it is also seen in several other works (e.g., Leventhal, 1979; Mikula, 1980; Schwinger, 1980). A prominent example is the distinction by Clark and Mills between the types of relationships known as communal and exchange (e.g., Clark, 1984; Clark & Mills, 1979). Their research has indicated that people keep track of respective contributions (cf. the distributive norm of equity) in exchange but not in communal relationships. The latter instead call for attending to the distributive norm of need. Fiske's analysis is consistent with this work but also extends beyond it in an important way germane to the current findings. On the one hand, Fiske's (1991) review of such work acknowledges that "as we would expect, there is some generality and stability in people's use of the models:" confronted with a novel experimental task situation in which they have to choose how to transfer things among themselves, people tend to use their preferred, criterion, or default model (p. 165). On the other hand, Fiske also refers to a number of problematic complications involving the "intricate embedding and concatenation of structures in a complex social relationship" (p. 166), so his position is much nuanced and need not be reduced to implying the necessity of one-to-one correspondences. Consider the following illustration, for example, showing that people can "link and nest multiple" sociality modes within the same basic type of relationship: "a group decision can be made by consensus (CS) that a king should be elected (EM) to rule arbitrarily over us (AR) because it is the most efficient, cost-effective system of government (MP)" (Fiske, 2000, p. 90). This illustration also shows how relational model theory accommodates both the distributive and procedural aspects of organizational justice. For instance, the CS model legitimates group decision-making procedures involving consensus as a procedural norm (deliberations do not end until all agree) in the sense of promoting one-for-all and all-for-one unity. Holding an election for purposes of designating who is to occupy official positions (e.g.,

6. HIDDEN BIAS

2I7

king), however, implies a separate procedural norm of one person, one vote—consistent with the EM model. Note that basing elections on plurality or majority rather than consensus would eliminate the CS overlay, but that this particular example shows a concatenation of two procedural norms (CS and EM) within a single vote! Such an example, although hardly uncommon or fundamentally unusual, is already much more complex than the typical discussions of procedural norms within the organizational literature. In fact, theories of procedural justice might be enriched considerably if they considered procedural variations in the light of relational models. Suppose, for example, that a research agenda began by designing the structural architecture of various decisionmaking or conflict-resolution procedures with each type of model in mind, thereby attempting to operationalize four distinct types of procedural arrangements (e.g., a CS procedure might use all of various interested parties to collect evidence according to their abilities, then allow the pooled evidence to be used selectively by any given party according to his or her needs, such as in arguing that a given decision is the correct one). Meanwhile, back to Fiske's example. The example continues by indicating that a candidate is elected to the position of king. This AR position has both distributive and procedural implications, as we noted earlier: Procedurally, further decisions can now be dictated by fiat; in the vein of the research agenda mentioned before, this would suggest that one operationalization of an AR procedure is a dictator role (one person is perhaps in charge of both gathering evidence and choosing among the alternatives so identified). Distributively, the dictated allocations as well as those made voluntarily by those at a subservient level will tend to honor the deferential high rank of kingship (e.g., getting the most, the best, first choice). Finally, note again that here AR also implies a form of institutional structure (hierarchy) that can be conceptualized as an alternative, legitimating norm for "systemic" justice—and yet, that the justification and legitimization stems from applying the MP model, and the associated norm of efficiency (raft'onality) as an ideal! Of course, this is not the only type of concatenation that might be employed involving efficiency. Consider Fiske's statement about the kingly-dictator role's having been chosen on grounds for "the most efficient, cost-effective system of government (MP)." The point is that the ends (cost-effectiveness, efficiency) were borrowed from MP reasoning, whereas the means for attaining those ends were borrowed from AR. Alternative models might keep the MP ends of efficiency but employ means guided more in line with either CS or EM (that the most cost-effective and efficient way to obtain evidence would be to allow equal search opportunities, in EM fashion, to parties with competing claims—the logic of counterbalanced bias, as in the prototypical symbol of EM-balanced scales in the hands of Blind Justice).

2 I8

CONNELLEY AND FOLGER

The embedded concatenations of sociality models, with linked and nested multiples of associated fairness norms, has a parallel in our findings at Corona. Recall that at a high level of abstraction, all Corona employees appeared to recognize the legitimacy of hierarchical structures as typically embodied in the modern corporation. At the more concrete level of implementation, however, they differed with respect to preferences across CS, MP, and EM. Both Fiske's theory and the present investigation, therefore, suggest the desirability of modifying the one-to-one correspondence principle that has become a standard assumption in much of the organizational justice literature. As mentioned in our introduction, different types of things may be distributed in different patterns within the same relationship, so that multiple distribution norms can be viewed as fair even within a single "type" of relationship—namely, the coworker and employer-employee relationship. Said differently, different norms may be seen as appropriate by different sets of people (yet all within the same general category of relationship) when fairness is viewed through the alternative lenses of contrasting relational models. Thus, we argue that Fiske's analysis and our research suggest the need for further theorizing beyond the approach traditionally adopted in the organizational justice arena. Quo Vadis?—Beyond Theory to Practice We end on the familiar note of the Lewinian dictum that nothing is so practical as a good theory. Earlier we speculated about the need for new theorizing in the realm of organizational justice. Here we consider what directions practice might take if guided by such theorizing. One moral of our story about Corona should be clear: Fairness in the workplace does not reduce to a single principle. Rather, different groups will view fairness issues in the light of their attitudes regarding an organization's existing workplace practices. The adage "If it ain't broke, don't fix it" is apt—those who think existing practices do not need fixing will invoke fairness norms consistent with these practices, whereas those disadvantaged by such practices will appeal to alternative norms instead. Management practice would be better informed by not viewing these alternative conceptions as based on arbitrarily and capriciously derived attitudes, or as thinly disguised attempts to cloak self-interest in the mantle of morality by invoking fairness principles. Fiske's work indicates that such principles derive from very basic approaches to sociality itself, rather than being conveniently invented ideologies. Managers who encounter diverging perspectives in fairness, therefore, should give credence to grounds for different viewpoints based on the cognitive architecture of humankind. Moreover, our analysis suggests that the use of norms derived from alternative relational models is symptomatic of fundamental dilemmas likely

6. HIDDEN BIAS

219

to be encountered with increasing frequency as a function of a multiculturally diversified labor force. One source of alternative norms, which we have not emphasized, is that culturally diverse groups might vary characteristically in their use of relational models as ways to process features of sociality. Collectivist cultures might more readily adopt a CS model, for example, whereas individualist cultures presumably might accommodate to MP with greater ease. In contrast, the comments of employees at Corona suggest that managers who attribute fairness outlooks solely to cultural differences might overlook other grounds for dissatisfaction—those more symptomatic of serious issues that cannot be dismissed as inevitable in a clash of cultures. True, the ingroup culture of CS favored by White males at Corona did clash with the MP values espoused by women and the EM values espoused by Blacks. Attributing these value discrepancies as flowing from inherently female or Afro-centric cultures, however, would discourage a closer look at the root causes of dissatisfaction with dysfunctional practices. For that reason, a Fiskian perspective promises to provide managers with new vistas of understanding.

REFERENCES Adams, J. S. (1965). Inequity in social exchange. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology (Vol. 2, pp. 267-299). New York: Academic Press. Clark, M. S. (1984). Record keeping in two types of relationships. Journal of Personality and Social Psychology, 47, 549-547. Clark, M. S., & Mills, J. (1979). Interpersonal attraction in exchange and communal relationships. Journal of Personality and Social Psychology, 37, 12-24. Connelley, D. L. (1994). Social identity as a barrier to understanding: The role of gender and race. Academy of Management Best Papers Proceedings 1994. Madison, WI: Omnipress. Connelley, D. L., & Goodrick, E. (2002). When a rose is not a rose: Understanding institutional conflict through an integration of social identity and relational models theories. Manuscript submitted for publication. Deutsch, M. (1975). Equity, equality and need: What determines which value will be used as a basis for distributive justice? Journal of Social Issues, 31, 137-150. Durkheim, E. (1933). The division of labour in society (G. Simpson, Trans.). New York: Free Press. (Originally published in 1893) Fiske, A. P. (1991). Structures of social life. New York: Free Press. Fiske, A. P. (1992). The four elementary forms of sociality: Framework for a unified theory of social relations. Psychological Review, 99, 689-723. Fiske, A. P. (2000). Complementarity theory: Why human social capacities evolved to require cultural complements. Personality and Social Psychology Review, 4, 76-94. Fiske, A. P., & Tetlock, P. E. (1997). Taboo trade-offs: Reactions to transactions that transgress the sphere of justice. Political Psychology, 18, 255-297. Gibson, C. B., Randel, A. E., &Earley, P. C. (2000). Understanding group efficacy: An empirical test of multiple assessment methods. Group & Organization Management, 25, 67-97.

220

CONNELLEY AND FOLGER

Gurin, P., Gurin, G., & Morrison, B. M. (1978). Personal and ideological aspects of internal and external control. Social Psychology Quarterly, 41, 275-296. Levanthal, G. S. (1979). Effects of external conflict on resource allocation and fairness within groups and organizations. In W. G. Austin & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 237-251). Monterey, CA: Brooks/Cole. Mikula, G. (1980). On the role of justice in allocation decisions. In G. Mikula (Ed.), Justice and social interaction: Experimental and theoretical contributions from psychological research (pp. 127-166). New York: Springer-Verlag. Nkomo, S. M., & Cox, T. Jr. (1996). Diverse identities in organizations. In S. R. Clegg, C. Hardy, & W. R. Nord (Eds.), Handbook of organization studies (pp. 338-356). Thousand Oaks, CA: Sage. Schwinger, T. (1980). Just allocation of goods: Decisions among three principles. In G. Mikula (Ed.), Justice and social interaction: Experimental and theoretical contributions from psychological research (pp. 95-125). New York: Springer-Verlag. Sheppard, B. H., Lewicki, R. J., & Minton, J. W. (1992). Organizational justice: The search for fairness in the workplace. New York: Lexington Books. Sondak, H. (1998). Relational models and organizational studies: Applications to resource allocation and group process. Journal of Organizational Behavior, 5, 83-103. Tolbert, P. S. (1988). Institutional sources of organizational culture in major law firms. In L. G. Zucker (Ed.), Institutional patterns and organizations: Culture and environment (pp. 101-114). Cambridge, MA: Ballinger. Tornblom, K. Y., & Jonnson, T. R. (1985). Subrules of the equality and contribution principles: Their perceived fairness in distribution and retribution. Social Psychology Quarterly, 48, 249-261. Tonnies, F. (1988). Community and society (C. P. Loomis, Trans.). New Brunswick, NJ: Transaction Books. (Originally published in 1887) Walster, E., Walster, G. W., & Berscheid, E. (1978). Equity: Theory and research. Boston: Allyn & Bacon. Weber, M. (1946). From Max Weber: Essays in sociology (H. H. Gerth & C. W. Mills, Trans.). New York: Oxford University Press. Whitehead, H. (1993). Morals, models, and motives in a different light: A rumination on Alan Fiske's structures of social life. Ethos, 21, 319-356.

C H A P T E R

7 Relational Models, "Deonance,' and Moral Antipathy Toward the Powerfully Unjust Robert Folger Rebecca Butz

The study of organizational justice has shed light on important issues (for a review, see Folger & Cropanzano, 1998). We argue that a subset of those issues stands to be enriched even more in light of Fiske's (e.g., 1991) theorizing about relational models. One of four such models, or prototypes of sociality by which people govern their relations, Fiske calls Authority Ranking (AR). The subset of organizational justice issues we address relates most directly to workplace AR when abusive authorities unfairly mistreat employees. We refer to one aspect of potentially abusive authority by using QC as an acronym from the first two words in an ancient Latin phrase: Quis custodiet ipsos custodes, or "Who will keep the keepers themselves?" That phrase's implications parallel those of "power corrupts and absolute power corrupts absolutely"—and also Saint Bonaventura's more colorful expression, "The higher the monkey climbs, the more you see his behind," which "warns of the exposure of character flaws in people free of the usual social constraints" (de Waal, 1996, p. 97). The Latin quis custodiet (QC) refers to the danger of insufficient constraint (a keeper-of-the-keepers deficiency) when no one opposes abusive authority to "keep" power in check. We link this topic to Fiske's (1991) relational models (especially AR), organizational justice, and a morally emotional state called "deonance," whose hypothesized effects may be crucial for social order. If the subordinate status of abused victims (V) equates to weakness, fear of retaliation from authority, and a lack of the power to oppose unjustly abusive transgressors (T), then V's lower rank in a hierarchy might pre-

221

222

FOLGER AND BUTZ

elude effective opposition against T's more dominant status. One potential answer to "Quis custodiet... ?" thus might entail opposition to an authority's power abuse by those other than its victims (e.g., London sweatshops made infamous at the start of the industrial revolution might still exist if not for eventual upper-class opposition stirred by the likes of writers like Charles Dickens). In that vein, we link QC issues to a specific type of custodes role for third parties, a "3QC" role as "keepers of the keepers": being vigilant observers prepared to exercise counterpower against the ipsos custodes as wrongdoers (e.g., abusive bosses). Using both anecdotes and research data, we focus particular attention on the role of third parties who gain nothing from opposing power abuse and, in fact, endure a costly sacrifice (e.g., the physical costs of a hunger strike against officials whose abuse does not bear on your self interest). Taking action despite negative, costly implications for self-interest, of course, violates commonsense psychological canons of pleasure seeking and pain avoidance. Standard economic analysis also has trouble with seemingly self-sacrificial behavior such as tipping in a restaurant to which you will not return. We adopt instead a deontic perspective (cf. Folger, 2001) in reference to the Greek root, deon, which means duty or obligation—as when people act in line with moral dictates. We develop this approach more fully by integrating it with Fiske's work on relational models and the literature on organizational justice. We address when, why, and how victims might count on the morally based emotion of "deonance" and related, deontically motivated actions as support from third-party observers eligible for the keeper-of-keepers role. Such 3QCs are neither power abusive, transgressive perpetrators of injustice (first-party Ts) nor abused victims of it (second-party Vs). As observers, however, 3QCs know about the power abuse and the abused V's harm. We define as a 3QC such an observer who opposes, or is at least in a position to oppose, the abuse. After an anecdote referring to one such 3QC-like event, we sketch an outline for the remaining discussion, then give another 3QC illustration based on research data. The anecdote and the data, however, should be regarded as "points in a 3QC space"; we add further elaboration throughout, but the full dimensionality of that space awaits much more theory and additional research. In particular, we refer only loosely to the qualities of deonance-based emotion as essentially encompassing forms of anger-like antipathy such as moral outrage and righteous indignation. AN ANECDOTE AND PREVIEW OF TOPICS

We begin with an anecdotal example of third-party antipathy and hostility toward an authority seen as having abused the prerogatives of power. One of us once heard his seatmate on a plane convey outrage regarding how the

7. MORAL ANTIPATHY TOWARD THE POWERFUL

223

(then) new owner of the Dallas Cowboys football team had fired Tom Landry, the team's coach. The man, a Cowboys fan, expressed righteous indignation because Landry learned of the firing only by reading a newspaper. This incident relates to topics discussed throughout this chapter: Organizational justice concepts. The fan's outrage came not from the perception of an undeserved outcome (the fan thought Landry needed firing; distributive justice); nor from a perception of a flawed decision-making process (procedural justice). Rather, the unjustified mistreatment involved a category of interpersonal conduct that includes insensitivity and the like (interactional justice; Bies, 1987). Note that the organizational context applies regarding the employer-employee relation between the owner and Landry. Third-party reactions. In contrast with Landry and the owner of the Dallas Cowboys as two parties both belonging to the same organization, the fan was an "outsider" or third-party observer of events taking place within that organization. Third-party observers hold special relevance for business organizations because of the public's role vis-a-vis potential sanctions against businesses (e.g., as jury members in suits involving punitive damages; as voters who can influence business-related legislation). Moral antipathy. The fan expressed anger and hostility regarding the owner, conveying more than the simple judgment that the owner had made a mistake. A teacher might not become angry when a beginning student writes down the wrong answer to a math problem, for example, whereas the fan was mad in a way that conveyed moral disapproval and condemnation of what he considered to be misconduct by the owner. Power, prestige, and status as aspects of AR. Each of Fiske's relational models represents a somewhat idealized, broad abstraction. As Fiske's (2000) subsequently related development of Complementarity Theory has made abundantly clear, the difference between the conceptual description of AR and its implications for conduct within a specific social setting is not unlike that of the difference between a theoretical construct and its operationalization in an experiment—or to use another analogy, the relational models have a fuzzy-logic or "family resemblance" quality (much like the difference between the features of any particular dog and the generic concept of dog). For that reason, we want to emphasize at the outset that not all instantiations of AR are alike. We focus rather exclusively on its power implications as implied by an evolutionary perspective we adopt. We see an evolutionary link to power, for example, both in transgression (as a forcefully coercive act) and in morally outraged opposition (as countercoercive or punitive in nature).

We discuss these topics in light of Fiske's (1991) four foundational or elementary models of social relations: Communal Sharing (CS), Authority Ranking (AR), Equality Matching (EM), and Market Pricing (MP). AR is espe-

224

FOLGER AND BUTZ

dally relevant to situations like the Landry example, where one party is dominant and the other subordinate in an organizational hierarchy: "Authority Ranking is a relationship of asymmetric differences, commonly exhibited in a hierarchical ordering of statuses and precedence, often accompanied by the exercise of command and complementary displays of deference and respect" (Fiske, 1991, p. ix). Fiske (1991) also says the four relational models have a "jural" (normative) feature giving them the "intrinsic imperative force" associated with constraints such as moral obligations (p. 170)—a feature we address later. Fiske called the jural quality a defining feature of the relational models that is more fundamental and elementary (psychologically primitive) than other conceptions potentially relevant to organizing and guiding social life. In the same vein, Fiske (1991) pointed to a possible evolutionary perspective: "any model that is fundamental, elementary, and universal in human social relationships has to be phylogenetically situated, and we must ultimately account for it as a product of natural selection, among other levels of explanation" (p. 199). Guided by this comment, we seek a conception of moral antipathy toward abusive organizational authorities consistent with the evolutionary past of the human species. The chapter proceeds as follows. We begin with a section highlighting some recent experiments on third-party reactions to unjust authorities. With results from those experiments as grist for the mill of further analysis, the remainder of the chapter considers various lessons that might be learned from studying why people get angry at others' social misconduct, particularly third-party antipathy toward abusive authority. That discussion includes speculation about the evolutionary roots of anger as a moral emotion (i.e., moral antipathy), relating back to Fiske's ideas about the jural properties of the relational models. We also expand on Fiske's treatment of AR and reactions to organizationally relevant transgressions in which that model plays a schematic role. In particular, we argue that Fiske's account of AR makes it uniquely suited for analyzing cases of moral disapprobation of authorities by third parties, such as the fan's reaction to the manner of Landry's firing. We spotlight the centrality of power as an organizing theme that quite explicitly underlies our analysis. We conclude by suggesting that this account of AR has even broader implications for moral antipathy in general. THIRD-PARTY, SELF-SACRIFICIAL PUNISHMENT OF AUTHORITY: MORAL ANTIPATHY IN AN AR SETTING

Experimental game theory economists sometimes use payoff structures in the laboratory to create an authority role for research participants—and in Fiske's terms, therefore, an AR situation. Specifically, these research partici-

7. MORAL ANTIPATHY TOWARD THE POWERFUL

225

pants become the highest-ranking authority on allocation by being given free reign to make decisions that have actual or potential consequences for their own financial well-being and that of others. The so-called Dictator Game, for example, allows Player A to determine—unilaterally—his or her own payoff as well as the payoff going to one or more other people (Player B, etc.). In other words, Player A's role makes him or her a hierarchically ranked superior or dominant player with respect to B and any other players, whose roles are subordinate. The payoffs from these "games" can be expressed as points exchangeable for money at the end of the experiment, or the points can be identified as the basis for monetary payoff to a randomly selected subset of participants. We mention the latter because it was the method used in several of the studies we report. The key Dictator paradigm described in our following discussion, however, has produced comparable results regardless of whether every participant gets paid or only a randomly selected subset. The "KKT" Studies by Kahneman, Knetsch, and Thaler (1986) Kahneman, Knetsch, and Thaler (1986—henceforth KKT) studied whether research participants as third parties would willingly sacrifice $1 to punish intended unfairness by someone who had previously played the Dictator role in determining game payoffs. In essence, participants were placed in third party positions which enabled them to punish intended-but-unimplemented unfairness (and simultaneously to reward intended-but-unimplemented fairness). Specifically, third parties were led to believe that previous students had been placed in a Dictator role and had been asked to choose how to divide $20 with an anonymous other. The Dictator's only options were either to designate (a) $18 for themselves and $2 for the classmate, or (b) to divide the $20 equally, each person receiving $10. With this information in hand, participants were advised that they had been matched with two of these Dictator-role participants: one (called U, for uneven, in KKT) who had opted for $18; the other (E for even), who had opted for $10. Participants were then asked to choose between taking $6 for themselves and giving $6 to the "U" Dictator, or taking $5 themselves and giving $5 to the "E" Dictator. Thus the options were either (a) to pay a dollar to split money with a stranger who endorsed fairness (equality) or (b) to split the money without cost with a stranger who had made the self-interested choice. The modal response (across variations, consistently averaging around 75%) was self-sacrificial punishment of the "U" Dictator. Results from the KKT paradigm thus indicated that by far the vast majority of the participants willingly sacrificed money that they could have allocated to themselves. They did so (at least in part—a point to which we re-

226

FOLGER AND BUTZ

turn momentarily) to punish the "U" Dictator who tried not to be fair (i.e., indicated the preference to deprive a partner from profit in a previous round). These results, therefore, suggest that many third-party observers of a previously attempted injustice will sacrificially punish an abuse of authority as simply "the right thing to do." KKT-Extensions by Turillo, Folger, Lavelle, Umphress, and Gee (2002) The KKT results indicate that people would rather reward individuals for previously being generous than for previously being greedy, even if that means making a personal financial sacrifice. Is it possible, however, that third-party participants in the KKT studies were motivated to punish or reward based on some sort of self-interest? For example, perhaps participants wanted to promote a certain public image of themselves, or wanted to send a public message to the Dictators to deter future unfairness. To investigate how variations in the instructions to participants might influence their reactions, Turillo, Folger, Lavelle, Umphress, and Gee (2002) conducted a series of conceptual replications and extensions based on the KKT paradigm. Each variation addressed a different type of rational, instrumental motive with respect to conceivable forms of self-interested benefit that a third-party observer might gain from self-sacrificially punishing an abusive authority. In particular, the Turillo et al. studies varied instructions regarding anonymity in several different ways and showed that those variations had no effect—across conditions, the results were virtually identical to those obtained in the KKT studies. That is, an average of 73% of their subjects chose to sacrifice a potential reward and share their lesser sum with someone who had previously been "fair" than split a larger sum with someone who had taken most of the money for him or herself in a previous round. Turillo et al. (2002) altered the original KKT methodology in several ways. First, their studies did not use the terms Even (E) or Uneven (U) to refer to Dictator decisions from the "first part." Instead, the descriptors were simply "$10-$10" or "$18-$2"-to delete a cue that might artificially heighten the salience of equality as a fairness norm. Second, participants did not sign their questionnaires. Because participants' names appeared only on consent forms, which were collected separately, the participants knew that they could collect money without identifying themselves. Turillo et al. even emphasized to students that if for any reason they wanted to take special precautions to remain anonymous, they could let a friend pick up the payment. Only numbers were used to award money—no names were recorded. This made it impossible for anyone to link participants' identities with their responses. Turillo et al. thereby eliminated the promotion of self-interest that might result from public impression-management and broadcasting

7. MORAL ANTIPATHY TOWARD THE POWERFUL

227

one's good deeds to others for the sake of basking in praise or establishing a reputation as a virtuous person. For punishing the abuse of authority to be "its own reward," no one else must know who did this good deed. Aside from ensuring anonymity, Turillo et al. also altered the descriptions of who would receive what kind of information about the consequences of their anonymous decision. Participants were told one of the following: (a) only $18-$2 Dictators would learn the results regardless of the payoffs the participant chose (i.e., $10-$10 Dictators would not); (b) only $10-$10 Dictators would learn the results regardless of the participant's payoff choice ($18-$2 students would not); (c) both the $18~$2 and the $10$10 Dictators would receive such information; (d) neither Dictator would know anything about the participant's choice unless scheduled to receive payment (i.e., the choice of the $0 punishment option would not be revealed); or (e) the Experimenter would require a signature at the time of payment, eliminating the participant's anonymity. There were no differences among the above conditions at an average (73%) nearly the same as KKT; F(4, 93) = 1.47, p = .22. The absence of differences among these replication conditions implies the lack of effects from motives germane to self-interest. If respondents thought they could change someone's future behavior by their allocation decision, for example, then responses would be confounded with a motive relevant to self-interest. If the absence of an opportunity to have any kind of instrumentally consequential influence produced the same results as those obtained when such an opportunity was present, however, then the relevance of self-interest is moot. Consider, for example, the variation in which the only person to learn the results from the participant's choice would be the Dictator from the "first part" whom the decision maker designated as his or her sole co-recipient (the prior $18/$2 or the prior $10/$10 Dictator). Sacrificial punishment of unfairness in that condition would be known by the fair person who benefited but would not be known by the unfair person punished, which tested for effects from not being able to "send a message" (i.e., the value of wanting to encourage a generic deterrence effect). Contrasting that condition with one in which both other people learned the results showed that a self-interested deterrence possibility had no impact. Hence, we can restate the relevance of these variations in terms of each of the possible sources of self-interest that failed to emerge as a significant determinant of behavior: Public Image. Remaining anonymous, respondents could not get any credit from anyone else for "doing good for goodness' sake." Public Message. Because no one other than the researchers would see the raw data, respondents were not able to use their punishment as a public message that "crime does not pay," thereby negating an explanation of

228

FOLGER AND BUTZ

general deterrence (sending a message to the general public) that might otherwise have been possible. Perpetrator Deterrence. To have any hope that the punishment might deter the perpetrator from committing a similar transgression in the future, the respondent would have to believe that the perpetrator was aware of being punished. But, as indicated above, this variation had no effect, indicating that self-interested deterrence was not influential. In addition, the replication result (73% self-sacrificial punishment) also speaks against self-interest because participants chose to punish Dictators who had not actually received payment, but who had indicated an intention to be unfair in their allocation of the money. In this way, punishment served to deprive the $18-$2 Dictator because of his or her presumed unfair intentions to collect $18 out of $20. Finally, the participants did not know the $18-$2 Dictator's identity. They were merely third-party observers of his or her unfair intentions, not participants in a long-term exchange relationship. Therefore, anonymity made it implausible to foresee any long-term consequences from their allocation decisions. For such reasons, the modal (73%) self-sacrifice choice of $5 (rather than $6) starts to indicate actually doing "good for goodness's sake" with "actions louder than words" (beyond mere fairness endorsement as a cost-free attitude).

AR, MORAL ANTIPATHY, AND EVOLUTION: ROOTS OF DEONTICALLY DRIVEN ACTION TENDENCIES

Much of social science relies for explanation on narrowly defined selfinterest and, therefore, has difficulty accommodating such data as those from the Turillo et al. (2002) studies reported before. Even equity theory, a model concerned with fairness perceptions, can be interpreted as making self-interest the ultimate foundation of justice. For example, Berscheid and Walster (1978) believed that groups develop equity norms to "maximize ... collective reward" (pp. 125-126). If the personal welfare of economic benefits were the only plausible source of human motivation, however, then data such as from Turillo et al. could not exist. If not economic self-interest, then what could have motivated participants? Our answer refers to a specific emotion, often simply called anger, which we label as moral antipathy and do not consider irrational at all. In fact, we think it is also the answer to the 3QC dilemma and that, for this reason, its "rationality" consists of allowing for one of the only available means of curtailing abuses of power otherwise free to run rampant. The alternative to morality is a might-makes-right use of coercion by the powerful to

7. MORAL ANTIPATHY TOWARD THE POWERFUL

229

maintain power. If absolute, unfettered power does indeed tend to corrupt absolutely, then a lack of any means for curtailing power would constitute a corrupting influence. We speculate that "corruption" in this sense would not enhance the survival of a species that required voluntary cooperation for sustaining effective access to resources—a feature, we believe, that characterizes humans. In other words, the existence of a 3QC checks-andbalances possibility relies on a human capacity for a type of moral affect related to disapproval of wrongdoing (a condemnatory emotion). Affect as a Product of Natural Selection The calculative self-interest seeking of rational choice and standard neoclassical economic models entails only a forward-looking, instrumental aim at future goals. That is, such models assume that people respond strategically based on what they assume will further their interests. A contrasting, "backward-looking" perspective (i.e., focused on antecedents, especially from an ancestral past history) comes from conceptualizing the function of affect in terms of adaptations favored by natural selection. This evolutionary viewpoint emphasizes emotions as adapted for emergencies: When environmental cues signal jeopardy, emotions save time by providing a builtin assessment of instrumental needs based on adaptively successful responses to such encounters in an ancestral past (cf. Fiske, 2002). We turn, therefore, to natural selection as a source of such affects as moral antipathy or indignation. We suggest that self-sacrificial behavior exhibited by participants in the Turillo et al. studies was driven by moral antipathy about injustice as the abuse of power. Furthermore, we suggest that this moral antipathy has evolutionary roots. We contend that moral antipathy has evolutionary roots because evidence suggests that natural selection shaped not only our physical machinery but also our mental machinery (e.g., neural circuitry relating to the experience of emotion). Bodies of literature across a diverse set of fields now provide both theory and research supporting that conclusion (e.g., Barkow, 1989; Katz, 2000; Tooby & Cosmides, 1990). Such assertions apply not only to emotions such as fear but also to those with moral connotations, such as guilt and moral outrage. To the extent that "people are motivated to behave morally not because they have rationally calculated that it is in their material interest to do that, but because they are emotionally predisposed to do so" (Frank, 1988, p. 288), it is important to consider evolutionary explanations about the possible origins of emotions. It seems plausible that the earliest precursor to moral outrage remains physiologically and neurologically encoded, much as Damasio (2001) has postulated can occur via "somatic markers" within the lifetime of a single individual. Specifically, at least two types of data imply the existence of an

230

FOLGER AND BUTZ

unthinking, prepotent, and simplified core to what Panksepp (1998) calls the RAGE system. First, humans exhibit rage-like behavioral tendencies when a neuroscience researcher electrically stimulates certain brain regions, whereas their own actions seem inexplicable to them both during and after the event. Common tendencies include verbalizations along the lines of "Get out of the way, I've got the urge to throw something at you." Second, infants as young as 14 months of age exhibit similarly expressive displays of rage when an experimenter pins their arms to their sides and holds tight (Stenberg & Campos, 1990). Based on such indications combined with several supporting lines of theory and research (e.g., reviewed by Panksepp, 1998), we speculate that the anger of moral outrage (a base ingredient in the moral antipathy syndrome) stems from adaptive responses to being attacked or having to fight against a resistant, opposing force—such as when a cornered prey cannot flee (e.g., perhaps partially and temporarily immobilized when grasped or pinned by a predator) and instead adopts the fight mode of the fight-or-flight response. In agonistic, competitive, or fighting situations against an opponent (conspecific or predator), the counterattacker gains an evolutionary advantage from using forcefully fierce displays in an intimidating fashion. Given the plasticity of behavioral adaptations garnered by humans, natural selection has operated so that it has also become possible to produce ritualistic displays of fierceness that can cause an opponent to cease and desist without an actual fight occurring. Resolving a dominance-submission confrontation by ritualistic displays of power and force allow conflict incidents to transpire at a reduced cost to reproductive resources—such as when the moral suasion of a sneer proves a sufficient deterrent. In general, then, humans have evolved beyond other species in the sophistication and subtlety by which they can apply negative sanctions against unwanted (e.g., transgressive, morally offensive) behavior. We suspect that the emotions we feel as humans when experiencing aspects of the moral antipathy syndrome build on the physiologically adaptive accompaniments of fighting back or fighting against. That is, we assume selection has favored bodily preparedness to attempt greater-thanordinary exertions of force when circumstances appear to deem it necessary (e.g., often because of instigations that take the form of an obstructive or coercive, countervailing source of opposition or resistance)—hence the tendency to get angry at a stuck window you cannot lift at first. Relatedly, fearsome expressiveness and a foreboding posture like the Darwinian (1872/1998) description of anger could motivate an opponent to quit rather than risk getting hurt. Put simply, an emotionally angry display and related forceful exertions can serve to help stop and possibly reverse an adverse trend from an opposing, resistant force. One could say, then, that outrage is the moral equivalent of "fighting fire with fire" because it betokens a coun-

7. MORAL ANTIPATHY TOWARD THE POWERFUL

231

terattacking thrust against a force that broke through the bonds of moral restraint commonly taken for granted by members of society. Because society's mores supposedly bind people to moral obligations and keep in check an overly aggrandizing disregard of others' interests (vis-a-vis specific persons or with regard to the moral community as a whole), moral transgressions get coded emotionally and psychologically by observers as excessive force willing to break those bounds. Thus, the need to meet force with force calls for an angry moral antipathy against violators. Evolved Antipathy as Moral in Nature We believe that natural selection favored the development of moral emotions enabling ways to prevent (or reduce the frequency of) actions apt to jeopardize humans' survival chances as a species. For a perspective on the moral aspect of such emotions, consider "moral goods" as "social, personal, or spiritual obligations (e.g., justice, social harmony, self-actualization, piety, chastity) to which one appeals to justify or criticize the practices and behaviors of others, and which are felt to be binding on all people" (Haidt, Rosenberg, & Horn, 2003). This type of statement about morality fits with our reference to deontics as the study of moral commitments experienced with the force of a binding obligation (cf. "I didn't really do it for any particular purpose—it just seemed like the right thing to do"). Note, moreover, that moral standards identify criteria for judging critically those who do not live up to moral dictates. We conceptualize moral antipathy as being offended by transgressions, and as a force that drives individuals to oppose and curtail abuses of the moral system. Just as morality reflects the deontic characteristic of binding obligation as a constraint limiting an otherwise unfettered autonomy, to impose a moral standard implies exerting one's will forcefully in opposition to the transgressor of that moral standard (cf. the "force of moral suasion" even when one hopes the mere appeal to a moral standard will alter conduct). Indeed, the language of power as an opposing force is part and parcel of the very concept of morality itself, once the moral domain is seen as being opposed to anyone's efforts to maximize his or her own self-interest exclusively (i.e., without considering potential harm to others as a result). Conscience means exercising self-restraint—opposing the power that pure self-interest would exert by forcing ourselves to act in another fashion instead. The Moral Status of Third Parties: Rising Above Self-Interest The reactions of third parties that we have highlighted also suggest that the experience of outrage in response to abuse of authority has a moral basis. Specifically, the lack of direct self-interest of third parties fits with how

232

FOLGER AND BUTZ

ethicists view morality: A canonical view suggests that morality "must bear on the interests or welfare either of society as a whole or at least of persons other than the judge or agent" (Gewirth, 1984, p. 978, quoted in Haidt, 2003). Third parties, as we use the term, are those in the role of "moral spectator," as Adam Smith (1759/1971) called it, whose judgments reflect adopting a perspective outside the realm of personal interest (often exercised in line with the interests of the moral community as a whole). Consider again the anecdotal and experimental examples provided earlier. The Dallas Cowboys fan enraged by Landry's firing can be defined as a third party to the employment relation between the owner of that team (first party, or injustice perpetrator) and its coach (second party, or injustice victim). Obviously the fan had self-interested involvement in how the team fared. With respect to the manner of Landry's firing, however, much less (if any) direct self-interest seems evident. It is hard to argue credibly that the fan's own self-interest would have been served better to an appreciable degree—worth the investment of emotional energy—if Landry's firing had been handled by the owner with greater consideration for Landry's feelings, such as by making sure Landry received notification before news reporters. Turillo et al. (2002) conducted their series of experiments in attempting to create third-party conditions even more removed from any form of direct self-interest. First note that no harm had ever actually injured a secondparty victim, so no basis existed for emotional empathy (no vicarious suffering to be removed as a means to promoting self-interest). Second, punishing the greedy, first-party Dictator could not make the world a better or safer place for the third-party: The respondents who chose punishment knew that neither the Dictator nor anyone else would ever learn about the punishment. The conditions of anonymity also provided a punisher with no benefits in terms of a reputation enhanced by displaying virtuous, selfsacrificial behavior. Thus, it appears that the tendency for third parties to react to the abuse of power even though they are not directly harmed stems from an evolved tendency to experience moral affect, specifically moral antipathy and indignation. Moral Antipathy and the Jural Property of Relational Models

Interestingly, Fiske also highlights the moral aspects in his relational models. For instance, similar to the deontic (moral obligation) approach we take toward explaining antipathy, Fiske describes his relational models as having jural qualities. The definition oijural relates to rights or obligations (the central deontic focus), and Fiske (1991) portrays his four relational struc-

7. MORAL ANTIPATHY TOWARD THE POWERFUL

233

tures as entailing such obligation: "The most important distinctive feature of this class of structures is that they are jural, or normative models. ... Specific rules and practices are held to be obligatory and legitimate because they are derived from these fundamental models" (p. 170). The models "have intrinsic imperative force" in that "it is reprehensible not to concern oneself with the performance of others in these social relationships: there is an obligation to monitor, intervene, and sanction where appropriate" (p. 170). Indeed, Fiske claimed "people value these forms of social relations ideologically and are committed to them as moral standards they impose on themselves and others" (p. 178). Additionally, Fiske's discussion about reactions to violations of mandates derived from the relational models is rife with language akin to power and force, or gaining the "upper hand" (cf. a higher and more dominant rank in an AR sense). Even the simple description of performing roles consistent with the implications of a relational model takes on this force-laden connotation, because these implications are also imposed requirements: "To maintain one's good standing in other linked relationships [viz., within the context of a relational model], one must perform each focal role adequately" (Fiske, 1991, p. 179). If you must do something, that means you are forced to do so. More than mere metaphor, the language of force (power directed toward proper alignment with the position dictated by a relational model) is absolutely central to Fiske's (1991) entire conception of the relational models and their jural feature as a defining characteristic: "Most moral precepts . . . are in some degree derivative from the fundamental underlying models. To the extent that people take such rules and values as axiomatically governing their own behavior, and as requirements which they must necessarily impose and enforce on others [emphasis added], the basic models are directive ideals" (p. 180). Although Fiske discusses the jural quality of all four of his relational models, we focus primarily on the power issues associated with transgressions in an AR setting. We do so because we think that moral transgressions involving the abuse of authority are especially germane to the context of organizations which are inherently AR in nature. A Caveat: The Many Faces of AR Before continuing our discussion, we must confess that we in no way claim to present AR's implications in a fashion 100% consistent with what Fiske has written or with what he might agree. Instead, we try to zero in on one very narrow set of AR features stressing power, consistent with Quis custodiet... dilemmas (without claiming that this emphasis reflects an orthodox Fiskian interpretation). At the same time, we deliberately limit our application of AR in that fashion precisely because we think of power as

234

FOLGER AND BUTZ

one particular "thin slice" of the AR totality that still has sweepingly broad implications when applied to fresh understandings about perceived organizational injustice and reactions to it. We focus specifically on the aspect of power within hierarchies, but we do so with the knowledge that there are at least several other conceptualizations of AR that focus on other aspects of hierarchy. For example, Henrich and Gil-White (2001) have made a case for focusing on prestige instead of power. They assert that "In humans, in contrast [to nonhuman dominance hierarchies], status and its perquisites often come from nonagonistic sources—in particular, from excellence in valued domains of activity, even without any credible claim to superior force" (p. 167). Einstein's obvious high status in terms of prestige, for example, could in no way be attributed to "might makes right" attempts to use his physical stature in a dominating, coercive, or intimidating fashion (we thank Alan Fiske for this point). Their analysis points to the enormous potential for impact from socially transmitted information and social learning, processes by which people willingly defer to others whom they assume have acquired greater expertise and wisdom. In a related type of analysis, Rubin (2000) suggests that coordination rather than power or prestige is the key aspect of hierarchy in contemporary work organizations. Rubin explains that calculative rationality in humans should cause people to appreciate hierarchies organized for production because they can "lead to sufficiently increased output so that all members can benefit" (p. 271). In essence, he proposes that humans use the efficiency of hierarchy in order to coordinate the production of goods. Although both these approaches offer plausible insights regarding human hierarchies, we have chosen to examine more closely those instances in which the hierarchical differences of rank are seen primarily in powerbased terms. We emphasize AR's power-based aspects rather than others it also encompasses because we think the former are more central to the modal cases in which perceived injustice has important ramifications in the workplace.

Moral Antipathy and the Domain of Organizational Justice The literature on organizational justice addresses factors influencing fairness perceptions in the workplace, as well employee reactions when they perceive themselves to have been treated unfairly. We see justice as part of the larger domain of morality. We define morality as a subject matter that also entails a hierarchy-like feature central to the very meaning of morality. By that token, organizational injustice and morality would relate to the AR

7. MORAL ANTIPATHY TOWARD THE POWERFUL

235

model even if organizations did not contain hierarchies—which, of course, they always do. Moral tenets and codes imply rules of conduct and the injunction not to break them. Descriptions of transgressive acts usually include the wrongdoing and a wrongdoer (the deed and the doer). Referring to the wrong itself, common usage will ordinarily have recourse to language about the state of the code or tenet at a pre-wrong and a post-wrong time: Prior to the wrong the code or tenet had not yet been broken (violated, etc.), whereas afterward the code or tenet has the status of brokenness (e.g., "she broke the rules"). Language describes the wrongdoer, therefore, as someone who breaks, violates, transgresses, and so on. Designating action as falling into the moral domain has the psychological connotation, therefore, that others feel legitimately entitled to hold the actor accountable and answerable for wrongdoing. The person's freedom, rather than conceived as infinitely open to the person's own discretion, has a circumscribed area within which its exercise needs no justification. Once outside the bounds of morality, however, the person becomes answerable to others. "I answer to no one" will strike members of the moral community as intolerable (cf. Rawls, 1971; Scanlon, 1998). We argue that in light of such conceptions of moral conduct, the story we crafted regarding the evolutionary bases of moral outrage is particularly applicable to understanding individuals' reactions to injustice in organizational settings. Although Fiske's AR model can be applied in an abstract way to many situations and settings, we believe that social relations in organizations, from large corporations to smaller businesses, are prime concrete representations of the AR model. Organizations are characterized by power laden dynamics, with clear hierarchical lines of authority and privilege. Those at the top of the organizational ladder have power and privilege over those at the bottom. Those at the bottom of the organizational ladder accept hierarchy as a legitimate form of organization, but as we might expect (given our evolved predisposition), they react with moral indignation at particular forms of injustice—those involving abuses of power. This is especially the case when a superior (boss) abuses his or her power over a subordinate (worker). In general in organizations, individuals assume and tolerate the presence of legitimate hierarchy in organizations. They likely do so for the benefits that it confers to all members of the group (Boehm, 2000). However, even those who are conferred dominance can overstep their bounds of authority. Superiors who abuse the power vested in them by forcefully pursuing their own interests to the detriment of their subordinates are no longer using their power to legitimate ends. Social influence that is legitimate or endorsed instead becomes coercion or intimidation. Although this type of social influence or wielding of power might be submitted to out of fear of

236

FOLGER AND BUTZ

negative consequences, it is ultimately less stable than power that is granted or endorsed (Blau, 1964). Perhaps the reason that it is less stable is that intimidation and coercion as abuses of power invoke strong moral responses of injustice and outrage that often compel individuals to react despite their position of subordination. Responses to this abuse of power can range from individual feelings of indignation and retaliatory behaviors (Bies, 1987; Skarlicki & Folger, 1997) to coalition formation and attempts to sanction undue domination and victimization (Boehm, 2000, p. 85). This is the case because abuses of power threaten to undermine "the useful web of co-operation that everyone profits from" (Boehm, 2000, p. 90). In essence, authority figures in organizations may act ruler-like (dominantly) so long as they are also "being ruled" by the moral code. Like all members of the organization, a superior's behavior is circumscribed by a moral boundary. Within the boundary, exercised power needs no justification. An abuse of power that has been legitimately conferred, however, symbolizes a step outside the bounds, an attempt to "break the bonds" of moral codes. And, a moral boundary crossing is met with a moral response—in this case, moral outrage or indignation. But How About the Other Relational Models? In this final section we amplify an earlier theme into a bold, admittedly very speculative proposal—namely that all moral violations (relational-model transgressions) implicitly invoke the AR "structure" as an "elementary form of human relations" (terms from the title of the book by Fiske, 1991). Assume that AR can influence consciousness without necessarily appearing in the conscious content of thought with all the trappings of that model. We might think of this as like the triggering of a node in an associative neural network, or we could conceptualize it as analogous to tacit knowledge (e.g., the "muscle memory" available after sufficient practice of a sport-relevant skill). We insist, in other words, that AR can affect emotional and behavioral tendencies without a person's literally thinking of a hierarchy at the time. Put another way, certain concepts and contexts implicitly connote aspects of hierarchy even if (a) no mention of hierarchy per se occurs and (b) the specific images of a formal hierarchy never come to mind as an object for reflection and consciously active consideration. In fact, we define morality as a subject matter that refers to hierarchy by the very nature of what morality must entail. True, some ethicists refer to "moral goods" (e.g., Haidt et al., 2003), and it might in that respect make some sense to think about allocations and exchanges. Certainly people often, at least in Western culture, will make judgments about the fairness of an exchange by using an MP form of logic (e.g., the outcome-input ratios specified by equity theory; Adams, 1965). We distinguish, however, between

7. MORAL ANTIPATHY TOWARD THE POWERFUL

237

(a) the relational-model source of a moral judgment that a transgression has occurred and (b) the feelings and behavioral tendencies spawned by awareness of such an event. The former can involve any of the four models, but the latter of necessity will always involve AR to at least some extent. Recall Fiske's (1991) remarks about the models' "intrinsic imperative force" in their normative or jural capacity: "Specific rules and practices are held to be obligatory . . . [when] derived from these fundamental models" (p. 170). The metaphor of force or power indicates how the very defining features of morality itself (having something known as a moral mandate implies the injunction not to rend the fabric of its constraints) also overlap with the psycho-logic of the situation (feeling the impulse to abide by the moral code as a force "holding me back" from wrongdoing). Normative injunctions, by definition, have constraint as their function. In a positive sense they serve as guides to correct behavior ("thou shalt")—but perhaps even more importantly, in a negative sense they limit acceptable behavior and thus place bounds around the permissible ("thou shalt not"). The tone of authority could not sound more loudly than when described along such lines of the logic of the term moral and the psycho-logic of response to it. Hence, any given model itself, whether CS, AR, EM, or MP, acts authoritatively as a superior force one is not supposed to "rise above"—the ultimate rank of dominance in the hierarchy. The authority force of a model (or rather, of a moral norm derived from one of them) has perhaps not been linked so explicitly to AR before simply because of reasons why other aspects of transgression have greater salience. AR stays in the background, implicitly taken for granted because of other language and modes of thought that influence the relative salience of various elements involved in a wrongdoing incident. In particular, descriptions of such an event often focused on deed and doer describe the latter with terms such as transgressor or harmdoer, thereby tending to invoke language about a victim. For that matter, the victim's harm might acquire special salience in numerous ways (e.g., similarity to the observer, personal relationship with the observer, or just the observer's capacity for empathy). We emphasize, however, that the "deed" done by wrongdoers actually has two victims rather than one: a person or persons harmed (in the sense that morality involves social relations and due regard for other persons) and the structural integrity of the moral system itself. On the one hand, of course, the moral structure itself does not actually suffer harm: It exists as an abstraction, in the world of pure thought, as an ideal. On the other hand, the "breaking the rules" connotations do imply a sense in which something once treated as if impenetrable (or supposed to receive that treatment) now shows itself as porous. Someone has escaped its boundaries, so "breaking" means "breaking through." We might use the metaphor of cell bars in a jail: When a jail "breakout" occurs, nothing in the

238

FOLGER AND BUTZ

jail might exist now in a literally "broken" state (e.g., the bars remain intact, but a prisoner slipped through them or used subterfuge that precluded their effectiveness). In using such metaphors and figures of speech within our argument, we obviously cannot claim the status of deductive proof for our conclusions. Conceding that point, we now play our psychological gambit with even greater transparency. The remaining discussion, in other words, relies even more heavily on taking for granted certain kinds of assumptions about human nature that others certainly might question. We hold that natural selection has favored an evolution of moral emotions. People experience strongly aroused feelings when they perceive events in moral terms. It may not involve too much exaggeration to suggest that humans treat morals anthropomorphically with respect to authoritative force: The moral injunction is "the final authority" on the matter. By this psycho-logic, then, "breaking" the "rules" gains the anthropomorphic status of regicide. If the AR connotations of morality (as supreme ruler) have indeed affected human moral emotion as we speculate, then the reactions to an unfair abuse of power in the dictator-game data collected by Turillo et al. (2002) make even more sense than ever. Moreover, the greater the truth to the speculations, the richer the implications. We close by noting just a few as related to organizational justice. First, we acknowledge explicitly and quite readily that more than one type of characterization could faithfully capture the nature of the transgression in the Turillo et al. (2002) studies as it affected the responses of the research participants. That is, we posit no need for them to have consciously perceived the violated norm as having to do with AR. Frankly, we would be surprised if any participant ever referred to the abuse of authority as the essence of the transgression. Mentioning allocation norms of distributive justice seems much more likely. On those grounds, the violation could have invoked EM (we both had an equal chance of being dictator—so your having received that role by random selection does not entitle you to any more than me) or perhaps a related MP variant (no differential level of inputcontribution qualifies you for more than me). Because such a situation involves a decision-maker's choice among alternative allocations, it makes sense to apply norms of distributive justice to assess the fairness of the amounts received by each party. The relevant norm, depending on various aspects of the perceiver interacting with the situation, might draw from any of the four relational models. Nevertheless, we argue that even when a perceiver compares the actual distribution with a normative standard for the distribution by applying a relational model other than AR, the AR model still operates in various ways for various reasons. First, Fiske has always insisted that no situation need invoke only a single model. Just because one model provides a normative standard for dis-

7. MORAL ANTIPATHY TOWARD THE POWERFUL

239

tributive justice, that does not make every other model ineligible from also having some relevance. Second, allocations made by decision makers automatically involve an AR-like (at least implicitly hierarchical) structure to the social situation because someone with the power and authority to decide the allocation amounts must necessarily rank higher that someone who can only wait until the decision maker makes the allocation results known. Third, as already implied, nothing prevents two violations from occurring at the same time, both as a direct consequence of the same event—that is, such an event violates norms drawn from two different models at the same time. Ire toward the $18-$2 dictator in the Turillo et al. studies, for example, might have encouraged punishment not only because that person violated an EM (or perhaps MP) norm by not making the evenly balanced choice of $10-$10, but also because that person acted as if corrupted by power (failed to show self-restraint when granted total autonomy over someone else's welfare). That, simply put, is the whole gist of our claim—that violating any moral mandate has an inherent structure analyzable in AR terms as a violation of rank privilege, thereby doing something that exceeds the authority of one's rank. Again, by definition, the moral mandate is the ultimate authority, so breaking a moral rule is analogous to killing a ruler. Using the language of regicide may sound just as overly anthropomorphic as we said it would, but we do so because Fiske emphasizes regicide as the quintessential or worst wrong in the AR category.

EPILOGUE In place of a conclusion or summary, we use this section to make some brief comments related to the tone of what we've tried to say but perhaps not so derivable directly from them. Much of what we have said boils down to a claim about a set of default tendencies that investigators should expect to find in the workplace when people refer to injustice, either as victims (e.g., employees who claim their bosses treat them abusively) or as observers (e.g., a public outcry because of perceptions about corporate misconduct at Enron or Arthur Anderson). First, recall our discussion regarding the "many faces" of AR. We noted that AR does not automatically entail considerations of power (e.g., coercive access to resources). For example the alternative perspectives that we mentioned (Henrich & Gil-White, 2001; Ruben, 2000) emphasized how prestige might be the salient ranking element in effect, or how the hierarchy might in a sense be simply almost a side-effect created out of the exigencies of finding ways to coordinate complex activity. If our central conjecture is correct, however, the employees who participate in such hierarchies might

240

FOLGER AND BUTZ

not view them through those lenses of prestige or productivity. Just because a hierarchy objectively is of a certain type (e.g., based on scientific fame, such that the physical strength of a Stephen Hawking would not allow him to exert coercive force abusively) does not mean people participating in it perceive it that way. We hypothesize that part of the ancient and ongoing tension between labor and management stems from the built-in tendency of preparedness for anger when subjugation seems possible. In part we took some initial inspiration from Boehm (1999), who has sifted through an impressive range of evidence to conclude that people have an inherent tendency to "proscribe the enactment of behavior that is politically overbearing" (p. 43). Another account of his summarizes the roots of the position this way: "Five million years ago, an ancestral ape was in a position to engage in protomoral behaviours as it dealt with its dislike of alpha bullying" (Boehm, 2000, p. 98). But Boehm does not claim that humans hate authority; in fact, he believes that human nature contains within it material whose expression can yield tyranny or democracy, despotism or egalitarianism—and by the same token, a more-or-less shared acceptance of either as legitimate under a given set of circumstances. No matter how strong the ordinary degree of acceptance for a hierarchically ranked authority as a governance structure, however, there can come a time when those lower down find those above to have exceeded their authority. Boehm focuses considerable attention on that key phenomenon as one that tends to draw a collective response such as an uprising. We, too, believe that human moral emotions such as judgmental antipathy and the condemnatory feelings about wrongdoers have emerged out of human history into an evolved set of dispositions. As Fiske (2002) notes, emotions can act as proxies for more calculative thought. Apparently natural selection did not favor having a cognitive committee give mental monologues in deep deliberation and debate over the Tightness or wrongness of some actions. Rather, a moral emotion evolved as a proxy that can take the place of such ruminations, thereby yielding a more-or-less instantaneous response. Think, for example, of how people were likely to have felt immediately after hearing that Susan Smith's children did not die from a kidnapping when she admitted that she drowned them by pushing her car into a lake. Or, in the organizational context, think about the time when a boss or colleague stole someone else's idea and tried to take credit for it. If we have conjectured correctly, the emotional reactions to such events show a commonality that stretches across (a) the Boehm anti-hierarchy thesis, (b) felt injustice over workplace mistreatment, and (c) the psychology of AR as a relational model. Nonetheless, we do not tie our position directly either to Boehm's or to Fiske's. Our idea of an AR-related response tendency in conjunction with violations of any and all relational models, for example, does not seem to fit

7. MORAL ANTIPATHY TOWARD THE POWERFUL

241

perfectly with some of Fiske's discussions concerning quite distinctive reactions depending on the particular model as a source for the violated norm. Similarly, we have no commitment to all of the assumptions Boehm makes in deriving his arguments for a human capacity in which subordinates rise up collectively for the sake of administering punitive sanctions against upstartism run amok (or as the dust jacket of his book puts it, he "postulates that egalitarianism is in effect a hierarchy in which the weak combine forces to dominate the strong"; Boehm, 1999; see www.hup.harvard.edu). Note, for example, that we referred to a boss or a colleague who steals an idea and tries to get all the credit for it or at least the lion's share (cf. AR as priority of access to resources, like the "I get to eat all I want first" disposition seen in the alpha-ranked among lions or other social species with dominance hierarchies). We do not, therefore, distinguish between (a) the lower ranked person who acts out-of-rank in a social climbing fashion and (b) the top-ranked person corrupted by power to the extent of becoming abusive. Conceptually, the form or process of both kinds of actions falls into the same type of AR mental category as a function of their antinormative or immoral quality. In closing, we note two final points. First, although crucial tests might well remain elusive as regards the specifics of evolutionary mechanisms responsible for current human proclivities, we think our speculations stand on more solid ground than most because the essence of our argument points to two parallel sources: On the one hand, we follow Panksepp (1998) in suspecting that the capacity for moral outrage stems from the nature of angry impulses selected for responding to predators. On the other hand, we also tie our analysis to the AR relational model that Fiske argues must have had evolutionary roots as well. Second, we echo Boehm (1999, 2000) in noting that human nature contains conflicting types of motivation: People have both "the innate tendency to dominate" and "the innate tendency to resent being dominated" (1999, p. 251). A wide range of sociopolitical forms can emerge as a result: "Some of our societies may be seen as 'cultures of rebellion' insofar as the limits of authority or domination are circumscribed. . . . Other societies amount to true 'cultures of dominance,' insofar as people not only accept but identify with a strong political authority that rules them" (1999, p. 252). We agree; after all, no evolved trait distinguishes humans more than the degree of plasticity and flexibility used to govern (though not always successfully) impulses from other evolved traits. We end on this note about inherent ambivalence and conflicting tendencies as a qualifier to anything else that we might have said earlier if it conveyed more rigidly inflexible tendencies. Nonetheless, despite Boehm's remarks about societies in which people accept and identify with authority, we lean more toward what he has written in other passages that also corresponds with Fiske's descrip-

242

FOLGER AND BUTZ

tion of AR. Although Fiske wrote of cultures whose acceptance of authority far outstrips the norm in Western cultures, he also (as is true of Boehm) noted that even extremes of allegiance have their limit. In our discussion of AR, we have likewise emphasized that humans seem to accept the legitimacy of authority only up to a point. Perhaps Fiske (1991) has put it best in the following words about the precedence and deference of rank among the Moose of Burkina Faso in West Africa: "Moose leaders have tremendous authority, authority that is, in principle, virtually unlimited within its traditional sphere of prerogatives" (p. 310). The catch lies in the qualifiers in principle and within its traditional sphere. In fact, a similarly hedged description fits modern workplace organizations just as readily as it does the political organization of those tribal peoples. Both phrases imply limits. Power and authority, no matter how nearly infinite, stop somewhere. Among humans, hoipolloi know that fact instinctively. When those in power forget and yield to the impulses conveyed by the quotes with which we opened this essay, punitive urges to put leaders "back in their place" help to solve the 3QC dilemma we made a central theme. Exploring the many factors that contribute to variations in the actual dynamics of such situations remains a topic worthy of much work yet to come.

REFERENCES Adams, J. S. (1965). Inequity in social exchange. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 2, pp. 267-299). New York: Academic Press. Barkow, J. (1989). Darwin, sex, and status: Biological approaches to mind and culture. Toronto: University of Toronto Press. Berscheid, E., & Walster, E. H. (1978). Interpersonal attraction (2nd ed.). Reading, MA: AddisonWesley. Bies, R. J. (1987). The predicament of injustice: The management of moral outrage. In L. L. Cummings & B. M. Staw (Eds.), Research in organizational behavior (Vol. 9, pp. 289-319). Greenwich, CT: JAI Press. Blau, P. M. (1964). Exchange and power in social life. New York: Wiley. Boehm, C. (1999). Hierarchy in the forest: The evolution of egalitarian behavior. Cambridge, MA: Harvard University Press. Boehm, C. (2000). Conflict and the evolution of social control. In L. D. Katz (Ed.), Evolutionary origins of morality: Cross-disciplinary perspectives. Bowling Green, OH: Imprint Academic. Damasio, A. R. (2001). Emotion and the human brain. In A. R. Damasio & A. Harrington (Eds.), Unity of knowledge: The convergence of natural and human science. Annals of the New York Academy of Sciences (Vol. 935, pp. 101-106). New York: New York Academy of Sciences. Darwin, C. (1872/1998). The expression of the emotions in man and animals. New York: Oxford University Press. deWaal, F. (1996). Good natured: The origins of right and wrong in humans and other animals. Cambridge, MA: Harvard University Press. Fiske, A. P. (1991). Structures of social life: The four elementary forms of human relations. New York: Free Press.

7. MORAL ANTIPATHY TOWARD THE POWERFUL

243

Fiske, A. P. (2000). Complementarity theory: Why human social capacities evolved to require cultural complements. Personality and Social Psychology Review, 4, 76-94. Fiske, A. P. (2002). Socio-moral emotions motivate action to sustain social relationships. Self and Identity, 1, 169-175. Folger, R. (1998). Fairness as a moral virtue. In M. Schminke (Ed.), Managerial ethics: Morally managing people and processes (pp. 13-34). Mahwah, NJ: Lawrence Erlbaum Associates. Folger, R. (2001). Fairness as deonance. In S. W. Gilliland, D. D. Steiner & D. P. Skarlicki (Eds.), Research in social issues in management (pp. 3-31). Greenwich, CT: Information Age Publishers. Folger, R., & Cropanzano, R. (1998). Organizational justice and human resource management. Thousand Oaks, CA: Sage. Frank, R. (1988). Passions within reason. New York: Norton. Gewirth, A. (1984). Ethics. Encyclopedia Brittanica. Haidt, J. (2003). The moral emotions. In R. J. Davidson (Ed.), Handbook of affective sciences. Oxford, England: Oxford University Press. Haidt, J., Rosenberg, E., & Horn, H. (2003). Differentiating diversities: Moral diversity is not like other kinds. Journal of Applied Social Psychology, 33, 1-36. Henrich, J., & Gil-White, F. (2001). The evolution of prestige: Freely conferred deference as a mechanism for enhancing the benefits of cultural transmission. Evolution and Human Behavior, 22, 165-196. Kahneman, D., Knetsch, J. L., & Thaler, R. H. (1986). Fairness and the assumptions of economics. Journal of Business, 59, s285-s300. Katz, L. D. (Ed.). (2000). Evolutionary origins of morality: Cross-disciplinary perspectives. Bowling Green, OH: Imprint Academic. Panksepp, J. (1998). Affective neuroscience: The foundations of human and animal emotions. New York: Oxford University Press. Rawls, J. (1971). A theory of justice. Cambridge, MA: Harvard University Press. Rubin, P. H. (2000). Hierarchy. Human Nature, 11, 259-279 Scanlon, T. M. (1998). What we owe to each other. Cambridge, MA: Harvard University Press. Skarlicki, D. P., & Folger, R. (1997). Retaliation in the workplace: The roles of distributive, procedural, and interactional justice. Journal of Applied Psychology, 82, 434-443. Smith, A. (1971). The theory of moral sentiments. New York: Garland. (Original work published 1759) Stenberg, C., & Campos, J. (1990). The development of anger expression in infancy. In N. Stein, B. Leventhal, & T. Trabasso (Eds.), Psychological and biological approaches to emotion (pp. 247-282). Hillsdale, NJ: Lawrence Erlbaum Associates. Tooby, J., & Cosmides, L. (1990). The past explains the present. Emotional adaptations and the structure of ancestral environments. Ethology and Sociobiology, 11, 375-424. Turillo, C. J., Folger, R., Lavelle, J., Umphress, E., & Gee, J. (2002). Is virtue its own reward? Selfsacrificial decisions for the sake of fairness. Organizational Behavior and Human Decision Processes, 89, 839-865.

This page intentionally left blank

P A R T

IV EMOTIONS, VALUES, AND MORALITIES

This page intentionally left blank

C H A P T E R

8 Proscribed Forms of Social Cognition: Taboo Trade-offs, Blocked Exchanges, Forbidden Base Rates, and Heretical Counterfactuals Philip E. Tetlock A. Peter McGraw Orie V. Kristel

Most psychological theories of judgment and choice work from the premise that people make judgments and decisions in their functional capacities as intuitive scientists or economists. As intuitive scientists, we are hypothesized to be on an epistemological mission of achieving as accurate a mental representation of the external world, and the causal relations that hold that world together as possible (Fiske & Taylor, 1991). As intuitive economists, we are posited to be on a utilitarian mission organized around the maximization of subjective expected utility (Kagel & Roth, 1995). One advantage of adopting these functionalist frameworks has been a ready supply of elegant normative models for assessing judgmental biases and errors, and for justifying attributions of irrationality that have the net effect of making researchers look smart and research participants look, well, not so smart. This chapter works from the premise that the aforementioned perspectives are far too restrictive (Tetlock, 2002). Ordinary people often resist the normative prescriptions of models anchored in the intuitive scientist-economist frameworks, even when these prescriptions have been carefully spelled out for them. We argue here that this resistance should not always be written off as incorrigible closed-mindedness. In some situations, people are better thought of not as intuitive scientists or economists but rather as intuitive theologians who are trying to protect sacred values from secular encroachments. We also propose a testable model of how people function as intuitive theologians. According to this sacred-value-protection model

247

248

TETLOCK, McGRAW, KRISTEL

(SVPM), certain categories of mental operations are anathema and off-limits because they require us to assign explicit finite valuations to values that our moral communities tell us should be treated as absolute, unquestionable, and unqualified commitments (Tetlock et al., 2000). People are just not supposed to think in certain ways, and as soon as one senses that others have broken the taboo, they become targets of moral outrage (as, indeed, so do those members of one's moral community who fail to censure those who violate the taboo—there is a meta-norm to enforce norms; Fiske & Tetlock, 1997). And, as soon as you sense that you yourself have started to stray down a proscribed mental pathway and begun to think the unthinkable, you are supposed not only to stop but to morally cleanse yourself: to reaffirm your commitment to shared values and your status as a member in good standing of the moral standing of the community. Of course, it can be objected that all this is not very parsimonious. It is simpler to work with a few functionalist frameworks that provide welldefined normative benchmarks for gauging rationality. The critics are right that there is a loss in parsimony. But the loss can be minimized by advancing reasonably precise and testable models that specify the conditions under which moral outrage and moral cleansing responses are likely to be activated. One of the best ways of doing that is to link the SVPM to a comprehensive theory and taxonomy of the relational schemata that people use to judge the acceptability of thought, feeling, and action in a wide range of cultural settings. Alan Fiske's (1991) relational theory fits this bill very well indeed, and we return to it in the closing section of this chapter to organize and, in certain senses, explain the sharp, qualitative content boundaries that people often place on the acceptability of certain forms of social cognition. We focus on four specific examples of such sharp qualitative breakpoints, contexts within which the SVPM predicts that patterns of thinking that, from an intuitive scientist or economist perspective, should be unproblematic but will provoke both moral outrage and moral cleansing. The four forms of proscribed social cognition are taboo trade-offs, blocked exchanges, forbidden base rates, and heretical counterfactuals. Taboo Trade-Offs Trade-off reasoning is fundamental to micro-economic and decision-theoretic conceptions of rationality. We live in a world in which we cannot always get what we want so we must strike compromises among our objectives: consumption versus saving, quality versus price, and work versus family. The SVPM posits, however, that people consider a willingness to entertain certain types of trade-offs, known as taboo trade-offs, not as evidence of cognitive sophistication, but rather as evidence of moral depravity

8. PROSCRIBED FORMS OF SOCIAL COGNITION

249

(Tetlock, Peterson, & Lerner, 1996). A taboo trade-off can be said to exist whenever entering one value into the trade-off calculus subverts or undercuts the other value. To compare is to destroy. The mental process of comparison can itself be morally corrosive. The longer one spends contemplating the dollar value of one's children, or friendships, or loyalty to one's country—values and obligations that are supposed to be sacred and categorically nonfungible—the more one reveals oneself to be the type of person who simply does not understand the meaning of parenthood, or friendship, or citizenship. It is useful to contrast taboo trade-offs, which characteristically entail comparisons of the relative importance of secular values (such as money, time, and convenience) and sacred values (that are supposed, at least for rhetorical purposes to be infinitely significant), with two other broad categories of trade-offs: routine and tragic. Routine trade-offs entail secular-secular value comparisons and are integral parts of everyday life. We execute such trade-offs every time one goes to the supermarket. Tragic trade-offs entail comparisons of sacred values—for example, one group of lives versus another, or life versus honor—and arise, almost by definition, only at dramatic turning points in life. Tragic trade-offs can be excruciatingly painful but they are widely recognized as a dreadful necessity. In one set of studies, Tetlock, Kristel, Elson, Green, and Lerner (2000) compared reactions to taboo and tragic trade-offs. In one scenario, people judged a healthcare administrator who had to make one of two choices: either between saving the life of one boy or another boy (a tragic trade-off) or between saving the life of a boy and saving the hospital a great deal of money (taboo trade-off). This experiment also manipulated two other independent variables in a 2 x 2 x 2 factorial design. The second independent variable was whether the administrator found the decision easy and made it quickly or found the decision difficult and took a long time. The third independent variable was whether the decision maker decided to save the life of a particular boy, Johnny, or he opted for the other alternative (either saving the other boy or saving money). The results revealed the predicted second-order interaction. The administrator in the taboo trade-off condition who thought the decision easy and chose Johnny quickly was judged least negatively whereas the administrator in the taboo condition who found the decision difficult and eventually chose the hospital was judged most negatively. By contrast, people judged the decision maker who worked through the tragic trade-off slowly more positively than the decision maker who made up his mind quickly, regardless of which boy he chose to save. These results show that lingering over a taboo trade-off, even if one ultimately does the "right thing" and saves the child's life, makes one a target of a significant moral outrage (relative to what one would have received if one had made the decision quickly). Lingering over a tragic trade-off simply under-

250

TETLOCK, McGRAW, KRISTEL

scores that one understands the gravity of the issues at stake. A quick decision would indeed be disrespectful to one of the competing sacred values (and to those who advocate those values in contexts of legal or political accountability). Taboo trade-offs are also morally contaminating. To observe a taboo trade-off without condemning it is to become complicit in the transgression. As a result, people often try to cleanse themselves. Tetlock et al. (2000) documented the greatest willingness to volunteer to increase organ donations among participants who had thought both that the decision maker had made the "wrong" choice in a taboo trade-off and that the decision maker had a hard time making up his mind. Thinking a long time about taboo trade-offs, even if one makes the "right" decision, taints one's social identity. It is as though observers reason "anyone who thinks that long about the dollar value of a child's life is morally suspect." By contrast, participants in the tragic trade-off condition, in which the administrator had to choose which child should live seemed to reason: "The longer the deliberation, the greater the respect shown for the solemnity of the decision." A big question, that economists are especially quick to raise, concerns the sometimes jarringly sharp disjunction between what people say about the importance of protecting sacred values and the compromises that decision makers must actually make in the real world of scarce resources. In the final analysis, someone must somehow set priorities, a process that— however distasteful we find it—requires attaching at least implicit monetary values to sacred values. We cannot devote 100% of GDP to taking care of the sick and elderly, or to educating children, or to uplifting the poor, or to advancing the cause of liberty and justice abroad. If these decision makers are to escape becoming targets of the moral wrath of the populace, some combination of two things must be true: Decision makers must be pretty skilled at obfuscating what they are doing or the citizenry must be willing to refrain from trying to peer through the rhetorical smokescreens that decision makers put up. The toxic-waste experiment, reported in Tetlock (1999), addressed these issues. The specific scenario used was inspired by an actual federal law— the Superfund Act: A government program cleans up toxic waste sites to the point where they pose zero risk to public health. Last year, the program saved an estimated 200 lives at a cost of $200 million. The Danner Commission—whose mandate is to improve the efficiency of government—investigated the program and recommended a set of reforms that were implemented. As a result, this year the program could save an estimated 200 lives—just as many as before—but at a cost of only $100 million, a saving of $100 million. If the government kept funding the program at last year's budget level of $200 million, the program could save an estimated 400 lives.

PROSCRIBED FORMS OF SOCIAL COGNITION

25 I

Now the Danner Commission recommends redirecting the saving of $100 million to other uses, including reducing the deficit, increased funding for programs to stimulate economic growth, and lowering taxes, (p. 254)

In one experimental version, the story ends with the Commission avoiding any reference to dollar-life trade-offs and simply offering a vague deontic rationale: "Based on its analysis of these options, the commission concludes that "morally this is the right thing to do." What do you think? Should the government keep funding the program at $200 million or should it redirect the $100 million saved to other priorities?" In another experimental condition, participants learned that "based on its analysis of the options, the Danner Commission has concluded that the cost of saving the additional 200 lives is about $500,000 per life—a cost that it still considers too high and one that cannot be justified given other needs and priorities. The Commission therefore recommends redirecting the saving of $100 million to other uses, including reducing the deficit, increased funding for programs to stimulate economic growth, and lowering taxes" (p. 255). Support for the Danner Commission recommendation hovers at 72% when the trade-off is masked by a deontic rationale but falls to approximately 35% when the trade-off explicitly violates the normative injunction against dollar valuations of human life. Additional experimental conditions address alternative interpretations. One argument is that people object not to attaching a dollar value to human life but rather to the particular cost estimate that the Danner Commission endorsed. There is nothing inherently wrong about pricing human life but one should get the price right. To explore this possibility, participants reacted to a Commission that engaged in cost-benefit reasoning and concluded that funding should remain at $200 million because $500,000 per human life is a cost worth bearing. In still other conditions, $500,000 is a bargain; the Commission would have been willing to pay either $1 million or $10 million (at which price the entire federal budget for fiscal year 1996-1997, $1.7 trillion, would allow us to save only 170,000 lives). Although people judge the Commission more favorably when it places a higher dollar value on human life, they still judge it less favorably, and substantially so, than when the Commission offers a deontic rationale either for keeping funding at the $200 million level or for redirecting spending to other priorities. Another possibility is that people object to utilitarian reasoning in general and not just to the violation of this specific taboo trade-off. To test this possibility, participants reacted to a Danner Commission that either redirected funding or kept funding constant and did so "after weighing all the relevant costs and benefits." Interestingly, vague utilitarian reasoning was

252

TETLOCK, McGRAW, KRISTEL

every bit as effective in providing political cover as vague deontic reasoning. Taboo trade-offs need to be cloaked (Calabresi & Bobbit, 1978) and either Kantian or utilitarian cloaks will do the job. People objected to the explicit spelling out of the trade-off. Doing so tarnished the image of a previously well-regarded reformist commission, rendering suspect its entire policy agenda. These methodological variations show that previously popular politicians and acceptable policies can be transformed into objects of scorn by revealing that the politicians performed taboo mental calculations in reaching their conclusions. The damage to one's political identity can be so severe that even possessing an otherwise winning utilitarian argument is not sufficient to prevail in the court of public opinion. The "contamination" findings have suggestive implications. As long as the public insists on treating certain values as sacred and beyond "monetization," politicians will be forced to resort to deceptive rhetorical subterfuges. To let down those subterfuges, to act as though Social Security or Jerusalem or Kashmir is just another issue to be compromised and bargained out, is to open oneself to demagogic attacks, vilification, and ridicule. A mark of a stable polity may well be the willingness to forgo the short-term advantages that come from accusing opponents of selling out secular values. Systems in which power rotates reasonably regularly among major contenders create incentives for such restraint. A mark of an unstable system may well be the escalation of charges of taboo trade-offs and the resulting "demonization" of adversaries that such charges can fuel. Blocked Exchanges and Relational Moderators of the Endowment Effect

Blocked exchanges are a special case of taboo trade-offs in which observers express their disapproval of the "commodification" of goods and services that they deem to have special moral or emotional significance. From a micro-economic point of view, there is really only one good justification for blocking mutually consenting parties from entering into contracts: namely, negative externalities or side effects on third parties. From the standpoint of the SVPM, however, these negative externalities need not take the form of tangible material damage to identifiable individuals; they can take invisible symbolic forms and the victims can be as abstract as the community, society, the nation, the Volk, disembodied moral abstractions, or, for that matter, God. In late 20th century America, there is a great deal of consensus about what types of social exchanges should be blocked. However, Tetlock et al. (2000) found that sharp, ideologically charged disagreements do occasion-

8. PROSCRIBED FORMS OF SOCIAL COGNITION

253

ally erupt. There is deep disagreement, for example, over the legitimacy of buying and selling medical care or access to high quality legal representation in court. For many on the left, these transactions come perilously close to buying and selling lives and justice. For those on the right, such transactions are not morally distinguishable from buying a house or food. Both left and right do, however, largely agree—in late 20th century American political culture—that buying and selling adoption rights for orphans or buying and selling draft obligations to perform military service are unacceptable (only within the rarefied ideological subculture of libertarianism to objections to these types of exchanges or trade-offs disappear). Tetlock et al. (2000) surveyed reactions to routine and blocked exchanges across a broad spectrum of student political activists at Berkeley in the early 1990s. Everyone but libertarians was outraged by efforts to "monetize" babies, body parts, and basic rights and responsibilities of democratic citizenship. There was far less outrage toward routine trade-offs but outrage gradually built as we moved leftward on the political spectrum toward socialists who not only found controversial topics like surrogate motherhood unacceptable but who also had problems with buying and selling uncontroversial commodities such as houses and food or borderline controversial commodities such as medical care and legal representation. Devout egalitarians tend to see such exchanges as inherently inequitable because they put the poor at a profound disadvantage (and because they seem to carry the implication that the lives and rights of the poor are worth less than those who can pay large sums for doctors and lawyers). A high profile example of a social exchange that people were unsure how to evaluate arose during the Clinton presidency when questions were raised about the large number of big campaign donors who were sleeping in the Lincoln bedroom (a vivid image of a sacred site subjected to secular contamination; McGraw & Tetlock, 2004). Reaction to the disclosures was initially negative among both supporters and detractors of the Clinton presidency. The unadorned facts elicited moderate outrage from both camps. But at least among Clinton supporters, the reaction was subject to quite dramatic political manipulation as a function of the semantic or relational framing of the exchange. Outrage among Clinton supporters was substantially attenuated when the facts were accompanied by a rationale—the one, incidentally promoted by the White House—that invoked a friendship as opposed to economic norm, a rationale that affirmed the right of friends to do and return favors for friends. Invoking this equality-matching or reciprocity rule reduced outrage to politically tolerable levels. Outrage among both supporters and detractors went up substantially, however, when the facts were accompanied by an explicit market-pricing rationale for the decisions as to whom to invite to stay in the Lincoln bedroom: Why not let the laws of supply and demand work their will?

254

TETLOCK, McGRAW, KRISTEL

We have also done some small-scale experiments that explore the political-psychological roots of objections to blocked exchanges. A controversial issue concerns exactly what taboo blocked exchanges and "taboo trade-offs" actually are. Are they taboo in the primal Polynesian meaning of the term, absolute, categorical, and unreasoned aversions that have the power to contaminate those who transgress boundaries that demarcate the sacred from the secular? Are sacred values better viewed as merely pseudo-sacred, as polite social fictions to which we pay lip service and honor at ceremonial occasions, but that we abandon as soon as upholding the sacredness of the value becomes burdensome and a convenient justification for jettisoning the value comes along? In the policy-revision paradigm, we have explored how easy or difficult it is to "talk people out of" their aversion to secular encroachments on sacred values (see also Tetlock, Peterson, & Lerner, 1996). The first step was to document, once again, that the overwhelming majority of respondents object, often emphatically, to proposals that would allow people to buy and sell body organs and adoption rights for children. The second step was to assess whether these taboo trade-offs are truly taboo in the primal Polynesian sense of the term, prohibitions, like that on incest, that require no further explanations and that permit no exceptions. We found that approximately 60% of those who object to "marketizing" babies and body organs saw no need for a reason—even when held accountable and pressed for one—beyond the blanket condemnation "the policy is degrading, dehumanizing, and unacceptable." We take this assertion to be a simple reassertion of the taboo—an "ugh" reaction that people find it odd to be asked to justify. It is as though we had reached the backstop of these belief systems—the de gustibus line of defense beyond which people find further queries to be absurd. But what about the 40% of respondents who offered rather specific reasons for judging the proposal unacceptable? Informal content analysis of these reasons revealed that many objections had an egalitarian character. One common concern about medical transplants was the theory that the poor would be driven into deals of desperation. They will need money so badly that they will feel compelled to submit to dangerous surgical interventions—donating part of one's liver or a whole kidney—that the well-off would never seriously consider. A widely held fear about baby auctions was that only healthy and attractive babies would draw bidders. A related concern was that the price of organs of babies would sky rocket and the market would cater only to the well-off. Another class of objections was more pragmatic. Some people were not convinced that all other options have been investigated for increasing the supply of body organs for desperately needed transplants or the supply of parents for babies in need of families. These respondents believed that "more civilized" solutions to the policy problem could be found.

8. PROSCRIBED FORMS OF SOCIAL COGNITION

255

Our research on the cognitive substrate of blocked exchanges assessed how much people change their minds about markets for body organs and for babies when we revised the proposals to address substantive objections that people raised. Each policy revision was designed to minimize what economists might call the moral externalities of permitting a particular taboo trade-off. For instance, one set of questions included: "Would you still object to markets for body organs: (a) if you lived in a society that had generous social welfare policies and never allowed the income for a family of four to fall below $32,000 per year (explaining the concept of inflationadjusted 1996 dollars)?; (b) if society provided the less well-off with generous organ-purchase vouchers that increased in value as recipient income decreased (the poorer the recipient, the larger the voucher)?; (c) if they could be shown that all other methods of encouraging organ donation had failed to produce enough organs and that the only way to save large numbers of lives was to implement a market for body organs?" Another series of questions probed reactions to permitting all qualified parents to bid for adoption rights in one scenario, participants were assured that if only attractive and healthy babies attracted high bids, then the money raised through the auction would go to create incentives to encourage other parents to adopt less attractive and less healthy babies as well as to improve the conditions of institutionalized life for these children. In another scenario, respondents were assured that poor people would not be prevented from adopting children because the program would provide generous vouchers to all poor people who want to adopt children, thereby permitting them to compete with the affluent would-be adopters (again, explaining the concept of vouchers). In a third scenario, participants were assured that all less radical solutions to the problem of non-adopted children had been explored and proven inadequate. How successful were we in eroding the taboo through this series of counterfactual thought experiments? Whereas nearly 90% of respondents initially objected to each proposal, opposition fell to approximately 60% in the hypothetical worlds with protection for the poor and unattractive and a guarantee that all other options had been thoroughly explored. For nearly half of the population, the term taboo is apparently too strong. They are willing to consider—in a quasi-utilitarian manner—conditions under which the benefits of the proposal might conceivably outweigh the costs. But for the other half of respondents, the term taboo still seems apt. They are prepared to ban exchanges that both donors and recipients believe would leave them better off and that, arguably would leave many third parties better off as well. In their eyes, taboo trade-offs remain an affront to the collective conscience of society. As one respondent protested: "what kind of people are we becoming?" Consequentialist arguments—whether they invoked egalitarian or efficiency concerns—missed the point: This woman's

256

TETLOCK, McGRAW, KRISTEL

sense of personhood required categorically rejecting the taboo trade-offs. In March and Olson's (1989) terminology, the logic of obligatory, not anticipatory, action governed her choices. She knew with existential certainty who she was and she knew that she did not fit into a social world that permitted such transactions. These results also have suggestive political implications. Once a tradeoff has become culturally defined as taboo, it is an uphill battle to render it respectable. Here is another reason why politicians in stable systems should be reluctant to initiate accusations of taboo trade-offs by opponents. Such accusations greatly complicate reaching integrative solutions to policy disputes. Research on blocked exchanges need not be limited to big, controversial questions bearing on the limits of markets; it can also proceed at more personal and interpersonal levels of analysis that focus on the social, moral, and emotional significance people invest in personal possessions. It is a truism of microeconomic theory that, when income effects are small, there should be negligible differences between a person's maximum willingness to pay (WTP) for a good and minimum compensation demanded for giving up the same good (willingness to accept or WTA). [Markets cannot function well, they cannot clear, if people suddenly insist that what they now happen to own is worth twice what they thought it was prior to owning it.] Daniel Kahneman, Richard Thaler, and colleagues have, however, established in a series of compelling experiments that people often demand much more to give up an object that they already possess than they would be willing to acquire it, the definition of the endowment effect (Kahneman, Knetsch, & Thaler, 1990; Thaler, 1980). This effect is closely related to, and may even be a by-product of, the well known status quo bias (Samuelson & Zeckhauser, 1988), a preference for the current state of the world that biases consumers in their functional capacities as intuitive economists, against both buying and selling goods. The most influential explanation for these anomalies invokes an asymmetry of value that Kahneman and Tversky (1984) call loss aversion, the disutility of losing an object is greater than the utility of gaining it. In a social environment strictly regulated by market-pricing principles, it does not matter to whom you sell something or from whom you buy it. The identities of buyers and sellers are interchangeable. Of course, in much of social life, objects take on deep emotional and symbolic significance as a function of the people or institutions from whom we obtain them. A commonplace object such as a pen may take on diverse meanings depending on whether the pen was a gift from a member of a group with one has communal sharing connections (e.g., a family) or a group to which one is bound by authority-ranking connection (such as a military or other institutional hierarchy) or from individuals to whom one is connected by equality-matching connections (reciprocity networks of friends and colleagues).

8. PROSCRIBED FORMS OF SOCIAL COGNITION

257

In a series of studies, McGraw, Tetlock, and Kristel (2003) have experimentally manipulated the relational context within which an object was originally acquired, and shown context to be a powerful moderator of the degree of discrepancy between WTP and WTA. Whereas the customary size of the endowment effect in the laboratory is a 2:1 ratio of WTA:WTP, the ratio soared for endowments charged with communal-sharing meaning. Figure 8.1 shows, for a related experiment, the cumulative percentage distributions of buying and selling prices, which are log transformed to preserve the original ordering of dollar values but still place them on a scale that incorporates all responses (no matter how extreme) into a reasonable display (McGraw & Tetlock, 2004). As expected by the endowment effect, the data almost perfectly exemplify stochastic dominance: Selling prices exceed buying prices throughout the range of values. The difference between the distributions, which is denoted by the shaded area, represents the disparity of WTP and WTA for each relational context. The area between the curves is greatest in the CS relationship and smallest in the MP relationship. These results indicate that people draw a sharp normative line in the sand between particularistic and universalistic relationships—that is, relationships in which people care very much about the identity of the actor from whom they receive goods and services and relationships in which people are indifferent to the identity of actors. The ever-increasing discrepancies between WTP and WTA, moreover, tell only part of the story. We also documented a sharp surge in cognitive confusion and moral outrageas well as principled refusals even to consider assigning dollar values— when people were asked to sell an object that had been acquired in communal-sharing, authority-ranking, or equality-matching relationships. Finally, as one would expect if basic norms were being breached, anger and resentment toward proposals to treat objects acquired in particularistic relationships as fungible was especially intense when observers believed that the person making the proposal was aware of the relationship history behind the possession of the object, and was insultingly implying that the pos-

FIG. 8.1. Cumulative percentage distributions of buying prices (left edge of shaded areas) and selling prices (right edge of shaded area). The size of the shaded area represents the difference in buying and selling prices.

258

TETLOCK, McGRAW, KRISTEL

sessor was the "sort of person" who would even think about selling a symbolically significant object. Forbidden Base Rates

Trade-off reasoning is widely viewed as a minimal prerequisite for economic rationality. Maximizing utility presupposes that people routinely factor reality constraints into their deliberations and explicitly weigh conflicting values. We find just as solid a normative consensus that good intuitive scientists and or statisticians should use base rates as we do that good intuitive economists should confront trade-offs. Decision theorists routinely invoke Bayesian probability theory as the appropriate set of principles for aggregating base rate and case-specific information (Koehler, 1996; Tversky & Kahneman, 1982). And researchers routinely find that people often deviate from Bayesian prescriptions and fail to use base rates as much as they should. The most influential explanation for this underuse of base rates is that people over-rely on error-prone heuristics such as representativeness (Kahneman & Tversky, 1972). The SVPM reminds us that over-relying on error-prone heuristics is not the only pathway to base-rate neglect. In a society committed to racial, ethnic, religious, and gender egalitarianism, forbidden base rates could include observations bearing on the disproportionately high crime rates and lower educational test scores of certain categories of human beings. People who use such base rates in judging individuals are less likely to be applauded for their skills as good intuitive statisticians than they are to be condemned for their moral insensitivity. Tetlock et al. (2000) report a set of experiments that demonstrate this point. In the red-lining experiment, an insurance executive must decide whether to sell policies in neighborhoods designated as high risk or as low risk. Sometimes the low risk towns have a disproportionately large AfricanAmerican population; sometimes they do not; sometimes there is no reference whatsoever to race. The results indicated that respondents who thought of themselves as racially liberal were especially likely to favor selling policies across neighborhoods for the same price (ignoring the base rate) when the base rate was correlated with racial composition of neighborhood. In other words, a base rate that was acceptable to use when race was not an issue became unacceptable when race became an issue. In a follow-up experiment, we examined what happens when people use a base rate to recommend a pricing policy for insurance policies and then later discover that the base rate either is or is not correlated with racial composition of neighborhood. It was hypothesized that racial liberals should find it especially morally threatening and dissonant to discover that they have recommended a pricing policy that works to the disadvantage of

8. PROSCRIBED FORMS OF SOCIAL COGNITION

259

a traditionally disadvantaged group. And this is exactly what happens. Racial liberals are more likely to change their pricing recommendations when they discover that the base rate that influenced those recommendations was correlated with race; and they were also more likely to engage in moral cleansing (express greater interest in volunteering for a civil rights cause as well as a cause unrelated to civil rights, helping to find a missing person). Heretical Counterfactuals Good intuitive scientists are supposed to use the covariation principle in gauging the plausibility of causal hypotheses, and be on the lookout for causal variables that regularly covary with effects of interest (Tetlock & Belkin, 1996). When full covariation information is not available, and in life it rarely is, they should be willing to engage in counterfactual thought exercises in which they try to figure out what would have happened if a cost that had been present had been absent or vice versa. The SVPM posits that cognitive theories of counterfactual reasoning need to acknowledge the emotionally charged normative boundaries that religious and political groups erect against what-if speculation. Particularly irksome are counterfactuals that apply the normal laws of human nature and of physical causality to heroic, larger than human, founders of the movement. Consider the reaction of the Ayatollah Khomeini to Salman Rushdie's heretical counterfactual that invited readers to imagine that the prophet Muhammed kept a company of prostitutes. For this transgression, the theocratic regime in Iran sentenced Rushdie to death (the ultimate expression of moral outrage). In one experiment, Tetlock et al. (2000) examined the reactions of fundamentalist Christians to counterfactual conjectures that undermine faith in the "unique historical status" of Jesus Christ—the view that Jesus was God made man, divine yet also human, and that the events of his life as revealed in the New Testament gospels were the product of a divine plan and shielded from the random contingencies that distort the lives of ordinary mortals. From a fundamentalist perspective, the life of Christ had to unfold as it did and devout believers should react indignantly to counterfactuals that imply otherwise: "If Joseph had left Mary because he did not believe she had conceived a child with the Holy Ghost, Jesus would have grown up in a one-parent household and formed a different personality." From a secular-intuitive scientist point of view, such counterfactuals are reasonable. They merely introduce causal schemata that virtually all of us apply reflexively in everyday life to a text that many people deem to be divinely inspired. Tetlock et al. (2000) compared the reactions of low, moderate, or high scorers on the religious fundamentalism scale both to heretical counterfactuals (that apply regular causal principles to exceptional figures) and to

260

TETLOCK, McGRAW, KRISTEL

secular counterfactuals (which apply regular causal principles to regular people). They found virtually no connection between fundamentalism and reactions to secular counterfactuals; but powerful tendencies for fundamentalists to resist heretical counterfactuals, to express moral outrage toward them, to express emotions such as disgust, to ostracize members of their religious in-group who endorse such counterfactuals, and to engage in moral cleansing by expressing renewed intentions to attend church regularly. It is also not hard to identify examples of heretical (or, at least, intensely controversial) counterfactuals that fall outside the religious domain. The historian John Charmley (1993) found one when he argued that Britain would have been better off //"Lord Halifax rather than Winston Churchill had been Prime Minister in the summer of 1940 when Hitler was expressing an interest in a peace settlement that, according to Charmley, would have demanded only fairly modest concessions and allowed Britain to remain more viable as a great power than it was in the world in which Churchill defied Hitler. The economist Walter Williams may have generated a heretical counterfactual when he argued against reparations for the enslavement of American blacks on the grounds that ancestors of Africans who were enslaved and brought to America have far higher per capita incomes than they would have had if they had not been enslaved and had remained in Africa. The novelist Stephen Frye generated an heretical counterfactual when he concocted a science-fiction scenario in which the best way for the heroes to save lives in the 20th century is to go back in time and prevent the death of Adolf Hitler because, it turns out, in the absence of Hitler, a far shrewder Nazi takes power in Germany and succeeds in accomplishing what Hitler failed to accomplish. Closing Thoughts

There do indeed appear to be sharp qualitative breakpoints in social cognition, and conceding this point does not mean conceding that the quest for parsimonious theory is hopeless. The breakpoints are not arbitrary. They become predictable when we join a psychological theory such as the SVPM to Alan Fiske's (1991) cross-cultural taxonomy of relational schemata. Taboo tradeoffs become taboo when they violate cultural conventions on norms that place limits of market-pricing rules; base rates become forbidden when they impede implementation of egalitarian efforts to repair past wrongs, thereby thwarting an equality-matching reformist agenda; heretical counterfactuals become heretical when they show disrespect for established religious or political authority, mocking authority-ranking prerogatives and sensitivities. These proscriptions make good relational and normative sense even if they look bizarre from a strictly intuitive-scientist or

8. PROSCRIBED FORMS OF SOCIAL COGNITION

261

economist framework. Here we are reminded of a quotation from Immanuel Kant that Isaiah Berlin (1991) adopted as the title for one of his books: "From the crooked timber of humanity no straight thing can be built." Grand reformist plans to straighten out thought, to induce all of us to apply the same normative guidelines across content domains and social context, are fated to run into sharp resistance from ordinary people who insist on thinking, feeling, and acting in seemingly incommensurable ways across spheres of activity and who appear content with ad hoc "compartmentalizing" solutions when conflicts are called to their attention.

REFERENCES Berlin, I. (1991). The crooked timber of humanity: Chapters in the history of ideas. New York: Alfred A. Knopf. Calabresi, G., & Bobbit, P. (1978). Tragic choices. New York: Norton. Charmley, J. (1993). Churchill, the end of glory: A political biography. London: Hodder & Stoughton. Fiske, A. P. (1991). Structures of social life: The four elementary forms of social relations. New York: Free Press. Fiske, A., & Tetlock, P. E. (1997). Taboo trade-offs: Reactions to transactions that transgress spheres of justice. Political Psychology, 18, 255-297. Fiske, S., & Taylor, S. (1991). Social psychology. New York: McGraw Hill. Kagel, J., & Roth, A. (1995). Handbook of experimental economics. New Jersey: Princeton University Press. Kahneman, D., Knetsch, J., & Thaler, R. (1990). Experimental tests of the endowment effect and the Coase theorem. Journal of Political Economy, 98, 1325-1348. Kahneman, D., &Tversky, A. (1972). Subjective probability: A judgement of representativeness. Cognitive Psychology, 3, 430-453. Kahneman, D., & Tversky, A. (1984). Choices, values, and frames. American Psychologist, 39, 341-350. Koehler, J. (1996). The base rate fallacy reconsidered: Descriptive, normative, and methodological challenges. Behavioral and Brain Sciences, 19, 1-53. March, J., & Olson, J. (1989). Rediscovering institutions. New York: The Free Press. McGraw, A. P., & Tetlock, P. E. (2004). Relational framing: The power to determine the acceptability of changes. Working paper, Princeton University, Princeton, NJ. McGraw, A. P., Tetlock, P. E., & Kristel, 0. V. (2003). The limits of fungibility: Relational schemata and the value of things. Journal of Consumer Research, 30, 219-229. Samuelson, W., & Zeckhauser, R. (1988). Status quo bias in decision making. Journal of Risk and Uncertainty, I, 7-59. Tetlock, P. E. (1999). Theory-driven reasoning about possible pasts and probable futures: Are we prisoners of our preconceptions? American Journal of Political Science, 43, 335-366 Tetlock, P. E. (2000). Coping with trade-offs: Psychological constraints and political implications. In S. Lupia, M. McCubbins, & S. Popkin (Eds.), Political reasoning and choice (pp. 239-263). Berkeley: University of California Press. Tetlock, P. E. (2002). Social functionalist frameworks for judgment and choice: People as intuitive politicians, theologians, and prosecutors. Psychological Review, 109, 451-471. Tetlock, P. E., & Belkin, A. (1996). Counterfactual thought experiments in world politics: Logical, methodological, and psychological perspectives. New Jersey: Princeton University Press.

262

TETLOCK, McGRAW, KRISTEL

Tetlock, P. E, Kristel, O., Elson, B., Green, M., & Lerner, J. (2000). The psychology of the unthinkable: Taboo trade-offs, forbidden base rates, and heretical counterfactuals. Journal of Personality and Social Psychology, 78, 853-870. Tetlock, P. E., Peterson, R., & Lerner, J. (1996). Revising the value pluralism model: Incorporating social content and context postulates. In C. Seligman, J. Olson, & M. Zanna (Eds.), Ontario symposium on social and personality psychology: Values (Vol. 8, pp. 25-51). Mahwah, NJ: Lawrence Erlbaum Associates. Thaler, R. H. (1980). Toward a positive theory of consumer choice. Journal of Economic Behavior and Organization, 1, 39-60. Tversky, A., & Kahneman, D. (1982). Evidential impact of base rates. In D. Kahneman, P. Slovic, & A. Tversky (Eds.), Judgment under uncertainty: Heuristics and biases (pp. 153-160). New York: Cambridge University Press.

C H A P T E R

9 Values and Emotions in the Relational Models Sonia Roccas Clark McCauley

In this chapter we examine the links between relational models and the psychology of values and emotions. Fiske (1991) introduced the relational models with considerable attention to their motivational implications, but we believe it is fair to say that the cognitive foundations and implications of the models have received by far the greater attention. Thus we bring the models together with recent research to explore two contentions: first, that the relational models express and embody different values, and second, that violations of different models elicit different emotions. As an introduction to the linkage between relational models and values, we begin with a classic study of conflicting values that can equally be interpreted as a study of conflicting models.

MODELS OF VALUES Solomon Asch's (1956) famous conformity paradigm forced subjects to choose between the evidence of their own eyes and the (supposed) evidence of their friends' eyes. This is usually interpreted as a value conflict, a conflict between honesty and independence on the one hand, and security and conformity on the other. Fiske's relational models offer an enlarged perspective in highlighting the relations among all three kinds of actor in Asch's experiments: the student subject, the students composing the bogus majority, and the experimenter himself. 263

264

ROCCAS AND McCAULEY

Subjects faced a conflict between their relationship with the experimenter, who asked them to report what they saw, and their relationship with the bogus majority. The first is Authority Ranking (AR) with experimenter as superior and subject as subordinate, and the second is Communal Sharing (CS) as the subject shares friendship or at least membership in a campus community with the bogus majority. The bogus majority faced a conflict between their AR relation with the experimenter and their CS relation with the subject. As experimenter, Asch faced a conflict between his AR relation with the students in his experiment, both the bogus majority and the subject, and his CS relation with these same students who, at least at the beginning of his experiments, shared his membership in the small community of Swarthmore College. This enlarged perspective helps understand an aspect of the results that is opaque to the usual interpretation about value conflict. A few subjects reported believing that the majority were correct, but gave the judgment of their own eyes anyway because this was the job the experimenter had given them. These few put their AR relation with the experimenter ahead of their CS relation with the bogus majority. A relational models perspective also leads to the recognition that the major power in Asch's experiment is the AR power of the experimenter with his student subjects. The students put AR with the experimenter ahead of CS pressure on two-thirds of the critical trials, and the experimenter deployed his AR power in recruiting the bogus majority and acknowledged his AR responsibility in debriefing the deceived subject after each session. The reader is invited to apply a relational-models perspective to Milgram's (1965) obedience experiment, where the real subject's relation to the supposed learner involves competition among three models: Equality Matching (EM; the rigged drawing for roles in the experiment), Authority Ranking (teacher responsible for maximizing learner progress), and Communal Sharing (concern for the welfare of the individual receiving shocks). Similarly the reader may apply a relational-models perspective to Darley and Batson's (1973) "good Samaritan" experiment, where the subject must choose between an AR relation with the experimenter who has asked the subject to hurry to another building, and a CS relation with the supposed victim lying moaning in a doorway on the path to that building. The point of these examples is that experiments justly famous in social psychology for posing value conflicts are easily and perhaps more revealingly interpreted as conflicts of relationship. This translation should not be surprising. Fiske presents his models as systems of reciprocating norms, norms that bring together prescription, expectation, and (usually) behavior. The prescriptive aspect of the models means that they are sources of value. Values play a role both in motivating and in justifying behavior (Seligman, Olson, & Zanna, 1996). They are beliefs about desirable goals and the

9. VALUES AND EMOTIONS

265

modes of conduct that promote them (cf. terminal vs. instrumental values, Rokeach, 1973). Values guide the selection or evaluation of actions, policies, people, and events (cf. Allport, 1961; Feather, 1995; Kluckhohn, 1951; keach, 1973). Many recent studies of values are couched at the group level, with comparison of the value priorities of different cultures or societies (Hofstede, 1991; Inglehart & Beker, 2000; Smith & Schwartz, 1997). In this chapter, however, we focus on the individual level to examine the links between relational models and individual differences in value priorities. In particular we focus on a theory of universals in the content and structure of human values developed by Schwartz (1992, 1994).

SCHWARTZ'S VALUE THEORY

Schwartz defines values as desirable, trans-situational goals, varying in importance, that serve as guiding principles in people's lives (see Schwartz, 1992, for a fuller elaboration; cf. Rokeach, 1973 and Kluckhohn, 1951). In developing a measure of values, Schwartz made active efforts to obtain values from many cultures; parallel versions were developed in three languages and at least two samples were obtained from each of 63 nations. The result was a paper-and-pencil inventory of 56 values, each represented as a word or phrase with brief explication in parentheses: For example, "POLITENESS (courtesy, good manners)." Respondents rate the importance of each value as a guiding principle in their life on a 9-point scale from opposed to my principles (-1), through not important (0), to of supreme importance (7). The asymmetry of the scale reflects the natural distribution of distinctions that individuals make when thinking about the importance of values, as observed in pretests when building the scale. Because values are typically seen as desirable, they generally range from somewhat to very important. The crucial aspect that distinguishes among values is the type of motivational goal each expresses. Schwartz derived 10 distinct motivational goals that he represents as value types, with each value type represented in terms of multiple specific values. Table 9.1 presents the definition of each value type in terms of its central goal, followed, in parentheses, by the specific values that primarily represent it. (Specific values are phrased as nouns or adjectives, corresponding originally to Rokeach's 1973 distinction between terminal and instrumental values, but this distinction has not been empirically supported in Schwartz's research.) Research with the value inventory led to a theory of the structure of relations among values (Schwartz, 1992). Actions undertaken in trying to achieve the goals expressed by values have psychological and practical consequences. Some values are incompatible, in the sense that actions un-

266

ROCCAS AND McCAULEY TABLE 9.1 Definitions of Motivational Types of Values in Terms of Their Goals and the Single Values That Represent Them

POWER: Social status and prestige, control or dominance over people and resources. (Social Power, Authority, Wealth, Preserving my Public Image) ACHIEVEMENT: Personal success through demonstrating competence according to social standards. (Successful, Capable, Ambitious, Influential) HEDONISM: Pleasure and sensuous gratification for oneself. (Pleasure, Enjoying Life) STIMULATION: Excitement, novelty, and challenge in life. (Daring, a Varied Life, an Exciting Life) SELF-DIRECTION: Independent thought and action-choosing, creating, exploring. (Creativity, Freedom, Independent, Curious, Choosing own Goals) UNIVERSALISM: Understanding, appreciation, tolerance and protection for the welfare of all people and for nature. (Broadminded, Wisdom, Social Justice, Equality, a World at Peace, a World of Beauty, Unity with Nature, Protecting the Environment) BENEVOLENCE: Preservation and enhancement of the welfare of people with whom one is in frequent personal contact. (Helpful, Honest, Forgiving, Loyal, Responsible) TRADITION: Respect, commitment and acceptance of the customs and ideas that traditional culture or religion provide. (Humble, Accepting my Portion in Life, Devout, Respect for Tradition, Moderate) CONFORMITY: Restraint of actions, inclinations, and impulses likely to upset or harm others and violate social expectations or norms. (Obedient, Politeness, Self-Discipline, Honoring Parents and Elders) SECURITY: Safety, harmony and stability of society, of relationships, and of self, (family security, national security, social order, clean, reciprocation of favors)

dertaken in order to fulfill one value may conflict with the pursuit of other values. For example, behaviors engaged in the pursuit of stimulation values, which emphasize excitement, novelty, and challenge, are likely to impede the attainment of security values, which emphasize safety and stability. The total pattern of relations of conflict and compatibility among values yields the circumplex structure represented in Fig. 9.1. Competing values emanate in opposing directions from the center; complementary values are in close proximity going around the circle. Tradition and conformity are located in a single wedge because the motivational goal of both types of values is submission of self to external expectations. Tradition and conformity differ however, in the object of subordination: For conformity values, subordination is directed toward persons with whom one is in current interaction, while for tradition values, subordination is directed toward transcendent authorities and received wisdom (Schwartz & Huismans, 1995). The structure represented in Fig. 9.1 emerged from two-dimensional spatial representations of intercorrelations among the 56 values. Separate analyses were conducted for 97 samples (Schwartz, 1992, 1994; Schwartz & Sagiv, 1995). These analyses provided substantial support for the distinc-

9. VALUES AND EMOTIONS

267

FIG. 9.1. Structure of the ten types of values. From Schwartz (1992). Adapted with permission of Shalom H. Schwartz.

tions among the 10 value types and their circular relationship in which proximity represents positive intercorrelations. Results indicated that 45 of the values had similar meaning across cultures, as indicated by their appearance in the hypothesized region of the circle in at least 75% of the samples. Schwartz (1992) uses these 45 values in standard indices to measure the priority given to the ten value types across cultures. Findings also indicate that there is widespread consensus regarding the hierarchical order of values. Comparing the mean value ratings across more than 50 nations indicates that benevolence, self-direction, and universalism are consistently rated as most important, and power, tradition and stimulation values are consistently rated as least important (Schwartz & Bardi, 2001). Nevertheless there are important individual differences in the importance attributed to each of the 10 types of values, and such differences are systematically related to consumer purchases, cooperation and competition, counselee behavioral style, delinquent behavior, environmental behavior, intergroup social contact, occupational choice, religiosity and reli-

268

ROCCAS AND McCAULEY

gious observance, and voting (Barnea & Schwartz, 1998; Bianchi & Rosova, 1992; Feather, 1995; Grunert & Juhl, 1995; Karp, 1996; Maio & Olson, 1994; Puohiniemi, 1995; Roccas & Schwartz, 1997; Sagiv, 1997; Sagiv & Schwartz, 1995; Sagiv & Schwarz, 2000; Schubot, Eliason, & Cayley, 1995; Schwartz, Sagiv, & Boehnke, 2000). In a social dilemma game, for instance, participants who emphasize benevolence values contribute more of their own money to their group's money-pool than participants who emphasize power values (Sagiv & Schwarz, 2000). Thus Fig. 9.1 and indeed the very concept of value importance refer to relative importance or prioritizing among the values.

THE MUTUAL INFLUENCE OF VALUES AND MODELS

We postulate that influence flows in both directions between values and relational models. Individuals' value priorities, grounded in their personal needs and experiences, including experiences in relational models, lead them to prefer relations based on a specific model (Fiske, 1991, pp. 99-114; Fiske & Haslam, 1998, p. 387). Conversely, cultural emphasis of a specific relational model can lead to emphasizing the value priorities that model expresses and downgrading the values that are inconsistent with that model (Fiske, 1991, p. 240). Individuals Actively Choose Social Relations That Are Compatible With Their Values People are more likely to favor social relations in which they can express and fulfill their value priorities. Each relational model offers a different set of opportunities for value expression, channeling some values but blocking others, and these opportunities will be more or less congruent with an individual's values. Social relations modeled according to CS, for instance, offer opportunities to behave in ways that express helpfulness, but block expression of power motivation. In contrast, individuals who value helpfulness will find scant opportunity to express this value in a market pricing relation, and individuals who value self-gratification cannot effectively express this value in a CS relation. Each relational model can also be viewed as a set of interlocking expectations and prescriptions. When individuals do not conform to the expectations implicit in the relational model in which there are engaged, they may be punished, ignored, or ostracized. Thus when an individual's values conflict with the expectations of a model, the individual may be more likely to violate the model and to suffer from social sanctions and feelings of isola-

9. VALUES AND EMOTIONS

269

tion. For instance, a child who tries to dominate or bully others may soon be isolated or ignored. In sum, we expect links between relational models and personal values because individuals will actively choose social relations that are consistent with their value priorities, and avoid social relations that are incompatible with their values. Relational Models Shape the Value Priorities of Individuals So far we have described a causal path from values to relational models, in which an individual's value priorities affect his or her willingness or ability to adhere to the norms prescribed by a relational model. Equally, however, experience in a model can shape value priorities. Social institutions based on a particular model are likely to inculcate value priorities that express motivations compatible with the behavior prescribed by that model. The relational model that governs an interaction induces compliance with behavioral patterns that are compatible with some values and incompatible with others. Once a person is committed to a habitual set of behaviors, two mechanisms will be entrained that will move her values toward consistency with her behavior. First, she may deduce her values from her behavior, and come to see herself as cherishing values compatible with the relational model most salient in her social context (Bern, 1972). This mechanism is familiar as the self-perception interpretation of dissonance results under circumstances where an individual is initially unsure about what value is relevant to a particular situation (Sabini, 1995, p. 553). The potential power of this mechanism is visible in the foot-in-the-door phenomenon (FTD; Freedman & Fraser, 1966), in which agreeing to a small request makes an individual more likely to agree with a subsequent later request. In an initial demonstration of FTD, agreeing to take and display a 3-inch square sign ("Be a Safe Driver") led to increased agreement to a subsequent larger request (staking a large poorly lettered sign, "Drive Carefully," into the subject's front lawn). Even when the first request involved a different issue (3-inch sign, "Keep California Beautiful"), increased agreement was found for the "Drive Carefully" billboard. Freedman and Fraser interpreted their results in terms of the subject's coming to see herself at the first request as the kind of person who helps charitable causes, that is, as someone who has and acts on a value of benevolence when asked to help a deserving cause. A relational models perspective suggests more explicitly that the subject chooses among relational models in responding to the first request. Agreeing to the first request instantiates CS as the model of relationship with door-to-door solicitors for charitable causes, and this model then controls response to the next solicitation.

270

ROCCAS AND McCAULEY

The second mechanism that moves values toward consistency with behavior is dissonance reduction. An individual may experience a threat to self-esteem as a result of inconsistency between her values and her behavior. This is the original understanding of dissonance as a kind of unpleasant arousal that follows upon the perception that one has behaved in a way that is stupid or sleazy (Sabini, 1995, p. 553). Dissonance reduction may then lead to changing value priorities to make them more compatible with the behaviors required by a specific relational model. For example, membership in a hierarchical organization (such as the military) leads to habitual deference to legitimate authority. With time, social and emotional commitment to promoting obedience will intensify and thereby increase the importance of obedience as a value. If we are correct in seeing mutual influence between values and models, it should be possible to draw out the value implications of each of the relational models. THE RELATIONAL MODELS AND THEIR LINKS TO VALUES Authority Ranking Fiske's (1991) definition of Authority Ranking makes the ranking relationship central: AR is any relation between individuals or groups in which one is higher, better, more powerful than the other. Nearly all of his examples of AR, however, include the mutual obligations that denote authority: the pastoral responsibility due from superior to subordinate and the respect and deference due from subordinate to superior. Ranking relationships do not always denote authority, however. Sports team rankings, for instance, reflect a competitive relationship without mutual obligations between superior and inferior teams, whereas parent-child relations typically include the obligations of care downward and respect upward. It seems to us that these two kinds of ranking relationships—"competitive ranking" and Authority Ranking—may represent importantly different models of relationship. In this chapter, we follow Fiske in focusing on the AR model that includes mutual obligations. Authority Ranking provides different affordances for behaviors and entails compliance with different norms for those in subordinate and superior roles. Most people in a hierarchy are in a subordinate role. Hence we first describe the values most compatible with a subordinate role, and then briefly extend the discussion to the motivations related to superior roles. In relations modeled as AR, subordinates defer, respect, and (perhaps) obey. Deference to the guidance of others and obeying directives is often

9. VALUES AND EMOTIONS

27 I

incompatible with making independent choices, and being guided by one's own ideas. Thus, social relations based on AR contrast with the pursuit of self-direction values, which emphasize independent thought and action, following one's own intellectual and emotional interests. Authority Ranking relations also contrast with attaching importance to stimulation values. The central motivational goal of these values is to experience excitement, novelty, and challenge in life. This goal is incompatible with self-restraint and compliance with the directives of authority figures. Authority-ranking relations are most compatible with the motivational goals of tradition and conformity values. These values share a basic motivational goal—submission of self to external expectations. For conformity values, subordination is directed toward persons with whom one is in current interaction. For tradition values, subordination is directed toward transcendent authorities and cultural ideas. Thus both value types are highly compatible with Authority Ranking relations. Tradition is most compatible when the relation is with a distant abstract authority (e.g., religious authorities; leader of a large social category). Conformity is most compatible with AR in interpersonal relations (e.g., leader of patriarchal family; mafia boss). Authority Ranking is also compatible with the central goal of security values. These values emphasize safety, and stability. Being subservient in an AR relationship helps obtain safety because it is the duty of those in the superior positions to look out for and respond to the problems of those who are in subordinate positions. The responsibilities of authority usually include both the material and social welfare of subordinates: not only food, shelter, and safety from outside threat, but ingroup harmony and safety from ingroup and outgroup conflict. In relations modeled as AR, superiors take precedence and take pastoral responsibility for subordinates. The main difference between the motivations of superiors and subordinates is that those in authority have opportunities to express power values. The central goals of these values are social status, dominance, and control over people and resources. Similarly superiors in an Authority Ranking have the opportunity to express achievement values, to the extent that leadership is won by successful performance rather than by birth or fortune. It is important to note that Schwartz's power values do not include the AR model's aspect of responsibility or pastoral care for subordinates; "responsibility" appears as one of the Benevolence values. Another way of expressing this point is to say that the AR model tempers power values with benevolence values, whereas the Market Pricing (MP) model discussed below encourages expression of Achievement and Power values with no concomitant responsibility. Bill Gates can join John D. Rockefeller in giving big money for public purposes, but neither would accept the idea that his posi-

272

ROCCAS AND McCAULEY

tion requires him to support public purposes—not at least while still building the empire that expresses Achievement and Power values. Communal Sharing In relations modeled according to CS, people treat members of a dyad or group as equivalent and undifferentiated with respect to the social domain in question. Relations of this type are most compatible with benevolence values, which express concern for others with which one is in frequent social contact. These values emphasize honesty, helpfulness and kindness, and CS relations give ample opportunities to express these values. In contrast, CS relations are incompatible with the motivational goal of achievement and power values. The motivational goal of these values is to pursue one's own interest even at the expense of others: Power values emphasize obtaining control over people and resources. Achievement values emphasize attaining personal success through demonstrating competence according to social standards. Both type of values contrast with the selflessness implied in CS relations. It is interesting to note that CS is also the model in which universalism values can be best expressed. The motivational goal in common for universalism values is securing the welfare of all people and the whole natural world. This goal amounts to an extension of benevolence values from the circumscribed world of personal contact to the whole of human kind, and even to the whole natural environment in which humans live. That is, the boundary of uncounted and unlimited obligation, in which all should receive according to their needs, is expanded from the local ingroup to a universal ingroup. In between the small group and the world are large and abstract groups such as nation and ethnic group; patriotism and nationalism are the application of CS to groups that are unknowably large but less than universal. There is something surprising about finding benevolence values and universalism values next to one another on Schwartz's circumplex structure, and to find both kinds of values expressed in the CS model. The surprise is associated with the tension between local, particularistic caring and universalistic caring: Are not local obligations of family or village or clan in competition with universalistic obligations of helping and caring for all humankind? Aren't patriotism and nationalism opposed to internationalism and the global village? This issue deserves more attention than we can give it here, but we suspect that the apparent opposition between local and universalistic caring may depend on the idea that caring more in one direction means caring less in another. If this "conservation principle" is false, if caring more about family and clan and village does not mean caring less about the world, then some of the surprise alluded to earlier would be resolved.

9. VALUES AND EMOTIONS

273

In sum, the distinction between benevolence values and universalism values has led us to suggest that CS can be experienced on a continuum of relationships from small group, to large but limited group, to unlimitedly large group. Future research might usefully inquire whether the CS model is applied or experienced in the same way across this continuum. Market Pricing Market Pricing relationships are oriented to socially meaningful ratios or rates such as prices, wages, interest, rents, tithes, or cost-benefit analyses. It is generally the individuals who value power and achievement who have most opportunities to attain their goals in Market Pricing (MP) relations. The central goal of power and achievement values is pursuit of self-interest even at the expense of others. They emphasize wealth, ambitiousness, and high accomplishment. Differences in wealth and in personal success are most meaningful in social contexts in which prices, market value and other ratio-value constructs are central. Relations based on market pricing are incompatible with expressing concern for others. Individuals are expected to avoid making special allowances for those who are most in need: such as conduct would be labeled as unfair, unprofessional. Thus, relations modeled according to MP are incompatible with the expression of universalism and benevolence values. Indeed the MP model, in its emphasis on ratios, tends to emphasize material exchanges where ratios can be calculated. The MP model is less easily applied to exchanges of such immaterial commodities as loyalty, love, and respect. Imagine trying to determine if my love for you is equal to your love for me, let alone whether my love for you is twice yours for me. It is the materialist emphasis of the MP model, as much as its focus on the individual, that puts it in conflict with universalism and benevolence values. Equality Matching In relations modeled according to Equality Matching (EM), people keep track of the balance or difference among participants and know what would be required to restore balance. Equality Matching relations entail limited obligations and limited benefits. Individuals do not get goods according to their needs and do not contribute according to their capabilities. The obligations pertain only to the specific social situation in which individuals interact. Thus, in equality matching receiving a service from others is not conceptualized as offering or accepting help; rather it is conceptualized as a limited transaction that entails reciprocity but not gratitude. Equality Matching is the relationship model that corresponds to the reciprocity rule studied by anthropologists and social psychologists and the tit-

274

ROCCAS AND McCAULEY

for-tat strategy beloved by economists and evolutionary psychologists (Buss & Kenrick, 1998). Reports from many different cultures around the world are consistent in describing strong norms for returning good for good and evil for evil; indeed the negative half of the reciprocity rule is so familiar as to have its own appellation: revenge. Unlike the other three relational models, which were easily mapped to value types, our analysis of the relations of EM and values is more complex. One is tempted to say that universalism values should be consistent with EM, as equality looms large in the discourse of rights; political rights, in particular, are often reduced to the slogan of "one-man, one-vote." But Fiske (1991, pp. 41-49) in his listing of "Manifestations and Features" of each of the four relational models, makes clear that the key to EM is turn-taking, reciprocity in kind, and balance over occasions. An EM transaction (e.g., car pooling) is independent of all other transactions, even with the same parties (e.g., exchanging Christmas cards). Further complexity arises from recognizing that EM limits the power of the powerful but also limits the help received by those in need. Consider two political policies consistent with the EM model: one-man, one-vote and a flat tax. One-man, one-vote expresses universalism values of equality when the alternative is oligarchy. But a flat tax expresses power and achievement values when the alternative is progressive taxation. Thus part of the complexity of mapping EM is that it has contrasting implications for values depending on the relational model it competes with. When the salient alternative to EM is that outcomes should be determined by status (AR—or perhaps in this case "competitive ranking"), EM expresses values of universalism and benevolence. But when the salient alternative is that relations should be regulated by need (CS), Equality Matching expresses values of power and achievement. These complexities, in our opinion, make impossible any clear linkage between the EM model and Schwarz's values; EM can have strong relations with values, but these relations are very much dependent on the particulars of the situation. Null Relationship Often neglected in discussion of relationship models is the possibility that no relationship is perceived, or, in other words, that the relationship is asocial. The Null model is likely to be my relation with the stranger coming out of the subway as I go in. For most Americans, the Null model is their relationship with distant and unknown groups such as Laplanders, Fijians, and Azerbaijanis. Efforts to enlist U.S. aid for foreign causes must often try to supplant the Null model with one of the other four. In a notable example of this kind of

9. VALUES AND EMOTIONS

275

appeal, President George Bush senior argued that the United States had to use military force against Iraq because Iraq had violated the norms of the community of nations in its invasion of Kuwait and even in its treatment of its own Kurdish citizens. Kuwait's oil reserves might suggest an MP argument about U.S. relationship with Kuwait, but President Bush relied principally on CS arguments: Iraq had violated its CS obligations to Kuwaitis and Kurds, and the United States could not violate its CS obligations to help these peoples (McCauley, 2000). The values most compatible with Null relational model are hedonism, stimulation, and self-direction. Hedonism and stimulation values express a desire for affectively pleasant arousal. Individuals who pursue these values place high importance on seeking pleasure and sensuous gratification for themselves, they wish for excitement and novelty. Self-direction values focus on personal autonomy; individuals who value self-direction wish to express independence in their thoughts and actions. Emphasis on these three types of values is incompatible with any relational model because any relationship entails mutual obligations, and these obligations and their routine behavioral expressions must circumscribe or inhibit the autonomy and selfindulgence pursued by individuals with values emphasizing self-direction, hedonism, and stimulation. Similarly, conformity/tradition and benevolence values (see Table 9.1) are most inconsistent with the Null model, because these values represent everyday obligations to those around us. Universalism values are not necessarily inconsistent with the Null model, because few understand universalism values to require the everyday obligations and self-control that are entailed by tradition and benevolence. Mother Theresa and Albert Schweitzer aside, most people vote for or contribute to universalistic causes only occasionally; most of us find it easier to care about humanity occasionally than to care for the fallible individuals who surround us every day.

THE STRUCTURE OF VALUES: A CONVERGENCE OF MODELS AND AXES

Figure 9.1 shows outside the perimeter of the Schwartz values circle our hypothesized linkages of relational models with values. Each of the three models for which we have hypothesized consistent and inconsistent values—Authority Ranking, Communal Sharing, Market Pricing—is represented twice: once with a "+" and once with a "-". The "+" and "-" refer to consistent and inconsistent values. So for instance "Communal Sharing+" is located next to benevolence and universalism, and "Communal Sharing-" is located next to achievement and power values. Equality Matching is not represented in Fig. 9.1, owing to the complexities cited earlier.

276

ROCCAS AND McCAULEY

Note that the models organize the circular value structure into two axes according to the values and the relational models that individuals prefer. The first axis is CS+ MP- vs. CS- MP+; the second is AR+ Null- vs. AR- Null+. The implication of these axes is that individuals who strongly favor CS relations are likely to oppose MP relations, and individuals who favor MP relations are likely to oppose CS relations. However, CS+ and CS- individuals can be equally comfortable with AR relationships. Similarly, individuals who strongly favor AR relations are likely to oppose Null relations and individuals who favor Null relations are likely to oppose AR relations, but AR+ and ARindividuals can be equally comfortable with CS or MP relations. The two axes defined by the relational models correspond exactly to the two axes suggested by Schwartz (1992). For Schwartz, each pole constitutes a higher order value type that combines two or more of the 10 types. One dimension opposes openness to change (self-direction and stimulation) to conservation (conformity, tradition and security). The other opposes self-transcendence (universalism and benevolence) to self-enhancement (achievement and power). Hedonism is related both to openness to change and to self-enhancement. We consider that the convergence of the relational models with the Schwartz value structure, particularly the match between the relational model axes and Schwartz's value axes, offers a striking confirmation of the findings of these independent programs of theory and research. The convergence is more notable insofar as evidence for Schwartz model comes from survey research in 50 cultures, whereas the evidence and examples cited by Fiske (1991) come initially from ethnographic observations and later from investigations in cognitive psychology.

INDIVIDUAL DIFFERENCES IN VALUES WITHIN RELATIONAL MODELS

Research on personal value priorities and well-being indicates that there are weak direct relations between value priorities and well-being. For example, individuals who give priority to benevolence values are not happier than individuals who give priority to self-direction values. However, congruity between values and environment is substantially and positively related to well-being regardless of the particular values to which people ascribe importance (Sagiv & Schwartz, 2000). That is, the individual giving priority to benevolence values is happier working for a social work agency than for an internet startup company. We follow Sagiv and Schwartz (2000) in suggesting that well-being and life satisfaction will be promoted by participation in a relationship governed by relational models congruent with an individual's personal value

9. VALUES AND EMOTIONS

277

priorities. Participating in this model leads to attaining valued goals, and is thus likely to promote positive well-being. In contrast, an individual is more likely to feel frustrated and unhappy when participating in interaction based on a relational model that hinders the attainment of their valued goals. Thus, an individual who cherishes self-direction values is likely to resent being subordinate in an AR relation, even when the superior tries sincerely to fulfill his responsibility for the subordinate's welfare. Furthermore, when personal values are incompatible with the relational model with which a person is engaged, he is more likely to reject normative expectations and behave in accordance with his own values. The mismatch between personal values and the values compatible with the relational model may then lead to punishment or rejection. For example, an individual who refuses to abide by the advice of his superior in an AR relationship will be deemed disrespectful and rebellious, and may be punished accordingly. Finally, we have suggested that the type of relational model in which an individual primarily operates can shape his or her values. Self-perception and dissonance pressures, as described earlier, can be in conflict with existing value priorities, and this internal value conflict is likely to be detrimental to well being. In sum, individuals with value priorities compatible with the relational model in which they operate are likely to experience positive well-being. Well-being is likely to be undermined when individuals operate in a relational model that hinders the attainment of their value goals.

EMOTIONAL IMPLICATIONS OF MODELS

The account of values as motivational goals offered by Schwartz is quite abstract; it does not offer any connection to the experience of emotion except the general suggestion that people are happy when they are expressing values they care about and unhappy when expression of these values is blocked. Similarly, the relational models are abstract patterns of mutual obligation. Fiske (1991, pp. 130-135) recognized that he has left relatively undeveloped the issues associated with violation of the models, although he speculated that there are specific emotions associated with seeking, experiencing, violating, and losing each model. A full accounting of the emotions associated with violation of relational models would begin by recognizing that each model admits of at least two kinds of violation: violation within the model and violation by applying different models to the same exchange. For example, violation within a model could occur in a marriage where the partners agree on CS as the model of their financial relationship but disagree about whether sharing requires a joint bank account (a culture-specific implementation rule). In contrast, vio-

278

ROCCAS AND McCAULEY

lation by application of different models could occur if one partner wants to extend CS to a decision about changing jobs when the other partner wants to introduce AR to this decision. Violation within a model can occur for each of the four relational models, leading to four kinds of violation to be considered (violation within a Null model is impossible). Violation by application of different models leads to 20 kinds of violation: each model, including the Null model, can be violated by the imposition of four other models. Thus a full accounting of the emotional aspects of model violation must deal with at least 24 kinds of violation. An additional complexity involves the perspective from which violation is perceived. The three major perspectives are observer, victim, and perpetrator. An observer who finds a model violation may have a reaction—specifically an emotional reaction—different from the person or persons directly involved in the model violated. Consider for example a child who addresses her teacher in the same way that she speaks to other children (Rozin, Lowery, Imada, & Haidt, 1999). The teacher is likely to feel anger for this AR violation, the child may feel embarrassment, whereas an observer may be more likely to feel contempt for the violator. In this chapter we focus on the perspective of the observer, not least because recent research on moral judgment appears to us to be almost exclusively drawn from this perspective (Haidt, Roller, & Dias, 1993; Rozin et al., 1999; Shweder, Mahapatra, & Miller, 1987). Thus we do not aim at a full accounting of the 72 (24 x 3) different cases of emotional response to model violation. Rather we aim more modestly at an introduction to issues of model violation by focusing on the work of Rozin et al. (1999), which suggests that there are characteristic emotions associated with observing violation of the three moral codes advanced by Shweder, Much, Mahapatra and Park (1997). MORAL CODES AND MORAL EMOTIONS Content analysis of interviews about norm violations with Hindu Indians led Shweder et al. (1997) to suggest that violations can be understood in terms of three moral codes: autonomy, community, and divinity. Most familiar in Western cultures is the code of autonomy. In this code, obligations emerge from recognition of individual worth and dignity, and moral violation is understood in terms of a discourse of harm, rights, and justice. Indeed many Westerners are reluctant to talk about moral violation in the absence of specifiable harm to an individual, as for instance when a family eats their dog after the animal has been killed in an accident (Haidt, Koller, & Dias, 1993). Less salient but occasionally discernible in the West are two other moral codes: community and divinity. Violation of community is any behavior that

9. VALUES AND EMOTIONS

279

puts self-interest ahead of the welfare of the community, including violations of respect for community authorities. The community is understood as an entity that is not reducible to the individuals in it, an entity with its own identity, history, and status. In this code, moral judgments are made in relation to obligations of duty, hierarchy, and interdependence, and moral violation is understood as harm to the community. Indeed violation of community is already entrained to the extent that the individual even imagines a personal welfare that is separable from the welfare of the community. Violation of divinity is any behavior that undermines an individual's status and responsibility as a bearer of divine essence. The divinity code supports discourse about moral violation in terms of violations of a sacred, natural, and traditional order, especially violations that pollute the divine essence of the individual. It is interesting to note that, of the twelve scenarios Shweder et al. (1997) found scoring high as violations of the divinity code, seven had to do with violation of norms about eating—a direct and specific form of self-pollution. The salience of the divinity code is very high in Hindu India, where almost every activity of the day is traced on a trajectory of advances and regressions on the scale of divinity. The instantiation of these codes is of course culturally dependent. Today many Westerners see divinity violations only in such extremes as incest and cannibalism (but see Haidt, Koller, & Dias, 1993, for reactions to having sex with a dead chicken), whereas violations of divinity for Hindu Indians encompass failures to observe a myriad of everyday rituals of prayer, avoidance, and purification. Similarly Westerners may acknowledge community obligations more narrowly in nuclear families, sports teams, and the military (but see Haidt et al., 1993, for reactions to using the national flag as a bathroom cleaning rag), whereas Hindu Indians recognize a broad array of family and caste obligations that include respect upward and responsibility downwards in status hierarchies. In contrast, autonomy is foregrounded in the West and is the foundation for most informal and formal (court) complaints about moral violation, whereas in Hindu India the scope of autonomy is relatively restricted (although Shweder et al., 1997, note the potential power of an autonomy code based on the presence of the divine essence in every individual). Taking up Shweder's three codes, Rozin et al. asked whether there might be specific emotions linked to the violation of these codes. Rozin et al. gave Japanese and U.S. subjects scenarios of code violation, emotional labels, and pictures of various emotional expressions. Subjects were asked to say which expressions or which emotional labels best expressed the emotional response of someone who saw or heard the violation described in each scenario of violation. Results indicated significant correlations such that harmrights-justice violations (e.g., seeing someone steal a purse from a blind person) were associated with anger, divinity violations (e.g., hearing about a

280

ROCCAS AND McCAULEY

70-year-old male who has sex with a 17-year-old female) were associated with disgust, and community violations (e.g., hearing a 10-year-old child say dirty words to his or her parents) were associated with contempt. These results lead us to hypotheses about the emotions associated with violation of the four relational models. FROM VIOLATING CODES TO VIOLATING MODELS

It is not difficult to link two of Shweder's three codes to relational models. The community code is larger than the CS model; indeed it includes both the CS and AR models. As Fiske (1991, p. 118) notes, many social theorists have described pre-industrial cultures in terms that join CS and AR. The autonomy code, which Shweder et al. (1997) identify with a view of the self as individual preference structure, includes both the EM and MP models. In contrast to the continuing and unlimited obligations of CS and AR models, EM and MP depend on repeated expression of individual preferences in exchanges that do not depend on previous exchanges of the same type. The relation of the divinity code to the relational models is less obvious, but our suggestion is that the divinity code provides a moral scale on which is projected an individual's success in meeting the demands of whatever models the individual undertakes. Failures to meet the obligations of any of the models pollutes an individual and lowers her on the scale, moving her further from the spiritual and transcendent and closer to the natural and mortal. In offering this translation, we acknowledge conflating the distinction that Shweder et al. (1997) make between the divinity code and the Hindu Indian concept of karma. Karma is the idea that actions have inherent effects on the actor, with good actions raising the actor's spiritual status and communication with the divine, and bad actions lowering this status and communication. Thus the malefactor punishes herself, which may make issues of apology and forgiveness moot for Hindu Indians (S. K. Menon, personal communication, February, 2002). Something of the same idea can be seen in the Christian conviction that sin demeans and pollutes the sinner. But if karma is the scale on which is registered the fruits of one's actions in relation to an order that is both natural and divine, if the natural and divine order includes both the autonomy code and the community code and if the focus of the divinity code is that same natural and divine order (Shweder et al., 1997, see especially pp. 147-150), then the divinity code must be the master code that yields the karmic scale of status in relation to obligations of the divine order. Thus we suggest that a violation of the divinity code is anything that pulls an individual down on a scale that has pure spirit and transcendence of the material world at the top of the scale and animal nature and death at the bottom. It follows that violation of the

9. VALUES AND EMOTIONS

28 I

obligations of any the relational models can potentially be understood as divinity violations. With these translations of codes to models, and code violations to model violations, we can hypothesize the emotional implications of violating relational models. Rozin et al. suggest that violation of the autonomy code should lead observers to anger, and violation of the community code should lead observers to contempt. Our identification of autonomy with MP and EM models then leads to prediction that violations of MP and EM models should move observers to anger. Similarly our identification of the community code with CS and AR models leads to the prediction that violations of CS and AR should lead observers to contempt. And if, as we have suggested, the divinity code offers a kind of common metric of violation, then violation of any of the four relational models should lead to disgust among those who understand the self-demeaning effects of moral violation. Conversely, heroic expression of the norms of any of the relational models should lead to "elevation," an emotion newly identified by Haidt (2000) as the human reaction to acts of moral beauty or virtue.

INDIVIDUAL DIFFERENCES IN RESPONSE TO VIOLATION OF A MODEL

So far we have discussed differences in the type of emotion felt as response to violations of the different models. However, different individuals are likely to feel different levels of emotion in response to the same violation. Going back to our example, not everyone is likely to feel contempt to the same degree when observing a boy talking disrespectfully to his teacher. We suggest that linking values to the relational models offers a way of understanding these individual differences: Individuals will have a stronger emotional reaction, the more compatible are their values with those of the relational model violated. Thus, for example, individuals who value self-direction over conformity/authority will feel less contempt for the disrespectful boy than individuals who value conformity/authority over self-direction. Individuals giving priority to benevolence values will have stronger emotional reaction to a violation of CS than individuals who give priority to hedonism values.

CONCLUSION

We began with a relational-models interpretation of Asch's conformity paradigm. Where the usual interpretation points to a conflict between the values of independence and conformity, we suggested a conflict between CS

282

ROCCAS AND McCAULEY

with fellow students and AR with the experimenter. An important advantage of the relational-models interpretation is that it cuts through the complexity of the choice subjects face if they experience the paradigm as a conflict of values. Which of the many values identified by Schwartz might be implicated in this choice? Self-direction vs. security? Achievement vs. conformity? Stimulation vs. conformity? Power vs. universalism? Like most behavioral choices, the subject's choice in the Asch experiment is susceptible to many value interpretations. The importance of the relational models is that they reduce the complexity of value conflict to something more compact and comprehensible: a choice between AR and CS. Thus the relational models offer a structural solution to the complexities of moral behavior. The difficulty of value-laden choices, of moral behavior in general, is that it is not obvious which values apply or how strongly. A salesman calls looking for your brother who uses your address while working abroad. Is this a test of universalist honesty ("he doesn't live here any more") or a test of benevolence ("he's not in")? How much easier and more concrete it is to choose between the (Null?) relation with the stranger salesman and your (CS or EM) relation with your brother. Or your parents die and you are the executor. Is this a test of hedonism, power, security, tradition, benevolence, or universalism? The problem is at least reduced to comprehensible size by framing it as a competition between CS (give more to your siblings who have less) and EM (everyone gets the same share). Do you even have the right to consider who needs more (an assumption of AR to the executor)? Thus consideration of values and models together leads to the recognition that the models are important because they parse situations into values. To the extent that individuals regulate their behavior according to their important values, they must be able to determine which values are relevant to which situations. But not every value need be considered in relation to every situation; the relevant values are the values embedded in the relational model that controls a situation. The models reduce the buzzing confusion of values to relatively simple choices between relational models. Consideration of the hedonic implications of the relational models led us to several predictions that may be tested in future research with the relational models. Following Fiske (1991), we hypothesize that violations of different models may be associated with different emotions, and we have extended the results of Rozin et al. (1999) to suggest that violation of CS and AR may lead observers to feel contempt for the violator whereas violation of EM and MP may lead observers to anger. An extension of the divinity code described by Shweder et al. (1997) led us to suggest further that violation of any of the four models will lead to disgust for observers who understand the self-polluting effects of violating the natural and divine order.

9. VALUES AND EMOTIONS

283

Finally, we were led to some predictions about individual differences within the relational models. We suggested that value priorities may modulate reactions to violations of a model, such that individuals with value priorities most consistent with a model may be least likely to violate the obligations of the model and have the strongest emotional reaction to violations observed or experienced. A related implication concerns the mental health and well-being of individuals participating in a relationship governed by a relational model: individuals with values most consistent with the model should be happiest. Most generally, we have sought to tie the relational models more explicitly to the psychology of motivation and emotion. Relational models are defined as cognitive schemas grounded in the distinctions of measurement theory, but their power to structure human interactions depends on their expression in values, motives, and emotions. Each model is a system of reciprocating norms, and it is the prescriptive quality of these norms that warrants the expectation and (usually) the performance of the normative behaviors. Without the social rewards associated with meeting normative expectations, and without the sanctions associated with failure to meet these expectations, the relational models would leave us paralyzed in cold cognition. This logic led us to explore some of the hedonic implications of the relational models. The particular predictions we have suggested may or may not be supported in future research, but we are confident in the general contention that the relational models will be most useful when their hedonic implications are as well articulated as their cognitive foundations. REFERENCES Allport, G. W. (1961). Pattern and growth in personality. New York: Holt, Rinehart & Winston. Asch, S. (1956). Studies of independence and conformity: A minority of one against a unanimous majority. Psychological Monographs, 70 (Whole No. 416). Barnea, M, & Schwartz, S. H. (1998). Values and voting. Political Psychology, 19, 17-40. Bern, D. J. (1972). Self-perception theory. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 6, pp. 1-62). New York: Academic Press. Bianchi, G., & Rosova, V. (1992). Environment as a value: Intraindividual, interindividual and intercultural differences. In H. Svodoba (Ed.), Culture, nature, landscape (pp. 37-45). Zdar n/ Sazavou: International Association of Landscape Ecology. Buss, D. M., & Kenrick, D. T. (1998). Evolutionary social psychology. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (4th ed., pp. 982-1026). New York: McGraw-Hill. Darley, J. M., & Batson, C. D. (1973). From Jerusalem to Jericho: A study of situational and dispositional variables in helping behavior. Journal of Personality and Social Psychology, 27, 100-108. Feather, N. T. (1995). Values, valences, and choice: The influence of values on the perceived attractiveness and choice of alternatives. Journal of Personality and Social Psychology, 68, 1135-1151.

284

ROCCAS AND McCAULEY

Fiske, A. P. (1991). Structures of social life. New York: Free Press. Fiske, A. P., & Haslam, N. (1998). Prerequisites for satisfactory relationships. In L. H. Meyer, H. Park, M. Grenot-Scheyer, I. S. Schwartz, & B. Harry (Eds.), Making friends: The influences of culture and development (pp. 385-392). Baltimore: Paul H. Brookes. Freedman, J. L., & Fraser, S. C. (1966). Compliance without pressure: The foot-in-the-door technique. Journal of Personality and Social Psychology, 4, 195-202. Grunert, S. C., & Juhl, H. J. (1995). Values, environmental attitudes, and buying organic foods. Journal of Economic Psychology, 16, 39-62. Haidt, J. (2000). The positive emotion of elevation. Prevention and Treatment, 3, np. Haidt, J., Koller, S. H., & Dias, M. G. (1993). Affect, culture, and morality, or is it wrong to eat your dog? Journal of Personality and Social Psychology, 65, 613-628. Hofstede, G. (1991). Cultures and organizations: Software of the mind. London: McGraw-Hill. Inglehart, R., & Beker, W. E. (2000). Modernization, cultural change, and the persistence of traditional values. American Sociological Review, 65, 19-51. Karp, D. G. (1996). Values and their effect on pro-environmental behavior. Environment and Behavior, 28, 111-133. Kluckhohn, C. (1951). Value and value orientations in the theory of action. In T. Parsons & E. Shils (Eds.), Toward a general theory of action (pp. 388-434). Cambridge, MA: Harvard University Press. Maio, G. R., & Olson, J. M. (1994). Relations between values, attitudes, and behavioral intentions: The moderating role of attitude function. Journal of Experimental Social Psychology, 31, 266-285. McCauley, C. (2000). How President Bush moved the U.S. into the Gulf War: Three theories of group conflict and the construction of moral violation. Journal for the Study of Peace and Conflict, 2000-2001 Annual Edition, 32-42. Milgram, S. (1965). Some conditions of obedience and disobedience to authority. Human Relations, 18, 57-76. Puohiniemi, M. (1995). Values, consumer attitudes and behaviour: An application of Schwartz's value theory to the analysis of consumer behaviour and attitudes in two national samples. Unpublished doctoral dissertation, The University of Helsinki. Roccas, S., & Schwartz, S. H. (1997). Church-state relations and the association of religiosity with values. Cross-Cultural Research, 31, 356-375. Rokeach, M. (1973). The nature of human values. New York: Free Press. Rozin, P., Lowery, L., Imada, S., & Haidt, J. (1999). The CAD triad hypothesis: A mapping between three moral emotions (contempt, anger, disgust) and three moral codes (community, autonomy, divinity). Journal of Personality and Social Psychology, 76, 574-586. Sabini, J. (1995). Social psychology (2nd ed.). New York: Norton. Sagiv, L. (1997). Process and outcomes of vocational counseling: The role of clients' values. Unpublished doctoral dissertation, The Hebrew University. Sagiv, L., & Schwartz, S. H. (1995). Value priorities and readiness for out-group social contact. Journal of Personality and Social Psychology, 69, 437-448. Sagiv, L., & Schwartz, S. H. (2000). Values, priorities and subjective well-being: Direct relations and congruity effects. European Journal of Social Psychology, 30, 177-198. Sagiv, L., & Schwarz. N. (2000). Do values influence behavior?Individual differences and temporary accessibility. Manuscript submitted for publication. Schubot, D. B., Eliason, B. C., & Cayley, W. (1995). Personal values and primary care specialty aspirations. Academic Medicine, 70, 952-953. Schwartz, S. H. (1992). Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries. In M. Zanna (Ed.), Advances in experimental social psychology, Vol. 25 (pp. 1-65). New York: Academic Press. Schwartz, S. H. (1994). Are there universal aspects in the structure and contents of human values? Journal of Social Issues, 50, 19-45.

9. VALUES AND EMOTIONS

285

Schwartz, S. H., & Bardi, A. (2001). Value hierarchies across cultures: Taking a similarities perspective. Journal of Cross Cultural Psychology, 32, 268-290. Schwartz, S. H., & Huismans, S. (1995). Value priorities and religiosity in four Western Religions. Social Psychology Quarterly, 58, 88-107. Schwartz, S. H., & Sagiv, L. (1995). Identifying culture specifics in the content and structure of values. Journal of Cross-Cultural Psychology, 26, 92-116. Schwartz, S. H., Sagiv, L., & Boehnke, K. (2000). Worries and Values. Journal of Personality, 68, 309-346. Seligman, C., Olson, J. M., &Zanna, M. P. (Eds.). (19%). The psychology of values: The Ontario symposium. Vol. 8. Mahwah, NJ: Lawrence Erlbaum Associates. Shweder, R. A., Mahapatra, M., & Miller, J. (1987). Culture and moral development. In J. Kagan & S. Lamb (Eds.), The emergence of morality in young children (pp. 1-83). Chicago: University of Chicago Press. Shweder, R. A., Much, N. C., Mahapatra, M., & Park, L. (1997). The "big three" of morality (autonomy, community, divinity) and the "big three" explanations of suffering. In A. Brandt & P. Rozin (Eds.), Morality and health (pp. 119-169). New York: Routledge. Smith, P. B., & Schwartz, S. H. (1997). Values. In J. W. Berry, M. H. Segall, & C. Kagitcibasi (Eds.), Handbook of cross-cultural psychology, Vol. 3 (2nd ed., pp. 77-118). Boston: Allyn & Bacon.

This page intentionally left blank

C H A P T E R

10 The Four Faces of Trust: An Empirical Study of the Nature of Trust in Relational Forms Leah D. Houde Dana M. Sherman Tiffany B. White Blair H. Sheppard

It's Friday night, and after a long week at work followed by hectic evenings with your child, you and your spouse decide you'd like an evening alone together and hire a baby sitter to watch your child. You have placed a great deal of trust in the sitter to take care of your child despite such potential risks as the sitter's neglecting your child's needs, allowing your child to do something dangerous, or worse yet, abusing him or her. As Deutsch (1958) suggested when he first presented this example, trust is evident only in situations where the potential damage from unfulfilled trust is greater than the potential gains if trust is fulfilled. With this assertion, he has captured two themes that have pervaded research and thinking regarding trust ever since: Trust entails the assumption of risks, and some form of trust is inherent in all relationships. Since Deutsch's beginnings, scholars have written much about trust and trust production. Many view trust with skepticism, claiming it as an irrational behavior. This viewpoint is at the center of much writing in agency theory (e.g., Eisenhardt, 1989; Fama & Jensen, 1983; Gomez-Mejia & Balkin, 1992; Jensen & Meckling, 1976), transaction cost economics (e.g., Bromiley & Cummings, 1995; Ring & Van de Yen, 1989; Williamson, 1981,1985), and game theory (Deutsch & Krauss, 1960; Kramer & Brewer, 1984, 1986; Luce & Raiffa, 1957; Murnighan, 1991). Others, however, view trust as a natural part of relationships (e.g., Gabarro, 1978; Rempel, Holmes, & Zanna, 1985). Sheppard and Sherman (1998) present a model asserting that risk is at the center of how people do and should think about trust but that risk var-

287

288

HOUDE ET AL.

ies distinctly as the form of the relationship varies. Following from this notion, these authors suggest that trust is therefore not a singular construct, but that it takes quite distinct forms. Using Fiske's (1990) relational forms as a basis for how relationships vary, and therefore how the nature of risk and trust varies within them, Sheppard and Sherman suggest that as the risks associated with the various relational forms are understood and managed they can be mitigated. Although Sheppard and Sherman use their typology to integrate a vast array of literature on trust and trust production, the original article used theory as the basis for the construction and defense of their model. Our goal here is to test this theoretical model. In the remainder of this chapter we describe in detail the model presented by Sheppard and Sherman. We then present a series of empirical studies we conducted to determine whether people do in fact behave in the manner suggested by the theory. Finally, we present the implications of this work for both practice and future research.

RELATIONAL FORM, DEPTH, AND RISK

After reading the other chapters in this book, you should all be familiar with Fiske's (1990) four models of human relationships: Communal Sharing (CS), Authority Ranking (AR), Equality Matching (EM), and Market Pricing (MP). Fiske's assertions about the pervasiveness and importance of these models, an essential element to the argument proposed here, suggests that the models are foundational. Specifically, Fiske (1990) states: My hypothesis is that these models are fundamental, in the sense that they are the lowest or most basic-level "grammars" for social relations. Further, the models are general, giving order to most forms of social interaction, thought, and affect. They are elementary, in the sense that they are the basic constituents for all higher order social forms. It is also my hypothesis that they are universal, being the basis for social relations among all people in all cultures and the essential foundation for cross-cultural understanding and intellectual engagement, (p. 25)

An implication of the foundational nature of these relational grammars is that they are well understood and well practiced. Although Fiske focuses on how relational partners understand the nature of exchange in the different relational models, he does not focus on the risks associated with each model or the mechanisms of trust production that might serve to mitigate these risks. Sheppard and Sherman (1998) propose a conceptualization of trust that flows from Fiske's models, suggesting that the nature of risk in each of the relational models is fundamentally different, and therefore the nature of trust in each of these models will also be distinctly different. Spe-

10. TRUST IN RELATIONAL FORMS

289

cifically, they examine the nature and depth of the interdependence inherent in the four models and then consider the risks in a relationship arising as a result of this interdependence. The two relational characteristics of interest, and the basis of Sheppard and Sherman's Grammars of Trust, are form and depth. Relational form, which can be conceptualized as either dependent or interdependent, refers to the type of interdependence in a given relationship. It is important to note that rather than the notion of interdependence as used by organizational theorists (Thompson, 1967), we and other social psychologists interested in trust (Deutsch, 1958; Kelley & Thibaut, 1978) believe it is most meaningful to approach interdependence from the vantage point of the person engaged in trusting behavior rather than a third-party observer or someone trying to manage the interactions of two or more parties. In this sense, dependence occurs when one's outcomes are contingent upon the actions of another. Although behavior of the engaged parties appears interdependent from the perspective of one trying to manage the relationship, it seems to be unidirectional dependency from the perspective of the person engaging in the trusting behavior. Relational depth, conceptualized as either shallow or deep, refers to a structural feature of a relationship that is a product of the importance, range, and number of points of contact among parties in the relationship (for more detailed discussions of relational depth, see Berscheid, 1994; Burgess & Huston, 1979; Levinger, 1980; Miller & Boster, 1988; Tuchinsky, 1997). Combining these two characteristics gives us four relationship types each of which maps quite closely to one of Fiske's models. As risk is at the heart of our argument, we briefly review the risks associated with being in each of the four types of relationships. Shallow dependence, which parallels Fiske's MP model, entails two main risks. The risk of unreliability is a concern that a relational partner will not behave as expected; in this case, a bargain or understanding is not upheld, and buyers risk not getting what they bargained for. The second risk, indiscretion, occurs when the trustor assumes that sensitive information will not be shared. Here, if Jim asks for help from Sally, thus creating a dependency on her, Jim would need to share information about himself for the task to be completed, and therefore has put himself at risk for Sally to disclose this information at his expense. Although the risks of unreliability and indiscretion are quite different, both arise from one party transferring responsibility to another, thus creating a dependency on the other. We equate shallow dependence with Market Pricing because it shares several features. First one actor is dependent upon another for the provision of a service or good without any expectation for reciprocation in kind. Second, this occurs in the context of a relatively narrow exchange typical of market transactions. Fiske adds the feature not essential to shallow de-

290

HOUDE ET AL.

pendence that there is an exchange of goods for money. There are probably instances of shallow dependence in which people provide a service or good without expectation of compensation such as in charitable circumstances. A community kitchen is providing a necessary service without expectation of compensation. In such an instance, the motivation is more communal in form and based on a concern for need. However, while this is true, the primary relationship in which one party provides a service or good upon which the other is dependent without greater authority is in market pricing exchanges. Shallow interdependence, which parallels Fiske's EM model, still entails the risks of shallow dependence, but as both parties must now coordinate their behavior in order to achieve desired results, this also creates the risk of poor coordination: that coordination will not occur to sufficient standards or with sufficient speed to be successful. As relationships become more intimate, and therefore typically more interdependent, one can see an increasing need to coordinate schedules, household chores, and vacations. Even if both parties try to maintain coordination, a missed step results in the consequence of this risk: too little too late. Shallow interdependence is most like EM as shallow interdependence involves a relationship between two somewhat independent parties who engage in reciprocal behavior. While Equality Matching behavior can be among parties with pretty deep interdependence, we would argue if the interdependence gets deep enough the parties come to develop a true sense of shared identity and shared fate typical of Communal Sharing. If two parties' successes are deeply, deeply intertwined they come to think of themselves as members of a community of common interest. Deep dependence closely resembles Fiske's AR model, and as such one can see the difficulty on the part of the trustor to monitor the trustee's behavior. Deep dependence in circumstances such as a boss determining one's raise, tasks or career opportunity, or one's teacher deciding grade or class placement often involves behavior outside the trustor's view. In such instances, the trustor is at risk of being cheated by the trustee. The risk of cheating is based on asymmetries in knowledge, which occur as the relationship broadens in range and points of contact, and when the party in control uses this additional knowledge to the dependent party's disadvantage. Rather than intentional cheating, behavioral invisibility more often entails risks associated with neglect or omission. For example, shareholders and other key stakeholders in organizations regularly risk that the managers of that organization will not consider their interests when making important decisions privately. Another risk associated with deep dependence is what Kelley and Thibaut (1978) termed fate control, and arises when one party can unilaterally determine the fate of another, as when a boss determines an employee's salary or a father sends his son to military school.

10. TRUST IN RELATIONAL FORMS

291

This fate control leads to the risk of abuse that may occur when the trustee overutilizes the control lever. Finally, as one's sense of self often directly follows from one's deep relationships, the risk to self-esteem is present and can manifest itself either through direct interaction with relational partners or through perceived lack of success in the relationship. The connection between deep dependence and AR should be somewhat self-evident. In the case of formally acknowledged authority roles such as teacher, boss, or disciplining parent, the dependency upon the actions and wishes of the other is clear. However, the argument works in the other direction as well; any deep dependence relationship entails an asymmetry of outcome control giving one party control and thus implied authority over the other. In deep interdependence, which parallel's Fiske's CS model, communication between the parties is essential. Not only do they need to coordinate their activities, but because the range of contacts is broad and important, this coordination becomes not only more challenging but also more crucial. Sometimes adequate or complete communication is not possible, however, such as when parties to the relationship are separated by a great distance or when decisions need to be made quickly. As a result, the central risk of deep interdependence is misanticipation, or the risk that, without specific instructions, one will not be able to anticipate what the other needs. For example, because Carrie only had enough time to leave her husband Scott a quick message that she needed to work late, he may not have realized that this meant he needed to pick their son up from school and cancel dinner with the neighbors. However, because they have been married for 5 years, Carrie anticipates that Scott will have enough foresight to cover these unexpected consequences. This leads to the second tenet of Sheppard and Sherman's theory, that proper selection of a trustworthy partner can mitigate the risks associated with the depth and interdependence of the relationship.

TRUSTWORTHINESS

In the deep interdependent situation described earlier, Carrie requires that Scott would foresee or guess that she would like him to pick up their child and cancel the dinner plans given her current situation (i.e., needing to work late). This ability to project what one's relational partner expects requires intuition and empathy, or what psychological researchers call "perspective taking" (Bern, 1993). However, as the previous discussion suggests, trustworthiness means different things depending on the nature of the risks being assumed in the relationship. This level of perspective taking is unnecessary in a simple, shallow dependence relationship, for example. Here, it is only necessary to

292

HOUDE ET AL. TABLE 10.1 Form of Dependence, Risks, and Qualities of Trustworthiness

Form of Dependence

Risks

Shallow Dependence

Indiscretion Unreliability

Deep dependence

Shallow interdependence

Cheating Abuse Neglect Self-esteem Poor coordination

Deep interdependence

Misanticipation

Qualities of Trustworthiness Discretion Reliability Competence Integrity Concern Benevolence Predictability Consistency Foresight Intuition Empathy

From Sheppard and Sherman (1998)

select partners for their history of discreet, competent, and therefore reliable behavior. In deep dependent relationships, a partner's trustworthiness must mitigate the risks of cheating, abuse, neglect, and harm to self-esteem. As the literature suggests, honesty (Butler, 1991; Larzelere & Huston, 1980) and integrity (Butler, 1991; Lierberman, 1981; McFall, 1987) are the most obvious mitigating factors as people with these characteristics are highly unlikely to take advantage of opportunities to cheat others. In addition, benevolence (Larzelere & Huston, 1980; Solomon, 1960; Strickland, 1958) and caring (Mishra, 1996) help mitigate the potential for neglect, abuse, or harm to self-esteem. Finally, while shallow interdependence also requires that one's partners be reliable and discreet (as in shallow dependent relationships), it also requires characteristics to mitigate the risk of poor coordination. In order to properly coordinate behavior, this behavior must be predictable; this suggests consistency (Butler, 1991), transparency (Good, 1988; Hart, Capps, Cangemi, & Caillouet, 1986), and predictability (Dasgupta, 1988; Gabarro, 1978; Giffin, 1967) as key attributes of trustworthiness in shallow interdependence.

EVIDENCE FOR THE GRAMMARS OF TRUST Sheppard and Sherman's (1998) original article used a construal of Fiske to account for an array of findings in research on the determinants of trust and perceived trustworthiness. As such it is only a theoretical argument. Thus, a

10. TRUST IN RELATIONAL FORMS

293

number of questions arise: What do we know about individuals' cognitive models of relationships? Do people think about various types of relationships in ways that the grammars suggest? Is this a useful framework for individuals to think about risks in relationships and the types of trust needed to overcome these risks? This chapter summarizes the findings of the initial studies in a series of studies directed at answering these questions. The first question concerned whether or not people construe relationships in a manner consistent with the dimensions of Sheppard and Sherman's model. While others have examined the structure of people's models of relationships and the capacity of Fiske's relational models theory to predict social cognition, (Fiske, 1993, 1995; Fiske & Haslam, 1996, 1997b; Fiske, Haslam, & Fiske, 1991; Haslam, 1994a, 1994b; Haslam & Fiske, 1992, 1999), we were seeking to determine whether individual's cognitive models reflected the dimensions suggested in the grammars of trust. We first decided to answer this question within the context of naturally occurring relationships that serve as prototypes for each relationship type. Brainstorming led to a set of 16 prototypes, four for each relationship type with two work-related relationships and two personal relationships. At this point, we conducted a study simply to examine how people viewed these 16 relationships in terms of their depth and interdependence. Participants consisted of 76 undergraduate students who received course credit for their participation. Each participant received a set of four relationships to consider, all either work-related or personal and one of each form. For relationships that were less common or straightforward, information was given to the participant to set up the context, however, no detailed information was given to describe the characteristics of the relationship itself. An example of a relationship description with a great deal of contextual description is, "Imagine that you are in a foreign country and do not speak the language. You are lost and need to ask directions to a local cafe where you are meeting someone. You see a uniformed officer and decide to ask him for directions. Think about your relationship with the officer." Examples of relationship descriptions without contextual description include, "Think about your relationship with your mother," or "Think about your relationship with your spouse." The participant was then asked to rate the relationship on a sevenpoint Likert-type scale in terms of its longevity and importance. Further, the participant indicated on seven-point Likert-type scales his or her agreement with seven statements about the relationship. Statements alluding to depth included, "In my relationship with the officer, he would consult with me on several issues of importance," and "In this relationship, I would have rich, interesting discussions with the officer." Statements alluding to dependence included, "In my relationship with the officer, I would depend on him more than he depends on me," and "In my relationship with the officer, he would have more power over me than I would have over him." Finally, in an

294

HOUDE ET AL.

attempt to distinguish dependence from interdependence, an additional statement was included: "In this relationship, I would consider the officer and me to be mutually dependent." Relationships were considered dependent if participants both agreed with statements indicating dependence and disagreed with the statement suggesting mutual dependence, while relationships were considered interdependent if participants both disagreed with statements indicating dependence and agreed with the statement suggesting mutual dependence. Relationships that scored high or low on both factors were not considered prototypical of any relational form. From the data it was clear that these evaluations were relatively simple for subjects to make and consistent with expectations. Eight relationships, in particular, fit the expected pattern as follows: shallow dependence—an officer one is asking for directions in a foreign country and the store owner where one is buying balloons, deep dependence—one's mother and the commanding officer in a military relationship, shallow interdependence—one's teammates in a pick-up game of basketball and an HR director, and deep interdependence—one's spouse and a book co-author (see Table 10.2 for mean depth and interdependence ratings). Although our main purpose was to identify naturally occurring relationships for the study of the risk-form relationship, it is encouraging that this task was so natural for subjects and that they invariably considered relationships as we expected. There were some interesting patterns in the ways in which perceptions of each dimension vary depending on the relational form. For example, as depth increases, so do perceptions of interdependence. In other words, in addition to interdependent forms being seen as more interdependent than dependent relationships, deep dependent relationships are viewed as more interdependent than shallow dependent relationships, and deep interdependent relationships are viewed as more interdependent than shallow interdependent relationships. The next study was designed to confirm our findings from Study 1 as well as examine if people perceive the risks associated with each type of relationship as suggested by the grammars. Participant again consisted of 76 undergraduate students participating to receive course credit. Each participant in this study received two relationship descriptions, both the workrelated and the personal prototype from one of the four forms. After reading the description, participants were asked to comment, on a seven-point Likert-type scale, as to how possible it would be that various risks could happen in the relationship. Two questions each were asked to ascertain the following risks: the risk of indiscriminate behavior on the part of the relational partner (a = 0.63), the risk of unreliability (a = 0.64), the risk of cheating by the relational partner (a = 0.66), the risk of neglect by the relational partner (a = 0.92), the risk of miscoordination (a = 0.59), and the risk of misanticipation (a = 0.83). In addition, participants were asked the same

10. TRUST IN RELATIONAL FORMS

295

TABLE 10.2 Mean Depth, Mutual Dependence, and Dependence Ratings for Each Prototypical Relationship

Prototype Officer in a foreign country M SD Store Owner M SD Mother M SD Commanding Officer M SD Pick-up game teammates M SD HR director M SD Spouse M SD Co-author M SD

Depth

Mutual Dependence

Dependence

2.39 1.14

1.16 0.39

5.25 1.76

2.39 1.54

3.24 2.25

3.94 1.85

6.85 0.40

3.13 1.93

5.38 0.81

5.67 1.78

3.33 2.10

3.92 2.19

3.55 1.71

6.07 1.16

1.87 1.46

4.33 1.53

6.33 0.98

1.58 1.73

6.78 0.33

6.29 1.10

2.18 1.81

6.43 0.78

6.81 0.43

1.69 1.40

questions as in Study 1, five to determine the depth of the relationship (a = 0.93), two to determine the asymmetry of the dependence (a = 0.77), and three to determine the interdependence of the relationship (a = 0.81). In terms of the manipulation checks, it once again seems that our prototypes differentiated the relationships among our variables of interest, namely depth and form. Specifically, the deep prototypes were seen as deeper than the shallow prototypes (f(l, 282) = 747.41, p < .0001), the dependent prototypes were seen as more dependent than the interdependent prototypes (Ffl, 282) = 411.80, p < .0001), and the interdependent prototypes were seen as more interdependent than the dependent prototypes (f^l, 282) = 282.56, p < .0001). However, we also had some interesting interaction effects once again. In this sample, not only did the perception of depth enhance the perception of interdependence (F(l, 282) = 12.68, p < .001), but perceptions of interdependence also enhanced perceptions of depth (P(l, 282) = 6.18, p < .01). Therefore, although these two factors certainly seem to

296

HOUDE ET AL.

underlie people's perceptions of relationships, they are not orthogonal but rather influence one another. With this in mind, we moved on to examine the nature of the risks perceived to be associated with each type of relationship. Unreliability, which according to the theory should be a risk in all types of relationships, was perceived similarly in all of the relationship types except for shallow interdependence (MSI = 4.1, MOTHERS = 3.4) where it was perceived as significantly riskier. Perhaps reliability is especially necessary when coordination is crucial in the relationship as it is in shallow interdependence. In terms of the other risks, however, deep dependence and shallow interdependence, while structurally different, are perceived similarly. For example, deep dependence and shallow interdependent prototypes were rated highest in terms of the risk of indiscretion (MDD = 3.7, MSI = 3.8). Shallow dependent prototypes were rated the lowest (MSD = 2.7) as suggested by the theory, and deep interdependence was seen as risky, but not as risky as deep dependence or shallow interdependence (MD1 = 3.5). Although we would expect deep interdependence to be the riskiest form, these prototypes are consistently rated as less risky than shallow interdependence and deep dependence prototypes. This is true for the risks of cheating and neglect, as well as miscoordination and misanticipation (where deep interdependence is seen as the least risky). This last result is particularly perplexing. Whereas deep interdependence should be the only relationship type where misanticipation should be a potential risk, it was perceived as significantly less of a threat here than in the other three relationship types (MSD = 4.5, MDD = 4.5, MSI = 4.7, MDI = 3.8). How do we explain these findings? Overall, it seems that across the various naturally occurring relationships, risk is perceived as an inverted U, with both shallow dependence and deep interdependence viewed as having little potential risk, and shallow interdependence and deep dependence viewed as similarly risky. A possible explanation is that shallow dependent relationships are perceived to have low potential for risk because although the probability of a problem may be greater, the cost of the problem is lower, whereas deep interdependent relationships have low perceived risk due to features a deep-interdependent relational partner would have. In other words, perhaps the risks of deep interdependence do not even occur to most people because they carefully select the people with whom they have deep interdependent relationships. One would not choose a spouse or coauthor for whom there is great risk. The cases of deep dependence used in this study did not permit such choice: One's boss is often not a voluntary relationship and one's parents certainly are not chosen. In shallow interdependent relationships, the selection of a relational partner may not be as carefully conducted (as this will be a short-term relationship). Both of these cases could thus lead to higher perceptions of risk. Of course, all these ex-

10. TRUST IN RELATIONAL FORMS

297

planations are speculation. In each instance, the problem is that the explanation for the results may have as much to do with the particular set of relationships chosen and their natural context as it has anything to do with the structure of the relationship. All we know for sure is that people can differentiate among depth and interdependence of relationships quite easily and that there is this quite intriguing possibility that deep interdependence does not behave in a manner similar to the rest of the relationship types. To confirm this result, one more experimental method is required. The next study was designed to manipulate relationship type to control for some of the limitations of the previous study. To permit a direct test of risk perceptions we decided to look at shifts in the type of relationship and the effect of these shifts on trust considerations rather than compare across relationship types as in the previous study. Thus, half the subjects first read of a relationship that began as shallow dependent and then grew in depth and interdependence. The other half read first of a deep, interdependent relationship that declined in depth and interdependence over time. The first set of subjects was asked the new characteristic they would most wish to have in their partner with the change in form. The second group was asked the characteristic they would be most willing to give up with the change in form. Our goal was to provide parallel tests of the risk considerations of subjects about to enter a relationship of a particular type. Participants in this study were given one of two versions of the study, each of which contained four relationship descriptions for them to consider. After reading each relationship description, participants were asked to answer questions, which varied depending on the experimental manipulation, regarding potential features of their relational partner. In the first version, we initially gave participants a description of a shallow dependent relationship. Specifically, the shallow dependent relationship stated, You own a small software company that sells very specialized software applications to large multi-national organizations. The department you deal with in one of these organizations is responsible for integrating your software into their product development activities and managing supplier relations for all technically oriented suppliers. Members of that department are recognized as reliable and discreet. They do what they say they are going to do, they pay their bills on time and they do not share private information you disclose.

We assumed that reliability and discretion would be necessary for any relationship, and therefore built it into all the relationship descriptions. Once they read this description, participants were asked to rank sets of features describing the relational partner as to which they believed was most important to the relationship. The sets of features included a set describing trustworthiness characteristics designed to mitigate the risks of deep dependence: "Thinks about my values and concerns when making decisions that

298

HOUDE ET AL.

affect me. Takes into consideration my values and interests. Uses power only when appropriate. Stays within bounds of authority. Doesn't use information I don't know to harm me," a set to mitigate the risks of shallow interdependence: "Says what he means. Easy to interpret. Doesn't have anything to hide. Only changes behavior when circumstances require. Will make similar decisions given similar situations," and a set to mitigate the risks of deep interdependence: "Understands me well. Socially perceptive. Knows how I'll respond without needing to ask. Can anticipate consequences of behavior. Can see how current actions will affect the future. Recognizes connections among events, that is, how one might affect another." Furthermore, participants were asked to rate each of seven features (pieces of the sets above) as to how important they are to the relationship. Participants then read a second relationship description; this scenario repeated the shallow dependent description and then added a section describing a change in the relationship from a shallow dependent form to a deep dependent form (i.e., the software company is acquired by the large organization). The participant was then given the same features and feature sets and was asked to rank and rate the features as to which would be most important given how the relationship had changed. Finally, participants were given two more relationship descriptions, one shifting the shallow dependent relationship to a shallow interdependent one (i.e., the two firms have been asked to run a conference together), and one shifting it to a deep interdependent one (i.e., the two firms have been acquired by a third, and the two have been asked to work together to design and then jointly run a new software development division). In the other manipulation, the first relationship description described a deep interdependent relationship. You have been co-leader of the senior management team of a software development division of a very large and dynamic multi-national organization. You and the other chair have to work with each other to manage the development and implementation of all strategic issues confronting the division. Your successful integration is a cornerstone of the success of the department. Your co-leader is reliable and discreet. Your co-leader does what he says he is going to do, he pays his bills on time and he does not share private information you disclose. Further, he thinks about your values and concerns when making decisions that affect you, takes into consideration your values and interests, uses power only when appropriate, stays within bounds of authority, and doesn't use information you don't know to harm you. He says what he means, is easy to interpret, doesn't have anything to hide, only changes his behavior when circumstances require, and will make similar decisions given similar situations. Finally, he understands you well, is socially perceptive, knows how you'll respond without needing to ask, can anticipate consequences of behavior, can see how current actions will affect the future, and recognizes connections among events, that is, how one might affect another.

10. TRUST IN RELATIONAL FORMS

299

In this case, we included all of the potential relationship characteristics as given in the relational partner. Again, we asked participants to rank and rate features and feature sets as to which they feel would be most important given the relationship described. Whereas in the first version we grew the relationship, in this version the relationship retreated through the types. Participants moved from deep interdependence to shallow interdependence (i.e., running a conference with a manager from the hardware division), to deep dependence (i.e., the current department and another are to be integrated and the other co-leader will now be head of the newly formed department), and to shallow dependence (i.e., the participant decides to go into business for him or herself selling software to the former employer). Parallel to the first version, as participants move through the relationship descriptions, they are asked to indicate, considering how the relationship has changed, which feature set they would be most willing to give up in the relational partner. While we are still in the process of collecting data for Study 3 (and therefore do not have the strength in numbers to give us the power to make solid statistical claims), the findings generally support our speculations from the theory and somewhat support the findings of Study 2. In the growth condition, where participants began in a shallow dependent relationship but shifted into a deeper or more interdependent relationship, they behaved much as the theory would suggest, but also in a manner consistent with the previous study. Specifically, as the relationship went from shallow dependence to shallow interdependence, 53% of participants rated the shallow interdependence features as most important, whereas 27% rated the deep interdependence features as most important and 20% rated the deep dependence features as most important. When moving from shallow dependence to either deep dependence or deep interdependence, 50% of participants rated the deep dependence features as most important while 30% rated the shallow interdependence features as most important. Deep interdependence characteristics were the lowest at 20%. In the retreat condition, in which participants were started in a deep interdependent relationship and shifted back through shallower and less interdependent relationships, people seemed to be least willing to give up the features of deep interdependence, regardless of the move being made. In moving from deep interdependence to deep dependence, 53% of participants rated the deep interdependence features as those they'd be least willing to give up, whereas only 27% rated the deep dependence features as those they'd be least willing to give up. And in moving from deep interdependence to a shallower form, 50% of participants were least willing to give up the deep interdependence features, while only 28% were unwilling to give up features designed to mitigate against the risks of shallow interdependence. These data suggest a fascinating paradox. People seem to be

300

HOUDE ET AL.

FIG. 10.1. In the growth condition, participants were asked to indicate the most important feature set as their relationship changed from shallow dependence to shallow interdependence or deep dependence or interdependence. The two lines show the difference between those moving to shallow forms versus deep forms.

quite concerned with mitigating the risks of greater depth and interdependence until a relationship is really deep and interdependent. Like in the earlier study, deep interdependence is perceived in quite different ways from the other types of relationship. Of course, this last study requires more subjects and other scenarios to be truly persuasive, but for now three clear conclusions arise from this collection of research. First, people find it natural and easy to think about interdependence and depth as characteristics of relationships. Second, shallow dependence, shallow interdependence, and deep dependence seem to be relationships for which people generally have risk considerations consistent with the model suggested by Sheppard and Sherman. In the third study, in particular, subjects expressed concerns for the risks of relationships as a function of interdependence and depth consistent with the model as the relationship grew from shallow dependence to other forms suggesting that they understand the risks associated with each change in form and desire elements of trustworthiness commensurate with those risks. These results are consistent with the theory. Even the results for deep interdependence are consistent, as subjects entering deep interdependence would be expected to be most concerned with mitigating the risks of greater depth before worrying about greater interdependence. Depth entails quite serious risks such as abuse and neglect and thus merits mitigation before

10. TRUST IN RELATIONAL FORMS

301

FIG. 10.2. In the retreat condition, participants were asked to indicate the feature set they would be least willing to give up as their relationship changed from deep interdependence to deep dependence or shallow interdependence or dependence. The two lines show the difference between those moving to shallow forms versus deep forms.

potential problems of coordination and anticipation are mitigated. The data from the second study are a little less clear. It seems that people were more concerned with the risks associated with shallow interdependent and deep dependent relationships than with shallow dependence, but not in a truly differentiated way. The dilemma with the data from this study is that it was drawn from relationships that differ on many factors other than just depth and interdependence. Both studies, however, suggest something interesting about deep interdependence. In theory, this is the relationship type with the greatest risk. There was no evidence of a concern about these risks in the second study as subjects consistently rated it lowest in terms of risk. In the third study, subjects did want to add characteristics related to deep dependence when the form shifted to deep interdependence. As just indicated this result is quite consistent with the theory. To tease out additional concerns in the move to deep interdependence would require a different methodology. Especially interesting, however, is the data from the one-half of the subjects in the retreat condition. In all instances no matter what the new relationship subjects retreated to they wished least to lose the characteristics typical of trustworthy behavior in deep interdependence. People want to keep the following characteristics once established above all others no matter how the relationship is transformed, "Understands me well. Socially per-

302

HOUDE ET AL.

ceptive. Knows how I'll respond without needing to ask. Can anticipate consequences of behavior. Can see how current actions will affect the future. Recognizes connections among events, that is, how one might affect another." How do we reconcile this asymmetry? The explanation most consistent with the data is that deep interdependent relationships are built with experience. Reliable, predictable, and benevolent behavioral experiences lead to a willingness and an ability to grow deep interdependence. Once a deep interdependent relationship exists, however, it is transformational. The relationship itself entails conditions that do not require the same protections against risks that were required to build the relationship in the first instance. People are not worried about the very real risks of deep interdependence once the relationship is in place, because the relationship itself provides the mechanisms for managing risk. Two characteristics, in particular, come to mind. First, the level of intimacy and resulting understanding about each other likely serves to mitigate most risks. The more one understands another, the easier it is to predict and develop effective mechanisms for anticipating and coping with any of the risks. Second, deep interdependent relationships entail the long arm of the future (Rempel et al., 1985; Sheppard & Tuchinsky, 1996a, 1996b). Any party has the potential for serious harm if he or she is to engage in inappropriate behavior with someone with whom he or she has significant dependence. There are real costs possible from either a sanction from the other or the removal of the relationship. Of course, we all know of relationships in which these protections do not exist, especially deep interdependent relationships that are perceived as purely dependent by one party. However, generally this type of relationship may be self-insuring. It is either a relationship in which the dependency is asymmetric or that is shallow that the need for protection arises, as these forms of self-insurance do not really hold. A direct implication of this logic is that we should see greater existence of institutional mechanisms to protect dependency relationships or shallow interdependent relationships than deep interdependent ones. If a relationship is self-insuring, then society does not need to insure it. This seems intuitively correct, as it is hard to think of the equivalent of labor laws, credit bureaus, or tort law for deep interdependent relationships. In fact, if there is protection it seems more to insist on the continuity of deep interdependency than the protection from risks in the relationship, such as in difficult divorce laws. However, this intuition deserves serious research. Our subjects' preoccupation with trust in shallow and dependent relationships versus deeply interdependent relationships is borne out by recent dramatic events at Enron, Tyco and WorldCom. In all instances shareholders who are dependent on the trustworthiness of the internal auditors, the external auditors in the form of Arthur Andersen, the audit committees

10. TRUST IN RELATIONAL FORMS

303

of the boards of those organizations, and the SEC were let down. It is the absence of trustworthiness and the resulting need for institutional trust mechanisms that is receiving most attention in the discussions of these events. How do we reduce the conflict of interest for the investment banks that both sell services to firms and do the research on those firms and the auditors who provide the audit intended to discover mistreatment of accounting practices, but who also make further income from consulting services provided to the clients being audited? The primary concern is that investors do not have the interactions to determine the trustworthiness of these actors, nor is there sufficient interdependence to insure trustworthy behavior. The deep interdependence is between the firm being audited or researched and the auditor or investment bank. In such instances there is an emerging outcry for some form of institutional control. Of course, this is problematic, as the presence of a public audit by a purportedly impartial outside firm was supposed to be just such a guarantee. Arthur Andersen was the institutional control mechanism. The question arises over who audits the auditors. The same holds for employees of these firms now in the news, who while more directly involved with the activities of the firm did exist in a context of deep dependence with invisible actors in the form of the senior management teams of the firms. An especially abhorrent violation of the trustworthiness required in deep dependence was the action by the senior leaders of Enron who while telling employees to continue their investments in the retirement fund at Enron were disposing of their own stock. Fascinatingly, the dramatic behavior of these few high profile instances of trust violation in dependency relationships has had huge consequences. One of the most prestigious professional services firms in the world in Arthur Andersen has entirely failed in the course of a few months. The whole value of U.S.-based stocks has declined dramatically and with it the value of the U.S. dollar. The absence of faith in institutional controls in key dependency relationships has huge consequences. Our subjects might have in the less dramatic circumstances of our research understood a basic notion illustrated by these high profile failures of trust—dependency without the ability to engage in discovery or without the ability to apply sanctions typical in deep interdependence is a circumstance fraught with potential trust problems. The scale of these consequences was not unlike the factors leading to the labor movements and immense labor law at the start of this century. Absence of faith in the ability of management to act in a trustworthy manner toward employees was a large factor behind the emergence of elaborate laws and institutional practices to permit the evolution of a more interdependent relationship. The ability to create unions and organized resistance with the possibility of legal strike behavior created a form of deep interdependency

304

HOUDE ET AL.

between management and labor that while introducing significant inefficiency into the system allowed the development of more trustworthy and employee centric managerial behavior. A second implication of our research speaks to the meaning of trust. It is highly likely that the trust that emerges from the relationship itself rather than from assurances or even characteristics of the other party has different meaning to people. Much of the literature on trust has related to elements of trust or trustworthiness. Less attention has been directed toward trustful relationships. The quite dramatic differences between deep interdependent relationships and the other types of relationship suggest this is an issue deserving greater merit. In some sense, our view of trust, a very social concept, may be undersocialized. We have focused potentially too much on contextual or personal factors in trust, when truly deep trust is actually inherently interpersonal. This insight might have some suggestions for how to begin to build a solution to the trust crisis in the economy today. In some sense any answer needs to create a feeling of relationship or connectedness between those who are acting on our behalf and us, their dependents. CEOs, CFOs, and auditors may have become too distant in their own and our minds to feel that trustworthiness is an essential component of their role. If we consider the example with which this chapter began, parents simply understand that they need to behave in a manner that supports the needs and interests of their children. The relational connection is tangible, visible, and inescapable. We may need to consider how to build proxies for this in our solutions to the creation of trust within huge organizations and within an extremely complex isolated market. In any case, the data provide encouraging evidence for the usefulness of the model built by Sheppard and Sherman (1998) and some potentially important qualifications of it. It further suggests the value of Fiske's taxonomy as a mechanism for providing clarity on another important aspect of social relations, that being trust. Finally, the data suggest a key insight about trust and relational form; while the risks of deep interdependence or communal sharing relationships may be highest, the relational form may be self-regulating. Trust that emerges from the features typical of deep interdependence is a thing that like our subjects we should desire to build and be loath to give up in any relationship of importance no matter what form it takes.

REFERENCES Bern, S. L. (1993). The lenses of gender: Transforming the debate on sexual inequality. London, England: Yale University Press. Berscheid, E. (1994). Interpersonal relationships. Annual Review of Psychology, 45, 79-129.

10. TRUST IN RELATIONAL FORMS

305

Bromiley, P., & Cummings, L. L. (1995). Transaction costs in organizations with trust. In R. J. Lewicki, R. Bies, & B. H. Sheppard (Eds.), Research on negotiation in organizations (Vol. 5, pp. 219-247). Greenwich, CT: JAI Press. Burgess, R. L., & Huston, T. L. (1979). Social exchange in developing relationships. New York: Academic Press. Butler, J. K. (1991). Towards understanding and measuring conditions of trust: Evolution of a conditions of trust inventory. Journal of Management, 17, 643-663. Dasgupta, P. (1988). Trust as a commodity. In D. G. Gambetta (Ed.), Trust (pp. 49-72). New York: Basil Blackwell. Deutsch, M. (1958). Trust and suspicion. Journal of Conflict Resolution, 2, 265-279. Deutsch, M., &Krauss, R. M. (1960). The effect of threat upon interpersonal bargaining. Journal of Abnormal and Social Psychology, 61, 181-189. Eisenhardt, K. (1989). Agency theory: An assessment and review. Academy of Management Journal, 14, 57-74. Fama, E. F., & Jensen, M. C. (1983). Agency problems and residual claims. Journal of Law and Economics, 26, 301-325. Fiske, A. P. (1990). Relativity within Moose culture: Four incommensurable models for social relationships. Ethos, 18, 180-204. Fiske, A. P. (1993). Social errors in four cultures: Evidence about universal forms of social relations. Journal of Cross-Cultural Psychology, 24, 463-494. Fiske, A. P. (1995). Social schemata for remembering people: Relationships and person attributes that affect clustering in free recall of acquaintances. Journal of Quantitative Anthropology, 5, 305-324. Fiske, A. P., & Haslam, N. (1996). Social cognition is thinking about relationships. Current Directions in Psychological Science, 5, 143-148. Fiske, A. P., & Haslam, N. (1997b). The structure of social substitutions: A test of relational models theory. European Journal of Social Psychology, 25, 725-729. Fiske, A. P., Haslam, N., & Fiske, S. (1991). Confusing one person with another: What errors reveal about the elementary forms of social relations. Journal of Personality and Social Psychology, 60, 656-674. Gabarro, J. (1978). The development of trust, influence, and expectation. In A. G. Athos & J. J. Gabarro (Eds.), Interpersonal behavior: Communication and understanding in relationships (pp. 290-303). Englewood Cliffs, NJ: Prentice-Hall. Giffin, K. (1967). The contribution of studies of source credibility to a theory of interpersonal trust in the communication department. Psychological Bulletin, 68, 104-120. Gomez-Mejia, L. R., & Balkin, D. B. (1992). Determinants of faculty pay: An agency theory perspective. Academy of Management Journal, 35, 921-955. Good, D. (1988). Individuals, interpersonal relations, and trust. In D. G. Gambetta (Ed.), Trust (pp. 131-185). New York: Basil Blackwell. Hart, K. M., Capps, H. R., Cangemi, J. P, & Caillouet, L. M. (1986). Exploring organizational trust and its multiple dimensions: A case study of General Motors. Organizational Development Journal, 4, 31-39. Haslam, N. (1994a). Categories of social relationship. Cognition, 53, 59-90. Haslam, N. (1994b). Mental representation of social relationships: Dimensions, laws, or categories? Journal of Personality and Social Psychology, 67, 575-584. Haslam, N., & Fiske, A. P. (1992). Implicit relational prototypes: Investigating five theories of the cognitive organization of social relationships. Journal of Experimental Social Psychology, 28, 441-474. Haslam, N., & Fiske, A. P. (1999). Relational models theory: A confirmatory factor analysis. Personal Relationships, 6, 241-250. Jensen, M. C., & Meckling, W. M. (1976). Theory of the firm: Managerial behavior, agency costs, and ownership structure. Journal of Financial Economics, 3, 305-360.

306

HOUDE ET AL.

Kelley, H. H., & Thibaut, J. W. (1978). Interpersonal relations: A theory of interdependence. New York: Wiley. Kramer, R. M., & Brewer, M. B. (1984). Effects of group identity on resource use in a simulated commons dilemma. Journal of Personality and Social Psychology, 46, 1944-1957. Kramer, R. M., & Brewer, M. B. (1986). Social group identity and the emergence of cooperation in resource conversation dilemmas. Frankfurt, Germany: Peter Lang. Larzelere, R., & Huston, T. (1980). The dyadic trust scale: Toward understanding interpersonal trust in close relationships. Journal of Marriage and the Family, 42, 595-604. Levinger, G. (1980). Toward the analysis of close relationships. Journal of Experimental Social Psychology, 16, 510-544. Lierberman, J. K. (1981). The litigious society. New York: Basic Books. Luce, R. D., & Raiffa, H. (1957). Games and decisions: Introduction and critical survey. New York: Wiley. McFall, L. (1987). Integrity. Ethics, 98, 5-20. Miller, G. R., & Boster, F. (1988). Persuasion in personal relationships. In S. W. Duck (Ed.), Handbook of personal relationships (pp. 275-288). London: Wiley. Mishra, A. K. (1996). Organizational responses to crisis: The centrality of trust. In R. M. Kramer & T. R. Tyler (Eds.), Trust in organizations: Frontiers of theory and research (pp. 261-287). Newbury Park, CA: Sage. Murnighan, J. K. (1991). The dynamics of bargaining games. Englewood Cliffs, NJ: Prentice-Hall. Rempel, J. K., Holmes, J. G., & Zanna, M. P. (1985). Trust in close relationships. Journal of Personality and Social Psychology, 49, 95-112. Ring, P. S., & Van de Ven, A. H. (1989). Formal and informal dimensions of transactions. In A. H. Van de Ven, H. Angle, & M. S. Poole (Eds.), Research on the management of innovation: The Minnesota studies (pp. 171-192). New York: Ballinger. Sheppard, B. H., & Sherman, D. M. (1998). The grammars of trust: A model and general implications. Academy of Management Review, 23(3), 422-437. Sheppard, B. H., & Tuchinsky, M. (1996a). Micro OB and the network organization. In R. M. Kramer & T. R. Tyler (Eds.), Trust in organizations: Frontiers of theory and research (pp. 140-165). Newbury Park, CA: Sage. Sheppard, B. H., & Tuchinsky, M. (1996b). Interfirm relationships: A grammar of pairs. In B. M. Staw & L. L. Cummings (Eds.), Research in organizational behavior (Vol. 18, pp. 331-373). Greenwich, CT: JAI Press. Solomon, L. (1960). The influence of some types of power relationships and game strategies upon the development of interpersonal trust. Journal of Abnormal and Social Psychology, 61, 223-230. Strickland, L. H. (1958). Surveillance and trust. Journal of Personality, 26, 200-215. Thompson, J. D. (1967). Organizations in action. New York: McGraw-Hill. Tuchinsky, M. (1997). The role of linguistic proxies in relational negotiations. Unpublished doctoral dissertation, Duke University, Durham, NC. Williamson, 0. E. (1981). The economics of organization: The transaction cost approach. American Journal of Sociology, 87, 548-577. Williamson, O. E. (1985). The economic institutions of capitalism. New York: Free Press.

P A R T

V RELATIONAL MODELS IN THE CLINIC

This page intentionally left blank

C H A P T E R

11 Depressed Mood as an Interpersonal Strategy: The Importance of Relational Models Nicholas B. Allen Paul Gilbert Assaf Semedar

Depressed mood represents something of a conundrum. It is primarily characterized by apparently dysfunctional decrements in positive affectivity and losses in the experience of pleasure (Clark, 2000), and yet it has many of the features of a behavioral adaptation (i.e., a species typical set of behaviors that have evolved to solve an adaptive problem). Depressed mood is ubiquitous, affecting most individuals from time to time (Izard, 1991). It is also reliably activated by specific kinds of contexts or circumstances, such as major defeats and losses (Oatley, 1992). Moreover, it appears that those who show a reduced capacity for depressed mood, or mood variation in general, such as those with hypomanic personality (Klein, Lewinsohn, & Seeley, 1996) or with specific types of damage to the prefrontal cortex (Damasio, 1994), are less able to adapt to changing social circumstances. For example, those with hypomanic personality appear unable to maintain stable interpersonal relationships and can be injudicious in their assessment of risk (Klein et al., 1996). However, how depressed mood might be adaptive remains unclear. There are now a number of models that propose that the determinants of depressed mood lie in the interpersonal or interactional realm (e.g., Gotlib & Hammen, 1992; Joiner & Coyne, 1999). Critical empirical observations here include findings demonstrating that depression is often precipitated by interpersonal events, and that interpersonal processes often mediate the exacerbation or resolution of depressive episodes (Joiner, Coyne, & Blalock, 1999). For instance, stressful interpersonal contexts are among the 309

3 I0

ALLEN, GILBERT, SEMEDAR

most reliable antecedents of depressed states (e.g., Brown, Harris, & Hepworth, 1995; Monroe, Rohde, Seeley, & Lewinsohn, 1999), suggesting they can have etiological significance. The prognostic significance of interpersonal processes is suggested by the findings that those recently treated for clinical depression are at greater risk of relapse if their family environment is characterized by high levels of negative affect expression (Coiro & Gottesman, 1996). Depressed individuals often generate stress within interpersonal contexts (e.g., Davila et al., 1995), tend to preferentially attend to signs of social threat (Matthews, Ridgeway, & Williamson, 1996), and inhibit the interpersonal expression of positive affect, while increasing interpersonal expressions of negative affect (e.g., Biglan et al., 1985; Hokanson, Sacco, Blumberg, & Landrum, 1980). Moreover low positive affect (the primary characteristic of depressed mood) is associated with reduced social engagement and time spent socializing (Watson, 1988), engagement in a more restricted range of social activities (Watson, Clark, Mclntyre, & Hamaker, 1992) and lower ratings of the quality of social relationships (Berry & Hansen, 1996). Adaptive Possibilities

Taking depressed mood as primarily being a lowered capacity to experience pleasure and be sensitive to reward (Clark, 2000), a key question concerns the possible adaptive benefits of low mood states. Before exploring some of the possible adaptive functions for low mood we should note that many of the affect control systems in humans (including those for anxiety and anger) were laid down over the long history of mammalian and primate evolution (Panskepp, 1998). This means that many adaptations that may once have been useful may no longer be so in new human contexts. There are two (not necessarily mutually exclusive) classes of adaptive explanations for human depression: Nonsocial and social. A number of nonsocial theories suggest that depressed mood is, or was, useful when an individual lost control over important external contingencies, such as aversive events. It creates a state of passivity and "giving-up" (Seligman, 1975). Klinger (1975) suggested that depressed mood serves the function of disengaging from pursuing unobtainable resources and allows for a redirection of efforts. More recently, Nesse (2000) suggested that depressed moods function to stop people investing in poor pay-off activities and to conserve resources (Beck, 1987). On the other hand, social theories have focused on the interpersonal elicitors and effects of mood, especially control of salient roles. For example, depressed mood may have evolved to facilitate the cessation of distress calling in animals (especially infants), in order to reduce the risk of drawing attention to themselves as unprotected when they have lost contact with a protecting parent (Bowlby, 1980). Social com-

11. DEPRESSION AND RMT

311

petition theory (Price, 1972; Price, Sloman, Gardner, Gilbert, & Rhode, 1994) suggests that depressed mood originally evolved with a set of behaviors (e.g., social withdrawal and submissiveness) that serve the function of inhibiting animals from pursuing resources when to do so could bring them into injurious contact with more dominant others. A toned down positive affect system therefore protects animals from acting out aspirations that could be dangerous when submissive behavior and passivity are more adaptive (e.g., in the context of defeats). This model seeks to account for the common experience of feeling inferior to others (low rank), submissive behavior, and inhibitions on confidence that are so common to depressed mood (Allan & Gilbert, 1997). Gilbert (1992) argued that the social competition theory was really one that focuses on the mechanisms for the control of behavior in relationships that are seen as socially ranked, with superior and inferior actors, and relabeled it social rank theory. This theory has focused on what happens when people find themselves or believe themselves to be in unwanted (involuntary) subordinate positions that limit their control in relationships and over social resources. It has little to say about drives for dominance as such, but focuses on involuntary subordination and the fear of subordination, because of the associated outcomes (e.g., attacks from above, rejection, exclusion). If subordination does not carry these negative outcomes, or if there are pay-offs from being subordinate (e.g., one is looked after by the more dominant) then individuals may not fear it. Indeed, in some cases people will work to not be dominant. In social risk theory (see below) the rank one wants, be it low, average, just above average, or dominant, depends on the perceived benefits and risks. Moreover, a number of people have noted that the tactics for gaining and losing social rank (being seen as a superior, acceptable, or inferior agent) are often not aggressive as such but depend on being seen as attractive by others (Barkow, 1989, Gilbert, 1992, 1997) or making a "good impression" (Leary & Kowalski, 1990). One can be placed in a subordinate position in a relationship because others see one as undesirable and thus ignore you, or reject you from forming a role (e.g., friend, lover, collaborator) with them. Nesse (1990) and Gilbert (1992) have argued that positive affect (and the physiological mechanisms that modulate it; Gilbert & McGuire, 1998) are responsive to the reception of social signals indicating social control and social rewards. In other words, successfully engaging in evolutionally meaningful desired roles (e.g., sexual, attachment, friendship) gives rise to positive affect, whereas losing control in these roles is associated with negative affect and lowered mood. Depressed mood is still linked to rank-controlling behavior and switches in when people perceive their social standing and social control in relationships (hence their ability to maintain a flow of positive social rewards) to be low (Sloman & Gilbert, 2000). More recently, Gilbert (2001) has

3 I2

ALLEN, GILBERT, SEMEDAR

argued that we need to distinguish the functions of depressed mood, which may be mild and transitory, from clinical depression, and that a key way to do this is to distinguish the functions of the mechanisms that may underpin mood regulation from mood regulation itself. Of specific relevance to this volume is the possibility that each of these theoretical positions appears to be emphasizing interactions within one of Fiske's relational models (Fiske, 1991), and describes the functions of depressive behavior in terms of the particular grammar and logic of that model. Although relational models theory was designed to describe social dynamics and forms of interaction (rather that evolved mental mechanisms and vulnerability to psychopathology); it is clear that social dynamics and cultural forms must, at some level, represent an expression of innate possibilities. For instance, the social competition theory (Price et al., 1994) emphasizes Authority Ranking (AR), whereas attachment approaches (Bowlby, 1980; Ingram, Miranda, & Segal, 1998) appear to be primarily cast within the logic of Communal Sharing (CS) relationships. Approaches that emphasize reciprocal exchange (e.g., Glantz & Pearce, 1989) appeal to Equality Matching (EM) and Market Pricing (MP) models. Thus, each of these theories offers important insights into the possible interpersonal functions of depression. Social Forms and the Meaning of Depressive Symptoms Certainly though, people may judge their social standing and thus control over social resources as low in any of a variety of social contexts. For example one might rate oneself as a good friend (communal) but a poor leader or poor at assertiveness (authority). If the cultural dynamic favors one form of social relating over another, then some people will be disadvantaged. For example, in social contexts where authority is salient, being able to be a willing member of a hierarchy, obey orders, compete to get ahead, and accumulate resources (as was the case for Middle Ages Europe under the Catholic Church with a very strict and clear hierarchy), would be advantageous. However, these same skills may work poorly in communal groups. There is now substantial evidence that for long periods humans existed in small close-knit groups and were hunter-gatherer egalitarians with few clear hierarchies. They lived in harsh ecologies and were highly dependent on each other. As Boehm (1999) has elegantly noted, in these groups highly competitive accumulators would be shunned and ostracized; indeed it was the ability of subordinates to gang up against the dominants and stop their exploitative behaviors by the use of shunning that has had such a major impact on human psychology (Boehm, 1999). It is important therefore that people track the behaviors that are valued by the group and do not act outside of their cultural form, for to do so risks losing social standing in the

11. DEPRESSION AND RMT

3 I3

group, being demoted, shunned, shamed, and ostracized. This "subordinate power" broke down with the advent of agriculture and the value of surplus and control of surplus. If people believe they will not be able to elicit important social resources and investments from others in their culturally accepted forms (e.g., by getting up the hierarchy or being seen as a good communal group member), they may be rejected as lovers, friends, or by authority figures and view themselves as having low attractiveness (low rank) in these domains. However, what is required is an analysis of the deeper level functional principles relevant to depression, in order to explore how this logic may play out within each of the relevant relational models. We believe that such an analysis is provided by examination of social risk assessment in depression. In this case the "risk" is continuing falls in social attractiveness. Moreover, we note that just like other defensive reactions (e.g., anxiety and anger), which can be alternately triggered by authority threats, exchange and reciprocal failures or disruptions to attachment relationships, there are a number of different social contexts that can trigger low and depressed mood. Depressed mood then can function differently in different contexts.

THE SOCIAL RISK HYPOTHESIS OF DEPRESSION The social risk hypothesis of depression (Allen & Badcock, 2003) suggests that depressed mood states, and clinical states of depression, are based on adaptive mechanisms rooted in brain design (Panskepp, 1998). This is not to say that depression is always adaptive in current human contexts, only that the mechanisms that form the substratum for depressed states have been preserved and shaped by natural selection because they solved adaptive problems. While some depressions may be adaptive, as discussed later, other depressive disorders can also be understood as "pathologies" based on poor regulation of adaptive mechanisms (Gilbert, 2001). This is no different from the idea that (say) panic disorder is a disorder of some alarm/anxiety system (Barlow, 1988). Moreover, as for any defensive behavior or affect (be it anxiety, anger, or decrements in positive affects) if it is triggered too easily, too intensely, too frequently, or lasts too long it risks becoming pathological (Gilbert, 2001). The social risk hypothesis of depression suggests that depressed states evolved to facilitate an extremely risk-averse approach (a kind of damage limitation strategy) to social interaction in situations where individuals perceive their social resources (e.g., status, affection, friendship, power) to be at critically low levels. There are probably a number of different tactics depending on the context. For example, animals operating at the bottom of the hierarchy, in the context of serious down rank attacks, might utilize a

3 I4

ALLEN, GILBERT, SEMEDAR

strategy of demobilization and highly inhibited challenge behavior. A timid depressed-like state might be sufficient to keep the animal out of trouble for a period of time as well as indicating to a dominant that there is no need to further attack it or try to suppress its resource competing behavior. On the other hand, in a context of social affiliation, where it is integration and support within a group that matters, individuals who have very low levels of such resources (e.g., are seen as inferior agents) must try to increase their share of social supports, but not to the extent that others are driven away— a situation that is rather common in some depressions and an example of poor regulation of the defensive behavior of help seeking (Segrin & Abramson, 1994). In this respect, the basic depressive strategy of conservative, risk avoidant, social behavior will appear in a variety of tactical variations depending on the particular social context. In this sense depressed mood (i.e., reduced positive affect) was (in the evolutionary past) a defensive-protective strategy. This may help explain why depression can be associated with paranoia, poor assertiveness, passivity, and strong desires to escape particular contexts (Brown et al., 1995), none of which are predicted by the conservation of resources, or attachment-loss models. The Social Context as an Adaptive Landscape How and why do social threats exert such a powerful influence on mood states? To explore this question, we have to consider both the ways that the social environment has exerted selective pressure on mental mechanisms, and the reasons for believing that social behaviors are often fundamental to common human depressed states. Many writers have noted that our forebears' social environment provided some of the most important adaptive problems influencing human evolution (Brothers, 1997; Byrne & Whiten, 1988; Cosmides, 1989; Humphrey, 1976). Cosmides, Tooby, and Barkow (1992) have defined an "adaptive problem" as a "problem whose solution can affect reproduction, however distally" (p. 8). Although classical Darwinian theory measured fitness in terms of individual reproductive success, in neo-Darwinian analyses of biological evolution the primary unit of selection is the gene, not the individual organism (Williams, 1992). Thus, it is the gene's fitness (i.e., its likelihood of being replicated in successive generations) that has become the focus of neo-Darwinian analyses. Because individuals are likely to share genes with their kin, those behavioral strategies that enhance the survival and reproductive success of an individual's kin will also be selected (Hamilton, 1964; Maynard Smith, 1964). This process of kin selection contributes to what is known as "inclusive fitness," and points to the fundamental involvement of interpersonal behaviors in natural selection.

11. DEPRESSION AND RMT

315

In this vein, Buss (1991) outlined the following social-reproductive tasks faced by humans: (a) intrasexual competition to attract desirable mates, (b) choice of the mate with the greatest reproductive value, (c) successful conception, (d) mate retention, (e) reciprocal dyadic alliance formation, (f) coalition-building and participation in co-operative groups, (g) ensuring survival and reproductive success of one's offspring, and (h) investing in kin other than one's offspring. Gilbert (1989, 1995) derived similar classes of biosocial goals, including care seeking, care giving, cooperation in dyads and groups, and status acquisition (competition). The pursuit of each of these goals is not without risk. For example, cooperating in a group of exploiters might mean they benefit at one's expense, and competing for status and access to resources with others who are more powerful might elicit constant attacks and humiliation. Thus it is important that individuals eva uate risks against benefits in whatever biosocial goal or role they are pursuing. Risks of attachment loss (separation) for a young child will be calculated and coped with in a different way to (say) risks encountered while competing for sexual partners. This means people will need to monitor the roles they are engaged in and to increase certain behaviors when opportunities exist, and to inhibit them when they entail too much risk of harm (see Gilbert, 1989, pp. 26-27). In order to characterize the social environment as an adaptive landscape, we need to describe how social interactions affect reproduction. In the case of mating relationships and inclusive fitness, such effects are quite proximal and (relatively) easy to track. In the case of nonsexual relationships and alliances, the critical mechanism is the effect of the relationship on an individual's access to nonplentiful resources (i.e., resources that others are seeking in competition with oneself). These resources, in turn, provide an advantage to the individual in solving adaptive problems more proximal to reproduction (e.g., attracting and retaining mates of high reproductive value, or caring for offspring). The kinds of nonplentiful resources that may be accrued as a result of relationships with others are quite varied. For example, Foa and Foa (1974) suggested that human social relationships variously result in the exchange of love, status, information, money, goods, or services; each of which may be used to enhance the solution of more proximal reproductive tasks. Therefore, the opportunity to engage in resource exchange and collective action in relationships has the potential to deliver significant fitness-enhancing benefits to individuals. For certain species, social relationships will be among the most powerful determinants of differential fitness within populations. This is particularly the case for humans (and many other primate species) who have highly developed cognitive skills, exchange goods and favors over long periods of time, form large structured coalitions for purposes such as defense from predation and foraging for food, and who give birth to offspring that require a very extended period of parental care in order to grow to maturity.

3 I6

ALLEN, GILBERT, SEMEDAR

Just as social relationships are a major determinant of fitness benefits, they must also determine who does not receive benefits, or who receives relatively few benefits. General terms used to describe the relative amount of benefit an individual receives in social relationships are social status and dominance, which have usually been defined among both humans and other animals in terms of an individual's control over nonplentiful resources (Ellis, 1993). When the pay-offs of a social interaction are highly asymmetric (perhaps even involving net benefits for one individual and net costs for another), dominant individuals will usually be able to ensure that they receive the better outcome, either by winning contests for resources (MaynardSmith & Price, 1973), or by forcing others to cooperate via a threat of punishment (Glutton-Brock & Parker, 1995). However, not all relationships are characterized by strongly asymmetric outcomes for the participants. Many interactions (perhaps the majority in the human context), involve either mutualism, where both individuals receive their greatest benefit from interacting cooperatively (Pusey & Packer, 1997), or reciprocal altruism, where individuals will deliver benefits to each other on the expectation that benefits will later be delivered to them in return (Trivers, 1971). In such contexts, the critical determinant of the fitness benefits an individual receives from social interactions will not be whether they can extract a competitive advantage at the expense of their social partners, but rather by which partners will conduct relationships with them at all (cf. Orbell & Dawes, 1993). In particular, the opportunity to engage in mutually beneficial relationships with more elite (i.e., resource rich) partners, and to participate in more elite groups, powerfully influences an individual's access to resources. Thus, in negotiating the adaptive challenges posed by the social environment, individuals must keep track of (at least) two matters: whether they are at risk of demotion within, or exclusion from, currently beneficial relationships, and whether they are maximizing the benefits that they can receive from their contribution to social relationships. The way in which an individual balances these concerns will reflect a trade-off between protective-defensive (risk minimization, or damage limitation) strategies (i.e., that protect social relationships from critical damage), and expansive (risky) strategies (i.e., that increases the benefits derived from social relationships by asserting greater dominance or control in current relationships and seeking access to more elite social relationships). Avoiding Demotions and Social Exclusion

A critical matter for the individual is to detect when the danger of demotion or exclusion from currently beneficial social groups or dyadic relationships is high. Hirshleifer and Rasmusen (1989) pointed out that in economic terms, exclusion or ostracism is an action that has costs for both the indi-

11. DEPRESSION AND RMT

317

viduals who are ostracized, and the group or individual who shuns them. This is the case because the group loses the resources brought to collective tasks by the ostracized individual. If an individual is high in such resources, the group will be loath to ostracize him or her, even when he or she is burdensome in other ways, because such ostracism will be too costly. However, if individuals are considered to be low in resources, it will cost the group little to exclude them, and may even be advantageous (i.e., reduce the number of competitors making claims on resources). This suggests that individuals would want to track the costs and benefits, to others, of their participation in relationships in order to ensure that such relationships were not at risk of critical damage. Demotions should also be avoided, even though a person may not be specifically ostracized. If their rank falls too low they will have little access to resources (e.g., high quality mates and allies won't bother with them). Theory of Mind. These exclusion and demotion threats may be tracked by monitoring one's relative attractiveness or desirability to others (Gilbert, 1992). There are a number of reasons why an optimal solution to the problem of detecting when relationships are at risk of termination, or a reduced flow of social resources (e.g., support, care, sexual interest), would involve complex judgments about the mental states of others. The first reason is that the cost to an individual of the resources they provide to others, and the benefit of these resources to those who receive them, may be very different. The resource may cost the giver almost nothing (as may be the case with some types of information), but provide the recipient with enormous benefit. In such cases it will be critical for an individual to judge the value of the resources they provide, as perceived by others (i.e., how desirable is the resource I offer?). The second reason is that reciprocity (Trivers, 1971), and long-term mutualism (Pusey & Packer, 1997; where a strategy with short term costs may become mutually beneficial when applied in the long term), mean that even though one is not providing resources (or extracting benefits) from a relationship currently, there may be an understanding that such benefits may be provided (or extracted) at a future time when needed. Once again, the perception of an obligation or debt by others is critical. Therefore, both of these factors suggest that it would be most effective to track the benefits and costs of one's participation in a group or relationship, as others perceive them. The ability to cognitively represent the mental states of others has been referred to as the capacity for a theory of mind. Byrne (1995) suggested that there are two areas where awareness of mental states in others might have provided a selective advantage—intentional deception (i.e., the capacity to manipulate the mental perspectives of others to one's own advantage), and intentional teaching (i.e., the capacity to transfer knowledge or skill to kin

3 I8

ALLEN, GILBERT, SEMEDAR

or a collaborator). Our analysis here suggests that the need to detect when critical social relationships were at risk may have also been an area where selective advantage was conferred by this ability. The critical variable to control in order to prevent the loss of beneficial relationships, therefore, is the ratio between the resources that are provided to others as a result of one's participation, as perceived by others (which we will call social value; Sv), and the costs to others (i.e., loss of current or potential resources) of one's participation in the relationship, as perceived by others (which we shall call social burden; Sb). When St; and Sb are approximately equivalent in a current relationship, danger of demotion, rejection, or ostracism is high, and a protective-defensive strategy in social interactions would be most adaptive. Because being accepted, chosen, and desired are so important to human interactions, Gilbert, (1989,1997) suggested that humans evolved evaluative mechanisms to track their attractiveness (and hence investment worthiness) to others by calculating their social attention holding power (SAHP) in specific relationships. External SAHP is simply a shorthand for estimates of how much attention and interest one believes one can elicit in others. It can be positive (e.g., we are able to attract positive attention-interest from others, stimulate positive feelings in others, with a low risk of ostracism) or negative (e.g., we attract negative attention, stimulate negative feelings in others such as contempt, anger, or fear with a high risk of ostracism). It is interesting to note that recent work on self-esteem has also conceptualized a similar process called a "sociometer," which monitors others' reactions and alerts individuals to the possibility of social exclusion (Leary, Tambor, Terdal, & Downs, 1995). The concept of Sv/Sb ratio may offer a more precise way of conceptualizing social attention holding and sociometer concepts. Among species that track their social value and burden as described earlier, the most important social resource becomes not strength, speed, or goods, but rather social prestige itself. Individuals who are highly regarded by others (because they are perceived as potentially capable of providing benefits) can utilize this prestige itself to obtain access to elite, resource rich groups and alliances. Foa and Foa (1974) noted that status or prestige can be exchanged within relationships, and therefore has a resource value of its own. The point is that much of the evolved psychology of social value has become focused on the perceptions of others per se, rather than the material resources that might underlay or be implied by such perceptions. Maximizing Return on Social Investments Management of the resources gained by one's current social investments can also be achieved by controlling the ratio between Sv and Sb. If one's social value far exceeds one's social burden in a currently beneficial relation-

11. DEPRESSION AND RMT

3 I9

ship, this can be a signal that one should attempt to extract greater value from social investments. Greater return on investments can be achieved either by making greater demands for resources in current relationships, or by dis-investing in the current relationship (freeing up resources to be invested in other, more rewarding, relationships). Although such behavior is risky, exposing an individual to the possibility of rejections, rebuffs, and defeats, such setbacks are probably not critical in a context when Su clearly exceeds Sb in one's current relationships (implying likely success in these social maneuvers). We therefore conjecture that humans will have evolved mental mechanisms that attempt to control the ratio of Su and Sb around a particular value. In the evolutionary past this value represented a balance between the need to ensure that social value exceeded social burden comfortably (implying that current social relationships were not at risk), and the need to ensure that one was obtaining the greatest benefit from one's social investments. When Sv is relatively high, opportunities to improve social opportunities or pay-offs loom larger than the costs of potential demotion, rejection, or ostracism, and thus a person will have an opportunity-seeking (and therefore more risky) approach to social interactions. However, as the cost of setbacks and losses increases, these costs will loom larger than opportunities, and it would have been adaptive for the organism to shift to a primarily cost-avoidant (and therefore risk-avoidant) strategy, one which sought to raise the value of Sv/Sb ratio conservatively, without risking further reductions. Indeed, a considerable amount of theory and research has shown that humans are strongly motivated to avoid social demotions and exclusions (e.g., Baumeister & Leary, 1995; Baumeister & Tice, 1990; Leary & Downs, 1995), and that social exclusion is associated with aversive emotional reactions (Leary, 1990, 2001; Williams & Zadro, 2001). Given that human groups contain concentrations of particular reproductive resources, including potential mates, kin to whom altruism can be directed, and nonkin with which to exchange resources, the effect of social exclusion on various proximal adaptive tasks can be considerable (Buss, 1990). In the Pleistocene period, social exclusion may have threatened one's survival by excluding the individual from the benefits of group-based forms of protection from predators or foraging for food. Under threat of social demotion-exclusion, scorn, or shunning, it would be very important to reduce the likelihood of further reductions in the Sv/Sb ratio, however minor or temporary. Such situations would require a qualitatively different strategy from that employed when such affiliations are not under threat. We propose that this situation provided a fundamental selective pressure for the evolution of new mechanisms to regulate interpersonal behavior and social outcomes.

320

ALLEN, GILBERT, SEMEDAR

A CYBERNETIC ANALYSIS OF CONTROL OF SOCIAL VALUE AND SOCIAL BURDEN

Having described the adaptive logic behind an evolved mechanism to control the ratio of social value (Su) and social burden (Sb) around a particular value, we can use the tools of cybernetics, or control systems theory, to describe the functions of such a mechanism. Control systems theory has been used in many areas of engineering to design and understand mechanisms that act on the environment to control a dynamic variable, and much is known about the basic functions required, and how they must interact (Cruse, 1996; Ford, 1987). Furthermore, although the organizational principles of functional design are the same in all control systems, the supporting structural design may be quite varied. A basic feedback control system usually consists of the following elements. The controlled variable represents the environmental variable that the system is attempting to control, while the command or set-point function defines the value at which the system is attempting to maintain the controlled variable. The system monitors the controlled variable via feedback transducers, which are required to represent the controlled variable in such a way that it can be compared to the set point. Due to biases and errors in transducer processes, the feedback signal is not the same as the controlled variable, but rather a representation of it. The feedback signal is compared to the input of the command or set-point function via a comparator, and differences between these two values are then sent to the control function, which "selects" a course of action intended to reduce the difference between the feedback signal and the set-point value. When this course of action is selected, this information is passed on to the action/effector function, which produces the appropriate actions on the environment. These actions, along with disturbance inputs (exogenous effects on the controlled variable) are then applied to an environmental process by which those variables manipulated in the environment by the action functions, have some effect on the controlled variable. Changes in the controlled variable, which can be due to the effects of both the action functions and the disturbance inputs, are then monitored, compared with the set-point value, and acted on again. Control of the social value/social burden ratio can also be expressed in terms of these cybernetic functions. We have already argued that Sv/Sb must be maintained at a value that balances the need for protection from social exclusion and social devaluations against the need to extract maximum value from social investments. This value should determine the setpoint value at which an adaptive system will attempt to maintain Sv/Sb. In order to make adjustments in social behavior, Sv/Sb must be monitored via some type of feedback transducers. In the case of monitoring one's social

11. DEPRESSION AND RMT

32 I

value and social burden as perceived by others, the feedback transducers will need to include a highly complex arrangement of mechanisms. Firstly, the feedback mechanism would need to monitor the resource implications, to others, of one's participation in the relationship. However, for reasons outlined earlier, the mechanism would also need to monitor the kinds of interpersonal behaviors from which mental states in others may be inferred. A useful catalogue of these behaviors can be obtained from Salter's (1995) ethological studies of dominance in institutional settings. He suggested that facial behavior, gaze direction, posture, limb and body movements, conversational tactics (e.g., interruptions, length of utterances, talking through others), voice characteristics (pitch, tempo, volume), and verbal content are all strong interpersonal signals of dominance and affiliation. In order to control Su/Sb, individuals need to process information about resource implications and interpersonal behaviors within relationships, and provide this information to a mechanism capable of inferring other's mental states from this kind of data. As noted earlier such skills have been referred to as theory of mind. Baron-Cohen (1995) and Leslie (1994) have both proposed the existence of a Theory of Mind Mechanism (ToMM) in humans. This mechanism infers the full range of mental states from behavior (e.g., thinking, knowing, believing), and relates these to actions. The ToMM should therefore be able to represent one's Su/Sb ratio for a given individual (or group of individuals) and predict the likely behavior of others on the basis of these mental states. We shall refer to this representation as Perceived Su/Sb, or P(Su/Sb), in order to clearly distinguish this feedback signal (i.e., the mental representation of one's social standing) from the controlled variable (i.e., the actual evaluations by others of one's social value and social burden). It is important to note here that information about the normative cultural and relational rules that govern a particular relationship or context will also be required in order to make inferences about the way social value and social burden are perceived by others. For example body thinness may be a potent attractiveness marker for some females in certain cultures, whereas body size may be important in others (e.g., Sumo wrestlers). In this respect relational models theory provides a description of some of the likely elementary forms. For instance, within Authority Ranking (AR) relationships, social value for up-rank individuals will be primarily defined by their ability to coerce or provide protection to down-rank individuals. Here the more one has or owns, the higher the status. Within a group, individuals who can win competitions (e.g., in sports, or leaders of industry) are given high value, especially if the group or team is competing with others. And within the group, individuals will compete for the top places to avoid being "left out." Such groups, as is common in Western capitalist societies, allow individuals to accumulate vast resources and thus to receive substantial at-

322

ALLEN, GILBERT, SEMEDAR

tention and social status. Within Equality Matching (EM) and Market Pricing (MP) relationships, social value will primarily be defined by an individual's capacity to repay debts incurred during social or economic exchange. Within a Communal Sharing (CS) context, social value will primarily be defined by the perceived veracity of one's claim to belong; to be "one of us." Status will be likely to be based on sharing rather than accumulating resources. This is also being the "good party member," the willing participant who works for the cause. In other words people can be valued (or feared) because they are an aggressive competitor, because they are trustworthy, altruistic and caring of others, or because they are a good team player and hold firm to the group's rules and values. In these domains people can increase their social value by expressing the traits appropriate to the role, but risk shame by failing to do so (Gilbert & McGuire, 1998). Note that competent behaviors in one domain of relating may not transfer to another. For example, being kind and caring, and putting the needs of others first may not be valued in competitive groups. Leaders will often loss status if they assert that their followers must be kind and understanding to the enemy! Competitive and self-focused individuals, who accumulate vast resources, may be shamed in highly cooperative groups and seen as exploitive. In each of these contexts, therefore, understanding the grammar of the social relationships in which social value and social burden estimates will be made is critical to adequate control of social resources. Once a value of P(Sv/Sb) has been established by these feedback mechanisms, it is compared to the set-point value, and the difference is passed on to the control function. When P(Su/Sb) is higher than the set-point value, the control function will instruct the action function to engage in social behaviors designed to realize greater return on social investments. This will include attempts to lower Sv (e.g., by offering less of value in the relationship, or dis-investing) or raise Sb (by making greater requests or claims on resources within the relationship). On the other hand, if P(Sv/Sb) is below the set-point value, the control function will instruct the action function to raise Sv (e.g., by offering more of value to others) or to reduce Sb (by making fewer requests or claims on resources within the relationship). In this way, the system should result in negative (i.e., discrepancy reducing) feedback. We conjecture that the ability to adjust social behavior in the face of changing social circumstances in this way is a key adaptive skill that allows one to maximize the utility of social relationships while also maintaining the degree of reciprocity necessary for relationships to be sustained over time. Clearly the efficacy of an individual's actions in controlling Sv/Sb will also critically depend on both the disturbance inputs (e.g., relevant life events) and the characteristics of the social environment (e.g., the current state and strength of one's social networks, the type of relationship being

11. DEPRESSION AND RMT

323

manipulated). In some circumstances, disturbance inputs and characteristics of the social environment may result in actions amplifying, rather than reducing, the differences between P(Sv/Sb) and the set-point value (cf. Segrin & Abramson, 1994). For instance, if one has a paucity of close social relationships, then attempts to reduce Sb by withdrawing from social contexts may also have the effect of reducing Sv (or at least the perceived signals of Sv) to such a extent that P(Sv/Sb) will actually drop as a result of the behaviors designed to raise it. This will result in positive (i.e., discrepancy amplifying) feedback, which may lead to escalation of mild depressive behaviors into pathological states. Indeed, an important consideration in evolutionary theory is the degree to which the current environment reflects the human environment of evolutionary adaptation (EEA). For instance, in many of the paradigms used in animal studies of the basic mechanisms of depressive disorders they obviously do not. Learned helplessness (e.g., Seligman, 1975; Weiss & Simson, 1985) is a highly artificial condition, as is caging defeated animals with dominant ones (e.g., Von Hoist, 1986). To the best of our knowledge the phenomenology noted in the learned helplessness and social defeat experiments is rarely noted in free living animals because they distance themselves from each other. If we need to create highly artificial environments to obtain these effects in animals (and they certainly can be observed in a variety of species under caged and trapped conditions; MacLean, 1990; Seligman, 1975; Toates, 1995) then we are not studying the adaptive function of evolved mental mechanisms, but rather their malfunctioning because they are operating in environments for which they were not designed. In the same way, we believe that the phenomena noted in clinical depression also represent the malfunctioning of evolved mental mechanisms in current environments. This logic not only applies to mismatch between the EEA and current social environments, but also to mismatch between the relational model that one person utilizes to perceive and control Sv/Sb and the relational model used by their interactional partner(s). For example, it is well established that excessive reassurance seeking is a prominent risk factor for the emergence of severely depressed states (Joiner, Metalsky, Katz, & Beach, 1999). This may well represent the tendency, on behalf of depression prone individuals, to overimplement the CS model in relationships that are not perceived as CS by the partner. In other words, the depressed person seeks help (and evidence of being cared for) but their partner sees giving help as a burden on their resources that is unlikely to be reciprocated. In this context the depressed person calls on help and support with the expectation that such help will be forthcoming, only to find that these requests lead to rejection by their interactional partner. In this situation the mismatch between the depressed person's attempts to control social resources, and the outcome engendered, will result in discrepancy amplifying feedback, and a

324

ALLEN, GILBERT, SEMEDAR

subsequent exacerbation of depressive behaviors. The person then feels less cared for, which resets their P(Sv/Sb) leading to more efforts at care eliciting until a more intense depressed mood forces them to abandon their efforts. If they are able to switch strategies (e.g., become more competitive in the outside world, or strike out on their own) or seek more rewarding partners/relationships, they may be able to recover. Indeed many people do recover from depression this way. But if this is not possible, mood may continue to go down because their strategies are failing. Feedback Versus Feedforward

Feedback systems of the type described in the previous model have a serious limitation with respect to the adaptive control of environmental variables; they reflect only the state of variables in the present moment (or the past), and are unable to predict or respond to future behavior of the variable (Ford, 1987). In order to adapt to future events a feedforward system is required. A feedforward system anticipates future changes in the controlled variable and instigates compensatory strategies by exerting control of action functions (and feedback transducers) that is independent of momentto-moment feedback from the environment. Feedforward control is able to anticipate when a variable is reaching a critical value, and is therefore required to respond to a situation such as when the values of Su and Sb are approaching equivalence, and exclusion from social relationships, or a serious reduction in the flow of resources is a risk. Feedforward systems have the advantage that once they are activated, they are not necessarily dependent on feedback for their functioning. As such, they are able to respond to anticipated circumstances quickly (whereas the feedback process often takes time), and usually have characteristics that have adapted to the dynamics of the system's processes. For instance, a feedforward mechanism should have a "refractory period" based on the time it takes for a feedforward action to have an effect on the controlled variable (i.e., it should not change its level of activation until the system has had time to respond to its initial actions; Cruse, 1996). As we have described, depression can be seen as a response to an anticipated threat to social relationships. For example, one predicts that one will not do well; that a lover may defect or lose interest, or give more time/resources to their job, or that one's efforts may not match those of competitors. It is the prediction of changes in Sv/Sb that is at stake. This strategy is best characterized as a feedforward mechanism. Indeed, it may be useful to think of moods (i.e., episodic affective states that last from minutes to days; cf. Ekman & Davidson, 1994) as feedforward mechanisms that "tune" the performance of feedback systems in response to anticipated environmental

11. DEPRESSION AND RMT

325

contingencies. The depression feedforward strategy is activated whenever the feedback signal indicates that Sv and Sb are approaching equivalence. When this state is detected by the depression control function, it instigates changes in both social perception (feedback transducers) and social behaviors (action functions). Furthermore, the strategy is activated with a refractory period that has, in the evolutionary past, allowed changes in behavior to achieve their goals (i.e., for the Sv/Sb ratio to rise above the critical region) before changes in activation can occur. In this respect the feedforward mechanism is functionally different from the feedback system's response to a low Sv/Sb value in a number of important ways. First, it is less responsive to feedback, and may remain activated for some time after the social environment is no longer critical. Second, the strategies used to raise the value of Sv/Sb can be different from those employed to minimize social risk. A strategy simply aimed at raising Sv/Sb will attempt to invest more in relationships, or reduce their burden on others. Although these strategies invite less risk of exclusion than strategies aimed at reducing the Sv/Sb ratio, they do not necessarily minimize risk. As already stated, depressed strategies functioned to ensure that individuals only sought to raise Sv/Sb in ways that keep the risks of social setbacks at a minimum. This will usually involve a strategy that reduces the variance of outcomes (both gains and losses), with the principal aim of reducing the likelihood of significant costs (Schoener, 1987). This does not mean that a risk-averse strategy will always succeed in avoiding critical losses in social relationships. Indeed, in the current context depression may damage relationships as often as it preserves them. Rather, we argue that on probabilistic grounds, the risk-averse strategy will be best in situations of critically low values of Sv/Sb. It is these probabilities, their influence distributed across a range of individuals and situations, that have exerted the strongest influence on the process of evolution by natural selection.1 'It is important to note that the situation we are describing here is not a situation where an individual has "nothing to lose," but rather a situation where the individual perceives that they have a great deal to lose by incurring further costs to their social reputation. This distinction is critical because results from optimal foraging theory (Schoener, 1987) and game theory (Grafen, 1987) have established that in situations that are so sub-optimal for an individual that starvation or continual, repeated, losses in agonistic encounters are likely, risk prone behaviors will be favored. The selective pressures that we describe for depression are fundamentally different to these "nothing to lose" situations. We propose that depressive behaviors evolved to deal with threat of critical loss, rather than after critical losses have occurred (although clearly past losses can increase the probability of future losses, and in that respect may activate depressive strategies). The occurrence of excessively risk-prone behaviors, such as are seen in many forms of psychopathology, including mania, antisocial personality, and borderline personality, may be an important issue with which to further explore the link between the psychology of risk and psychopathology.

326

ALLEN, GILBERT, SEMEDAR

THE EFFECTS OF DEPRESSION ON SOCIAL BEHAVIOR AND COGNITION

We conjecture that the "depression mechanism" has three broad classes of action that have, in the past, served to minimize the risk of social exclusion (Allen & Badcock, 2003). These include (a) vigilance for indications of social risk, (b) signaling in order to reduce threats and to elicit safe forms of support, and (c) a general reduction in behavioral propensities to social risktaking. Although each of these functions has significance for interpersonal processes, they have been extensively described elsewhere (Allen & Badcock, 2003). In this section we explore the signaling functions of depressed states, as this is where the importance of interpersonal contexts, such as the use of relational models to coordinate social interactions, is the greatest. As pointed out earlier, although previous evolutionary accounts of depression have emphasized communicative functions of depressed states, they have tended to emphasize signaling in one type of relationship, be it competitive (e.g., Price et al., 1994), close attachment (e.g., Bowlby, 1980), or reciprocal (e.g., Glantz & Pearce, 1989). From the perspective of relational models theory, these accounts fail to capture the diversity of relational contexts. As noted earlier, the way in which risk to social value will be perceived and managed will vary according to how relational contexts are perceived, and as such the strategy enacted to protect against losses in social value should differ. Put simply, the social risk hypothesis would predict that there would be differential changes in social behavior depending on the relational context. In communal or other strongly attached relationships, expressions of dependency or care eliciting should predominate, whereas depressed individuals should withdraw from exchange-oriented relationships, and be submissive or exploitive (depending on rank) in AR relationships. One method of protecting against loss might be to re-categorize relationships to lower risk categories (lower risk in the sense that they expose the individual to a lower likelihood of losses in social value). For instance, relationships that are normally conducted according to rules that implicitly govern CS relationships might be cognitively shifted to the EM category during depressed states. This would protect against making excessive calls on the favors of others who may not tolerate them (potentially resulting in rejection). It would also facilitate keeping track of favors given, further protecting against short-term exploitation by others in these contexts. Such recategorization may well have considerable opportunity costs, especially if the change of social behaviors that results from these re-categorizations causes some distress and confusion in others due to changes in the implicit rules of the relationship. The depressed person's resultant anger at favors not immediately returned, while helping to protect against a short-term imbalance in exchange of favors, may in fact act to the detriment of the longer

11. DEPRESSION AND RMT

327

term interests of the relationship. However, as the function of the depression mechanism is to protect social resources from further harm in the short term, this may nevertheless be an adaptive form of re-categorization during times of low social resources. As already noted, a less effective form of re-categorization might be to see exchange oriented or AR relationships as appropriate targets of help seeking behavior during times of depression. Although such help seeking behavior will usually be tolerated, and even effective, within CS relationships, behaviors such as dependency and excessive reassurance seeking (Joiner et al., 1999) will often result in greater risk of rejection within other types of relationships. In this respect it is interesting to note our recent empirical results, which have shown that individuals who are vulnerable to depression tend to overimplement the CS relational model, even in relationships where it is not normative such as one's relationship with one's boss (Allen, Semedar, & Haslam, 2003). Evidence on Social Signaling in Depression There is little question that depressed behavior represents a strong social signal, and empirical research has outlined a range of changes in interpersonal behavior associated with depressed states (Gotlib & Robinson, 1982; Libert & Lewinsohn, 1973; Youngren & Lewinsohn, 1980). In interpersonal situations depressed persons tend to communicate self-devaluation and helplessness (Biglan et al., 1985; Hokanson et al., 1980). The analysis presented earlier, however, suggests that the interpersonal behavior of depressed persons should vary according to the relational context in which it occurs. Although there has been little work specifically directed at this issue, some research has provided support for the notion that the interpersonal context affects depressive behaviors. For instance, Hinchliffe, Hooper, and Roberts (1978) found that interactions of depressed persons with their spouses contained more negative affect than their interactions with strangers. Likewise, Hale, Jansen, Bouhuys, Jemmer, and Hoofdakker (1997) found that depressed outpatients showed less "involvement" (i.e., more withdrawal or sadness) when interacting with their partners than when interacting with a stranger. Relatedly, Clark, Pataki, and Carver (1996) have reviewed evidence suggesting that people do strategically present sadness to elicit help from others, but primarily do so in communal relationships. By contrast, research by Hokanson et al. (1980) has shown that depressed persons are relatively exploitative and noncooperative when in a high-power role. A critical matter, however, is how such behaviors affect others. The social risk hypothesis predicts that mild depressive states, if effective, should act to elicit signals of support and reduce signals of threat in a per-

328

ALLEN, GILBERT, SEMEDAR

son's social environment. There are three important caveats to this. First, there is no reason to expect that depressive behaviors are effective in eliciting support and reducing threat across their entire range of severity. Indeed, we hypothesize that low levels of depressive behavior act to bring about these changes, whereas more severe levels of depressive behavior may well exacerbate the very difficulties that instigate them (i.e., low perceived social resources) through eliciting signals of rejection (Coyne, 1976; Segrin & Abramson, 1994). The second caveat is that, as noted, the effect of depressive behaviors will vary according to relational context. The third caveat is that we are chiefly interested in the behavioral responses of others. If depressive behaviors induce negative or mixed mood states in others (e.g., are annoying), but still elicit behavioral signals of support and reduce signals of threat, then the adaptive argument will be satisfied. Although research on the effects of depressive behaviors on others has tended to treat both depressive behavior and social relationships as fairly homogeneous concepts, there are a number of studies germane to the hypothesis and its caveats. Although some research has suggested that depressive behavior results in the depressed person being rejected or avoided (Coyne, 1976; Segrin & Abramson, 1994), the majority of these studies have utilized clinically depressed samples, to which our adaptive hypothesis need not apply. Other research, however, has pointed to a more functional role for depressed behaviors, particularly within communal types of relationships. For instance, Biglan et al. (1985), Hops et al., (1987) and Nelson and Beach (1990) have all shown that depressive behavior reduces the likelihood of aversive responses from family members. Clark, Ouellette, Powell, and Milberg (1987) found that expressions of sadness strongly elicited helping, but only in communal relationships. This is not surprising given research that has shown that there is a strong tendency to offer help within communal relationships (Williamson & Clark, 1992; Williamson, Clark, Pegalis, & Behan, 1996), and that affective expression is accepted more readily within these relationships (Clark & Taraban, 1991). Other research, chiefly using mildly depressed college samples, has also demonstrated that depressed behavior elicits signals of support from others. Hokanson, Lowenstein, Hedeen, and Howes (1986) examined the interactions of mildly depressed college students with their roommates over 11 weeks and showed that depressed students elicited progressively increasing care-taking behavior from their roommates. Burchill and Stiles (1988) also examined interactions between depressed college students and their roommates and found that although the mood of these dyads was worse than controls before the interaction, it actually improved after the interaction. They also found that the depressed student roommate interactions were more personally involved, with higher levels of self-disclosure. In an

11. DEPRESSION AND RMT

329

experimental study, Stephens, Hokanson, and Welker (1987) found that depressed confederates received higher levels of advice and support than did those in anxious or normal roles. It is likely that supporting and helping a depressed person is rewarding when that person shows some gratitude (recognition of debt) and apparent efficacy of the help given (i.e., signaling that the giver is competent and valued). This increases the Su of the giver. Things can all go badly, however, when help given does not seem to be helpful to the depressed person and investments given appear to fall on barren ground. In everyday language these patients are known to their general practitioners as "heart-sink" patients, and do produce a lot of rejection. It remains an important research question of how people feel when the help given is not helpful and what other processes may be involved in the depressed person's inability to experience help as helpful (e.g., they don't see the help as enough). A number of studies have also made the point that although depressed or distressed behavior does induce a mixture of positive and negative feelings in others, it does (at least in mild forms) elicit overt signals of support. Biglan, Rothlind, Hops, and Sherman (1989) and Lovejoy and Busch (1993) both showed that although distressed behavior elicited mixed affective reactions, it also elicited supportive behavioral responses. Peterson (1993) showed that in response to helpless behavior "the most frequent response was trying to help the helpless individual feel better. As the frequency of helpless behavior increased, people were less likely to try to make him or her feel better,... (they were) more likely to become angry and more likely to ignore or avoid this individual" (p. 289). Joiner, Alfano, and Metalsky (1992) found the degree of reassurance seeking engaged in by depressed persons was crucial to their likelihood of rejection by their college roommates, with those seeking greater levels of reassurance being more likely to experience rejection. These latter results are consistent with the notion that lower levels of helpless or distressed behavior may be effective in eliciting social support, whereas higher levels may actually alienate such support. It also emphasizes the importance of directing help seeking behavior to those targets who will tolerate it well (i.e., CS partners). Although clearly much more research is needed on the functions of normal depressed mood, these results do suggest that some levels of depressed behavior can be successful in reducing threat and eliciting support. CONCLUSION

Depressive behaviors, especially those associated with mild, transient depressed mood states, can operate as powerful signals to conspecifics, and have probably functioned to reduce social risks in ancestral environments. This discussion suggests that although the potential for depression proba-

330

ALLEN, GILBERT, SEMEDAR

bly evolved many millions of years ago and grew out of social defeat and challenge inhibition strategies, with successive adaptations in social living the problems of handling social conflicts have become more complex. This appears especially to be the case with respect to tailoring depressive behaviors to relational contexts such that they reduce risks. The addition of relational models theory to this conceptualization of depression makes it clear that although the overall strategy of depressive behaviors might be to minimize social risk, tactical variations will be required in order to achieve this within specific relational contexts. CS relationships are particularly important, as calls on help and support without the immediate prospect of reciprocity are generally well tolerated. As such, these relationships are probably vital to allowing social resources to be bolstered during times of threat. On the other hand, submissive behaviors may be especially important to defuse aggression from up-rank individuals within AR relationships. Indeed, a critical factor distinguishing adaptive from nonadaptive forms of depression may be the extent to which the cognitive models used to coordinate the interpretation of, and actions within, social relationships match the cognitive models utilized by interactional partners. We believe that this conceptualization has many advantages over previous adaptive/evolutionary models which have tended to treat the social environment as if depressive behaviors were only relevant to one category of interactional behavior, be it communal, authority, or reciprocity based. Furthermore, this conceptualization points not only to the way in which interpersonal tactics of depressed persons might vary across relational contexts, but also suggests novel ideas about how depressed people change the way in which they think about and perceive the social world. For instance, the tendency to recategorize relationships in order to protect social resources, and the resultant dislocation of social interactions, might lie at the heart of the dysfunctional tendency for depressed states to be exacerbated into clinical disorders in some circumstances. A dislocation between the grammar one applies to a relationship during a depressed state and the way in which one usually thinks about it (and the way it is usually thought about by the interactional partner), might explain why some depressive behaviors exacerbate, rather than ameliorate, the very problems they were designed to address. In this way relational models theory suggests a number of unique hypotheses about interactional aspects of depression that can enrich research in this area for some time to come.

REFERENCES Allan, S., & Gilbert, P. (1997). Submissive behaviour and psychopathology. British Journal of Clinical Psychology, 36, 467-488.

11. DEPRESSION AND RMT

331

Allen, N. B., & Badcock, P. B. T. (2003). The social risk hypothesis of depressed mood: Evolutionary, psychosocial, and neurobiological perspectives. Psychological Bulletin, 129, 887-913. Allen, N. B., Semedar, A., & Haslam, N. (2003). Relationship patterns associated with vulnerability to psychopathology. Manuscript submitted for publication. Barkow, J. H. (1989). Darwin, sex and status: Biological approaches to mind and culture. Toronto: University of Toronto Press. Barlow, D. (1988). Anxiety and its disorders. New York: Guilford Press. Baron-Cohen, S. (1995). Mindblindness: An essay on autism and theory of mind. Cambridge, MA: MIT Press. Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117, 497-529. Baumeister, R. F., & Tice, D. M. (1990). Anxiety and social exclusion. Journal of Social and Clinical Psychology, 9, 165-195. Beck, A. T. (1987). Cognitive models of depression. Journal of Cognitive Psychotherapy: An International Quarterly, 1, 5-38. Berry, D. S., & Hansen, J. S. (1996). Positive affect, negative affect, and social interaction. Journal of Personality and Social Psychology, 71, 796-809. Biglan, A., Hops, H., Sherman, L, Friedman, L. S., Arthur, J., & Osteen, V. (1985). Problem solving interactions of depressed women and their husbands. Behavior Therapy, 16, 431-451. Biglan, A., Rothlind, J., Hops, H., & Sherman, L. (1989). Impact of distressed and aggressive behavior. Journal of Abnormal Psychology, 98, 218-228. Boehm, C. (1999). Hierarchy in the forest: The evolution of egalitarian behavior. Cambridge, MA: Harvard University Press. Bowlby, J. (1980). Depression. Vol. 3: Attachment and loss. London: Hogarth Press. Brothers, L. (1997). Friday's footprint: How society shapes the human mind. New York: Oxford University Press. Brown, G. W., Harris, T. O., & Hepworth, C. (1995). Loss, humiliation and entrapment among women developing depression: A patient and non-patient comparison. Psychological Medicine, 25, 7-21. Burchill, S. A. L., & Stiles, W. B. (1988). Interactions of depressed college students with their roommates: Not necessarily negative. Journal of Personality and Social Psychology, 55, 410-419. Buss, D. M. (1990). The evolution of anxiety and social exclusion. Journal of Social and Clinical Psychology, 9, 196-201. Buss, D. M. (1991). Evolutionary personality psychology. In M. R. Rosenzweig & L. W. Porter (Eds.), Annual review of psychology (Vol. 42). Palo Alto, CA: Annual Reviews. Byrne, R. W. (1995). The thinking ape: Evolutionary origins of intelligence. Oxford, England: Oxford University Press. Byrne, R. W., & Whiten, A. (1988). Machiavellian intelligence: Social expertise and the evolution of intellect in monkeys, apes, and humans. Oxford, England: Oxford University Press. Clark, L. A. (2000). Mood, personality and personality disorder. In R. J. Davidson (Ed.), Anxiety, depression and emotion (pp. 171-200). New York: Oxford University Press. Clark, M. S., Ouellette, R., Powell, M., & Milberg, S. (1987). Recipient's mood, relationships type, and helping. Journal of Personality and Social Psychology, 53, 94-103. Clark, M. S., Pataki, S., & Carver, V. (1996). Some thoughts and findings on self-presentation of emotion in relationships. In G. J. 0. Fletcher & J. Fitness (Eds.), Knowledge structures in close relationships: A social psychological approach (pp. 247-274). Mahwah, NJ: Lawrence Erlbaum Associates. Clark, M. S., & Taraban, C. (1991). Reactions to and willingness to express emotion in communal and exchange relationships. Journal of Experimental Social Psychology, 27, 324-336. Clutton-Brock, T. H., & Parker, G. A. (1995). Punishment in animal societies. Nature, 373,209-216.

332

ALLEN, GILBERT, SEMEDAR

Coiro, M. J., & Gottesman, 1.1. (1996). The diathesis and/or stressor role of expressed emotion in affective illness. Clinical Psychology: Science and Practice, 3, 310-322. Cosmides, L. (1989). The logic of social exchange: Has natural selection shaped how humans reason? Studies with the Wason selection task. Cognition, 31, 187-276. Cosmides, L., Tooby, J., & Barkow, J. H. (1992). Introduction: Evolutionary psychology and conceptual generation. In J. H. Barkow, L. Cosmides, & J. Tooby (Eds.), The adapted mind: Evolutionary psychology and the generation of culture (pp. 3-15). New York: Oxford University Press. Coyne, J. C. (1976). Depression and response to others. Journal of Abnormal Psychology, 85, 186-193. Cruse, H. (1996). Neural networks as cybernetic systems. Stuttgart: Georg Thieme Verlag. Damasio, A. R. (1994). Descartes error: Emotion, reason, and the human brain. New York: Grosset/ Putnam. Davila, J., Hammen, C., Burge, D., & Paley, B. (1995). Poor interpersonal problem solving as a mechanism of stress generation in depression among adolescent women. Journal of Abnormal Psychology, 104, 592-600. Ekman, P., & Davidson, R. (1994). The nature of emotion: Fundamental questions. London: Oxford University Press. Ellis, L. (1993). Social stratification and socioeconomic inequality, Vol. 1 (A comparative biosocial analysis). Westport, CT: Greenwood. Fiske, A. P. (1991). Structures of social life: The four elementary forms of social relationship. New York: Free Press. Foa, U. G., & Foa, E. B. (1974). Societal structures of the mind. Springfield, IL: Charles C. Thomas. Ford, D. H. (1987). Humans as self-constructing living systems: A developmental perspective on behavior and personality. Hillsdale, NJ: Lawrence Erlbaum Associates. Gilbert, P. (1989). Human nature and suffering. Hillsdale, NJ: Lawrence Erlbaum Associates. Gilbert, P. (1992). Depression: The evolution ofpowerlessness. Hillsdale, NJ: Lawrence Erlbaum Associates. Gilbert, P. (1993). Defense and safety: Their function in social behaviour and psychopathology. British Journal of Clinical Psychology, 32, 131-154. Gilbert, P. (1995). Biopsychosocial approaches and evolutionary theory as aids to integration in clinical psychology and psychotherapy. Clinical Psychology and Psychotherapy, 2, 135-156. Gilbert, P. (1997). The evolution of social attractiveness and its role in shame, humiliation, guilt and therapy. British Journal of Medical Psychology, 70, 113-147. Gilbert, P. (2001). Evolutionary approaches to psychopathology: The role of natural defences. Australian and New Zealand Journal of Psychiatry, 35, 17-27. Gilbert, P., & McGuire, M. (1998). Shame, social roles and status: The psychobiological continuum from monkey to human. In P. Gilbert & B. Andrews (Eds.), Shame: Interpersonal behavior, psychopathology and culture (pp. 99-125). New York: Oxford University Press. Glantz, K., & Pearce, J. K. (1989). Exiles from Eden: Psychotherapy from an evolutionary perspective. New York: Norton. Gotlib, I. H., & Hammen, C. L. (1992). Psychological aspects of depression: Towards a cognitive interpersonal integration. Chichester, England: Wiley. Gotlib, I. H., & Robinson, L. A. (1982). Responses to depressed individuals: Discrepancies between self-report and observer rated behavior. Journal of Abnormal Psychology, 91,231-240. Grafen, A. (1987). The logic of divisively asymmetric contests: Respect for ownership and the desperado effect. Animal Behavior, 35, 462-467. Hale, W., Jansen, J., Bouhuys, A., Jemmer, J., & Hoofdakker, R. (1997). Non-verbal behavioral interactions of depressed patients with partners and strangers: The role of behavioral social support and involvement in depression persistence. Journal of Affective Disorders, 44,111-112. Hamilton, W. D. (1964). The genetic evolution of social behaviour. Part I and II. Journal of Theoretical Biology, 7, 1-52. Hinchliffe, M., Hooper, D., & Roberts, F. J. (1978). The melancholy marriage. New York: Wiley.

11. DEPRESSION AND RMT

333

Hirshleifer, D., & Rasmusen, E. (1989). Cooperation in a repeated prisoner's dilemma with ostracism. Journal of Economic Behavior and Organization, 12, 87-106. Hokanson, J. E., Loewenstein, D. A., Hedeen, C., & Howes, M. J. (1986). Dysphoric college students and roommates: A study of social behaviors over a three month period. Personality and Social Psychology Bulletin, 12, 311-324. Hokanson, J. E., Sacco, W. P., Blumberg, S. R., & Landrum, G. C. (1980). Interpersonal behavior of depressed individuals in a mixed motive game. Journal of Abnormal Psychology, 87,322-332. Hops, H., Biglan, A., Sherman, L., Arthur, J., Friedman, L, & Osteen, V. (1987). Home observations of family interactions of depressed women. Journal of Consulting and Clinical Psychology, 55, 341-346. Humphrey, N. K. (1976). The social function of intellect. In P. P. G. Bateson & R. A. Hinde (Eds.), Growing points in ethology (pp. 303-317). Cambridge, England: Cambridge University Press. Ingram, R. E., Miranda, J., & Segal, Z. V. (1998). Cognitive vulnerability to depression. New York: Guilford Press. Izard, C. E. (1991). The psychology of emotions. New York: Plenum Press. Joiner, T. E., Alfano, M. S., & Metalsky, G. I. (1992). When depression breeds contempt: Reassurance seeking, self-esteem, and rejection of depressed college students by their roommates. Journal of Abnormal Psychology, 101, 165-173. Joiner, T. E., & Coyne, J. C. (1999). The interactional nature of depression: Advances in interpersonal approaches. Washington, DC: American Psychological Association. Joiner, T. E., Coyne, J. C., & Blalock, J. (1999). On the interpersonal nature of depression: Overview and synthesis. In T. E. Joiner & J. C. Coyne (Eds.), The interactional nature of depression: Advances in interpersonal approaches. Washington, DC: American Psychological Association. Joiner, T. E., Metalsky, G. I., Katz, J., & Beach, S. (1999). Depression and excessive reassuranceseeking. Psychological Inquiry, 10, 269-278. Klein, D. N., Lewinsohn, P. M., &Seeley, J. R. (1996). Hypomanic personality traits in a community sample of adolescents. Journal of Affective Disorders, 38, 135-143. Klinger, E. (1975). Consequences and commitment to aid disengagement from incentives. Psychological Review, 82, 1-24. Leary, M. R. (1990). Responses to social exclusion: Social anxiety, jealousy, loneliness, depression, and low self esteem. Journal of Social and Clinical Psychology, 9, 221-229. Leary, M. R. (2001). Toward a conceptualization of interpersonal rejection. In M. R. Leary (Ed.), Interpersonal rejection (pp. 3-20). New York: Oxford University Press. Leary, M. R., & Downs, D. L. (1995). Interpersonal functions of the self esteem motive: The self esteem system as a sociometer. In M. H. Kernis (Ed.), Efficacy, agency, and self-esteem (pp. 123-144). New York: Plenum Press. Leary, M. R., & Kowalski, R. M. (1990). Impression management: A literature review and twocomponent model. Psychological Bulletin, 107, 34-47. Leary, M. R., Tambor, E. S., Terdal, S. K., & Downs, D. L. (1995). Self-esteem as an interpersonal monitor: The sociometer hypothesis. Journal of Personality and Social Psychology, 68,519-530. Leslie, A. M. (1994). Pretending and believing: Issues in the theory of ToMM. Cognition, 50, 211-238. Libert, J., & Lewinsohn, P. M. (1973). The concept of social skill with special reference to the behavior of depressed persons. Journal of Consulting and Clinical Psychology, 40, 304-312. Lovejoy, M. C., & Busch, L. M. (1993). Emotional and behavioral responses to aversive interpersonal behaviors. Journal of Abnormal Psychology, 102, 494-497. MacLean, P. D. (1990). The triune brain in evolution. New York: Plenum Press. Matthews, A., Ridgeway, V., & Williamson, D. A. (1996). Evidence for attention to threatening stimuli in depression. Behaviour Research and Therapy, 34, 695-705. Maynard Smith, J. (1964). Group selection and kin selection. Nature, 201, 1145-1147. Maynard Smith, J., & Price, G. R. (1973). The logic of animal conflict. Nature, 246, 15-18. Monroe, S. M., Rhode, P., Seeley, J. R., & Lewinsohn, P. M. (1999). Life events and depression: Relationship loss as a prospective risk factor for the first onset of major depressive disorder. Journal of Abnormal Psychology, 108, 606-614.

334

ALLEN, GILBERT, SEMEDAR

Nelson, G. M., & Beach, S. R. H. (1990). Sequential interaction in depression: Effects of depressive behavior on spousal aggression. Behavior Therapy, 21, 167-182. Nesse, R. M. (1990). Evolutionary explanations of emotions. Human Nature, 1, 261-289. Nesse, R. M. (2000). Is depression an adaptation? Archives of General Psychiatry, 57, 14-20. Oatley, K(1992). Best laid plans: The psychology of emotions. Cambridge, England: Cambridge University Press. Orbell, J., & Dawes, R. (1993). Social welfare, cooperators' advantage and the option of not playing the game. American Sociological Review, 58, 787-800. Panskepp, J. (1998). Affective neuroscience. New York: Oxford University Press. Peterson, C. (1993). Helpless behaviour. Behaviour Research and Therapy, 31, 289-295. Price, J. S. (1972). Genetic and phylogenetic aspects of mood variations. International Journal of Mental Health, 1, 124-144. Price, J., Sloman, L., Gardner, R., Gilbert, P., & Rhode, P. (1994). The social competition hypothesis of depression. British Journal of Psychiatry, 164, 309-315. Pusey, A. E., & Packer, C. (1997). The ecology of relationships. In J. R. Krebs & N. B. Davies (Eds.), Behavioural ecology: An evolutionary approach (pp. 254-283). Oxford: Blackwell Science. Salter, F. (1995). Emotions in command: A naturalistic study of institutional dominance. London: Oxford University Press. Schoener, T. W. (1987). A brief history of optimal foraging ecology. In A. C. Kamil, J. R. Krebs, & H. R. Pulliam (Eds.), Foraging behavior. New York: Plenum. Segrin, C., & Abramson, L. Y. (1994). Negative reactions to depressive behaviors: A communication theories analysis. Journal of Abnormal Psychology, 103, 655-668. Seligman, M. E. P. (1975). Helplessness: On depression development and death. San Francisco: Freeman. Sloman, L., & Gilbert, P. (2000). Subordination and defeat: An evolutionary approach to mood disorders and their treatment. Mahwah, NJ: Lawrence Erlbaum Associates. Stephens, R. S., Hokanson, J. E., & Welker, R. (1987). Responses to depressed interpersonal behavior: Mixed reactions in a helping role. Journal of Personality and Social Psychology, 52, 1274-1282. Toates, F. (1995). Stress: Conceptual and biological aspects. New York: Wiley. Trivers, R. (1971). The evolution of reciprocal altruism. Quarterly Review of Biology, 46, 35-57. Von Hoist, D. (1986). Vegetative and somatic components of tree shrews' behavior. Journal of the Autonomic Nervous System (suppl.), 657-670. Watson, D. (1988). Intraindividual and interindividual analyses of positive and negative affect: Their relations to health complaints, perceived stress and daily activities. Journal of Personality and Social Psychology, 54, 1020-1030. Watson, D., Clark, L. A., Mclntyre, C. W., & Hamaker, S. (1992). Affect, personality and social activity. Journal of Personality and Social Psychology, 63, 1011-1025. Weiss, J. M., & Simson, P. G. (1985). Neurochemical mechanisms underlying stress-induced depression. In T. M. Field, P. M. McCabe, & N. Schneiderman (Eds.), Stress and coping. Hillsdale, NJ: Lawrence Erlbaum Associates. Williams, G. C. (1992). Natural selection: Domains, levels, and challenges. New York: Oxford University Press. Williams, K. D., & Zadro, L. (2001). Ostracism: On being ignored, excluded, and rejected. In M. R. Leary (Ed.), Interpersonal rejection (pp. 21-53). New York: Oxford University Press. Williamson, G. M., & Clark, M. S. (1992). Impact of desired relationship type on affective reactions to choosing and being required to help. Personality and Social Psychology Bulletin, 18,10-18. Williamson, G. M., Clark, M. S., Pegalis, L. J., & Behan, A. (1996). Affective consequences of refusing to help in communal and exchange relationships. Personality and Social Psychology Bulletin, 22, 34-47. Youngren, M. A., & Lewinsohn, P. M. (1980). The functional relationship between depression and problematic behavior. Journal of Abnormal Psychology, 89, 333-341.

C H A P T E R

12 A Relational Approach to the Personality Disorders Nick Haslam

In "The awful party," cartoonist Michael Leunig shows a person looking glumly out the window of a bleak room, nursing a drink. On the bare floorboards behind him sit several large baskets, each containing a few people of distinctly off-putting appearance. Attached to each basket is a sign: "Too hard," "Too pathetic," "Too boring," "Too strange," and so on. Through the window, the night sky looks sullen and starless. Many of the themes that run through the study of the personality disorders (PDs) are contained in this little tableau. Like the wicker-bound cartoon figures, people with PDs are classified into diagnostic groups and labeled accordingly. What sets them apart from the ordinary run of humanity is usually understood to be an excess of some disposition, too much of what might otherwise make them normal. At the same time, there is a lingering suspicion that this classification is somehow arbitrary and questionable, a concern that it may be subjective and biased. After all, who is this pigeonholing misanthrope to judge what is abnormal? Finally, and most importantly for my purposes, the grounds on which PD is assigned are primarily forms of interpersonal deviance, ways in which people are socially problematic. Disordered individuals are those with whom we are prone to have troubled interactions and relationships. In this chapter I work through some of the implications of these themes for the study of PDs and propose an account of some of them based on the relational models (RMs). I argue that researchers and theorists have frequently failed to recognize the intrinsically relational and interpersonal nature of the PDs, and that the exceptions have some significant limitations 335

336

HASLAM

themselves. In particular, I suggest that for all their merits the currently dominant strands of interpersonal research on the PDs proceed from flawed and restricted understandings of sociality which leave a theoretical space for a complementary relational account. I briefly lay out such an account in general terms and in relation to a few exemplary disorders, describe the interpersonal processes that would, in theory, generate particular forms of relational disturbance in each disorder, and present a preliminary study that tests some of the predictions of the account. In closing, I discuss a few connected issues raised by this analysis, specifically the possible links between the RMs and normal personality variation and the unique relational context of psychotherapy as an arena for interpersonal conflict and transference.

PERSONALITY DISORDERS The PDs were defined as a distinct domain of psychopathology in 1980 when DSM-III assigned them to a separate diagnostic "axis," apart from the very large majority of recognized mental disorders. Unlike these disorders, the PDs were conceptualized as enduring and inflexible aspects of character rather than conditions that are in some sense extraneous to the personality. The diagnosis of a PD on "axis H" is therefore intended to provide a characterological context for any primary "axis I" disorder that is present, and it will usually be taken by clinicians to be a complicating factor in treatment. Although the classification of PDs remains controversial, 10 recognized disorders have survived three revisions of DSM since the inception of axis II, some of them with a long history in psychiatric thought, often associated with familiar concepts (e.g., histrionic and narcissistic PDs) or axis I disorders (e.g., obsessive-compulsive, paranoid and schizoid PDs), and some of more recent vintage (e.g., avoidant, borderline, and dependent PDs). It is widely accepted that PDs are primarily disorders of social being. Interpersonal features figure prominently in clinical descriptions of many PDs, including the diagnostic criteria specified in ASM-TV. Dependent personalities go "to excessive lengths to obtain nurturance and support from others," narcissists are "interpersonally exploitative," borderlines have "a pattern of unstable and intense interpersonal relationships," schizotypals "have a lack of close friends or confidants," and avoidant personalities are "preoccupied with being criticized or rejected in social situations" (APA, 1994). Indeed, every DSM-IV PD has diagnostic criteria that refer to interpersonal disturbance. If PDs are pre-eminently interpersonal in their manifestations, one would not know it from many of the prevailing theoretical accounts. A useful collection of papers outlining alternative theories of PD (Clarkin & Lenzen-

12. PERSONALITY DISORDERS

337

weger, 1996) is a case in point. Neurobiological theories (Depue, 1996) emphasize disturbances in intrapersonal processes such as affect regulation and behavioral inhibition, anchored in neurochemical abnormalities. Psychodynamic theories (Kernberg, 1996) emphasize intrapsychic conflicts and deficits. Although the inflexible schemas and strategies that cognitive accounts of PD invoke often have some interpersonal content (Pretzer & Beck, 1996), they tend to focus on beliefs about the self that are at best only incidentally social. Borderline PD, for example, is discussed by neurobiological theorists in terms of impulsivity and emotional lability mediated by defective modulation of neurotransmitter systems; by psychoanalytic theorists in terms of primitive ego defenses, weakened superego structure, and developmental fixation at the separation-individuation stage; and by cognitive theorists in terms of poor emotion regulation and stress tolerance. Although any of these abnormalities might indeed tend to produce interpersonal difficulties, they are all at a substantial explanatory distance from them. None of the explanatory variables invoked by these theories are intrinsically social or capable of accounting in a direct manner for the sorts of interpersonal and relational characteristics listed in the PDs' diagnostic criteria. Instead, these characteristics are understood to be incidental phenomena that are causally downstream of the true etiological influences, interpersonal derivatives of more basic aberrations. INTERPERSONAL ANALYSES OF THE PDS Researchers and theorists of an interpersonal bent have taken issue with this understanding of the PDs, and have attempted to account for their interpersonal features in correspondingly and irreducibly interpersonal terms. Several distinct interpersonal approaches to the PDs have arisen, and the relational account developed later in the chapter is intended as an addition to their ranks. Although there is no need for a new approach to place itself in a competitive stance toward established ones, it is important to survey some of these approaches to assess their limitations. An account based on the RMs can justify itself by showing how it might fill some of the gaps and resolve some of the problems that these alternatives present. By doing so, such an account can establish its credentials for offering complementary insights. With this rationale in mind, I turn to brief analyses of prevailing approaches to the interpersonal study of PDs. The Interpersonal Circle

Chief among such approaches is the interpersonal circle (1C) or circumplex, a model of interpersonal traits and behaviors that traces its history to Harry Stack Sullivan and Timothy Leary. This model proposes that interper-

338

HASLAM

sonal behaviors, and dispositions to perform them, can be located around the perimeter of a circle defined on bipolar axes of Control (dominance vs. submissiveness) and Affiliation (warmth vs. coldness). The circle is normally divided into octants representing the four poles and four intermediate blends. Such is the model's popularity and taken-for-grantedness that research on the circle is sometimes virtually identified with the interpersonal approach (Pincus & Ansell, in press), and the "interpersonalness" of traits is gauged by the extent to which they project a distinctive profile onto it (Gurtman, 1991). Several research groups have sought to establish circumplex locations for the PDs, which would capture their distinctive interpersonal signatures (Sim & Romney, 1990; Soldz, Budman, Demby, & Merry, 1993; Wiggins & Pincus, 1989). The results, reviewed by Widiger and Hagemoser (1997), are generally quite consistent, with 7 of the 10 PDs recognized in the DSM-IV having more or less robust locations distributed around the circle. Clearly the circle successfully captures something of the distinctive flavor of many disorders. Even so, several problems accompany the 1C account of PDs. First, it fails to supply robust locations for several disorders—obsessive-compulsive, schizotypal, and borderline PDs—that clearly have a distinctive interpersonal style. Obsessive-compulsive PD, for instance, is marked by stubbornness, rigidity, competitiveness, a preoccupation with interpersonal control and status, a preference for work over informal social activities, and a deferential attitude to respected authority figures. If the circle cannot represent the interpersonal signature of this disorder and others, it surely reflects a limitation of the model rather than a conclusive demonstration that the disorder is purely /n/rapersonal. A second problem with the 1C account is that some of the PDs that do have demonstrated locations share these locations with other disorders that seem to be interpersonally distinct. For example, both avoidant and schizoid PDs are linked to the cold submissiveness octant, reflecting their shared tendencies for social withdrawal and apparent aloofness, but they are clearly differentiable by the urgent but conflicted desire of the former for close relationships to which the latter is indifferent. Similarly, narcissistic and paranoid PDs share a disengaged hostility that places them within the cold dominance octant, but differ in that the former is marked by an arrogant need for admiration and an entitled sense of being above reciprocal obligations, whereas the latter is marked by suspiciousness and a preference for autonomy and calculative reason in human affairs. Although it is true that the diagnostic boundaries between some PDs are indistinct and comorbidity between them is rampant, the failure of the 1C to represent these basic differences indicates its limited capacity to capture the finer texture of interpersonal style.

12. PERSONALITY DISORDERS

339

A more basic, third problem with the 1C account of PDs is that it offers a map rather than an explanation or theory. The circle fits the interpersonal characteristics of each PD into a spatial schema, but this schema says nothing about the psychological structures, processes, motives, or mechanisms that give rise to distinct forms of interpersonal deviancy. Several quite distinct explanatory accounts, employing a variety of such concepts, might be able to make sense of a person's position on the circle. For example, submissive tendencies might result from strong dependency needs, weak dominance striving, a strategic attunement to social status, or a tendency to attribute hostile intentions to authority figures. According to the 1C, however, PDs are represented simply as distinctive patterns of interpersonal behavior, whose sources are left unspecified. Such a formal redescription can be an important step toward theoretical understanding, but it is not itself a theory. A final critique of the 1C account of PDs identifies something that is less a flaw than a limitation of scope. As it is a fundamental issue that motivates the account of PDs that I develop later it merits a somewhat extended discussion here. My basic argument is that the cognitive, motivational, and behavioral processes that underpin interpersonal relationships can be divided into (at least) two distinct levels of analysis. On one level, interpersonal phenomena can be analyzed in an individualist fashion, with a focus on the behavior of individuals and on cognition that takes individual attributes—traits and other personal characteristics of self or other—as its objects. On another level, interpersonal phenomena have a distinctly relational aspect, which requires a focus on how relationships between people proceed, understood in terms of the emergent properties of dyads or groups, and on cognition and motivation about such properties. An individualist analysis of an interaction might point to how one interactant behaved in a dominant way and the other in a submissive way, and speculate on how each cognizes the other in terms of personal attributes such as gender or youth or hostile intentions. A relational analysis, in contrast, might characterize the interaction in terms of the asymmetry of the implied relationship and speculate about how the interactants cognitively construe their shared relationship in terms of its level of inequality, closeness, durability, and so on, properties that inherently apply to relations between people rather than to the individuals themselves. It should now be plain to see that the 1C account of PDs embodies a thoroughly individualist approach, representing disorders as ways in which deviant individuals behave and saying little or nothing about how their relationships, or their ways of thinking about them, are aberrant. The circle can be used to make sense of relationships by studying patterned links between the circumplexes of two interactants, and useful analyses of interpersonal

340

HASLAM

complementarity and other relational laws have been developed in this manner (Kiesler, 1983; Wiggins, 1982). The interpersonal disturbance associated with PDs might in principle be explained at this level, with each disorder characterized by a particular deviation from normal patterns of interpersonal complementarity. However, 1C accounts of PDs have not taken this path, and research on circumplex-based laws of relational complementarity and the like has sometimes found them lacking in any event (Haslam, 1994b; Orford, 1986). None of this takes anything away from the value of the 1C model as an account of PDs: People diagnosed with PDs clearly manifest abnormalities in their personal dispositions in ways that the circle picks up. However it does indicate that the 1C model can only be one part of the interpersonal understanding of PDs. People with PDs may conduct, construe, and have motives for their relationships that are abnormal in ways that are not reducible to the sort of interpersonal dispositions captured by the 1C. We might therefore ask whether new understandings of the interpersonal disturbances that characterize the PDs might come from seeking the relational aberrations that are associated with them. This chapter argues that RMs theory offers a solid backing to such an effort. In addition to elaborating a theory of how relationships might be represented and construed, the theory makes a strong case for the importance of a distinctly relational level of social cognition. A large body of research generated by the theory has found evidence that cognition at the relational level structures a variety of social-cognitive phenomena at least as strongly as a range of individual-level variables (see Fiske & Haslam, 1996, for a review). RMs theory promotes an understanding of social relations as a field that must be apprehended within the terms of its own emergent principles and grammars, as well as offering a critique of the bias toward individualist analyses within psychology. Its implications for the study of PDs are discussed at greater length later on, but for now the important thing to note is that interpersonal accounts of PDs that take an individualist approach, such as the 1C approach, are restricted in their focus.

STRUCTURAL ANALYSIS OF SOCIAL BEHAVIOR Another prominent interpersonal account of personality disorders has been developed by Lorna Smith Benjamin (1996). Her structural analysis of social behavior (SASB) model superficially resembles the 1C in proposing circular structures as the descriptive foundations for its analysis. However it also deviates in several important ways. First, the circumplex it employs differs fundamentally from the 1C in its vertical dimension. Instead of the latter's dominance versus submission, it contrasts enmeshed or interdependent ways of relating—which include those based on dominance and

12. PERSONALITY DISORDERS

341

submission—with independent or autonomous ways. Notably, this dimension is more relational in its focus, given that interdependence, unlike dominance or submission, is intrinsically a condition of dyads or groups rather than individuals, and as a result the SASB model is better equipped to capture dyadic phenomena than the 1C. Second, the SASB model includes three distinct circumplexes, referring to transitive or "parent-like" behavior toward others, intransitive or "child-like" behavior in response to others, and "introjective" behavior directed toward the self. These circumplexes are also linked by various formal rules or "predictive principles," such as complementarity (e.g., transitive "protect" linked to intransitive "trust"). Third, the SASB characterization of PDs is not purely descriptive, but embeds them in a theoretical context of developmental histories, defense mechanisms, affects, and self- and other-representations that owes a great deal to psychoanalytic and attachment theory. The result is a set of interpersonal formulations of the PDs that is unparalleled in its richness. The SASB model offers an influential account of the interpersonal patterns associated with PDs. It has also proven its usefulness as an instrument for characterizing the fine detail of interpersonal communication, as in psychotherapy process research. However, like the 1C it too has limitations as an account of the interpersonal bases of PDs. First, many of the SASB formulations of particular PDs are baroque in their complexity, often positing multiple locations on each of the three circumplexes. Second, although the model affords textured formulations, it does so at the expense of parsimony, employing eight descriptive terms on each of the three circumplexes as well as numerous predictive principles. It might also be added that the SASB model appears to violate its proposed circumplex structure (Pincus, Gurtman, & Ruiz, 1998). Third, the model's implicit commitment to a psychodynamic approach—in which concepts such as introjection and distinct parental and child forms of behavior sit comfortably—limits its appeal to many researchers and practitioners. Finally, despite accommodating dyadic phenomena better than the 1C, the SASB model still does not fully address the relational level of analysis. That is, SASB formulations describe how people behave actively (e.g., blaming) and reactively (e.g., sulking) toward others and toward themselves (e.g., self-blaming), but not how they enact or think about relationships in distinctly relational or dyadic terms. Its focus remains on individual behavioral tendencies, and to capture relational phenomena two circumplexes are required, one for each interactant (e.g., A blames and B sulks). Although the enmeshment axis is implicitly relational, its pole on each of the circumplexes is still thoroughly monadic in the manner of the 1C, labeled Control on the transitive circumplex and Submit on the intransitive. In short, formulating a person or a PD in SASB terms remains primarily an exercise in mapping their characteristic behavioral dispositions in individualist

342

HASLAM

terms. This is not intended as a criticism, as there is ample evidence that PDs can be fruitfully understood in such terms, but it does point to a restriction of scope that a truly relational account of PDs might complement. Other Interpersonal Approaches Although the 1C and the SASB are the preeminent interpersonal approaches to the study of PDs, several other approaches have also been developed. These accounts—based on attachment style, the "interpersonal octagon," and traits of dependency and autonomy—each merit a brief review. The first of these approaches derives from Bowlby's work on attachment in infancy and childhood. Research applying classifications of attachment style to adult relationships has recently undergone a massive acceleration, and whole handbooks have been assembled to encompass the quantities of work that have been generated (Cassidy & Shaver, 1999). Clearly attachment theory has touched a nerve and come to occupy a conceptual space that was previously vacant. Although very little research has addressed PDs from this perspective (Fonagy et al., 1996), it would seem to show promise partly because it offers a clinically relevant classification that in the hands of some exponents incorporates a truly relational focus. Attachment styles such as secure, dismissive, and preoccupied are proposed to refer not only to characteristic forms of interpersonal behavior and to images of self and (generalized) other, but also to "working models" of relationships. Despite these virtues, the attachment style approach to PDs has several limitations. First, its focus is exclusively on close relationships to which the attachment concept applies, whereas an interpersonal formulation of PDs would ideally address problems arising in relationships of all sorts. Interpersonal difficulties are by no means confined to close relationships, and PDs would be expected to manifest themselves elsewhere, such as in dealings with acquaintances or colleagues at work. Second, despite the relational focus implied by the working models concept, attachment styles appear to be substantially reducible to an individualist formulation involving simple negative versus positive images of self and other (Bartholomew & Horowitz, 1991). Third, although the attachment styles have often been conceptualized as discrete categories or types, in contrast to the assumed continuity of the 1C and SASB models, research evidence (Fraley & Waller, 1998) indicates that they too reflect dimensional continua. Finally, given that people with PDs are unlikely to present with a secure attachment style, the remaining two or three insecure variants seem insufficient to encompass in a differentiated way the many forms of PD. The second less prominent interpersonal approach to the study of PDs is the "interpersonal octagon" formulation developed by Birtchnell (1996).

12. PERSONALITY DISORDERS

343

This schema has many superficial similarities to the 1C, proposing a plane whose two axes resemble the familiar affiliation and control dimensions and which is cut into eight segments, but Birtchnell is at pains to emphasize its distinctiveness. For example, octagon positions are interpreted in terms of interpersonal competencies and objectives rather than behavioral dispositions or preferences, the two axes are not assumed to be bipolar, and two separate octagons are generally employed to embody a distinction between positive and negative relating. A study by Birtchnell and Shine (2000) found that all PDs were associated significantly with at least one point on the negative octagon, although some of their projections onto the octagon were relatively nonspecific and redundant. Although its degree of resemblance to the 1C can be disputed, the interpersonal octagon has most of the same limitations in view of its primarily descriptive character and its individualist focus. In short, although early evidence suggests that the octagon helps to characterize the interpersonal style of several PDs, the model has familiar limitations of scope. A third relatively undeveloped interpersonal approach to the PDs can arguably be found in research and theory on the personality dimensions of sociotropy/dependency and autonomy. These dimensions, the former labeled sociotropy by Beck (1983) and dependency by Blatt (Blatt & Zuroff, 1992), refer to interpersonal dispositions, preoccupations, and motives surrounding nurture and approval on the one hand, and self-definition, independence, and individual striving on the other. These dimensions have generated a considerable amount of research, most of it in relation to depression. Research relating them to PDs is sparse (Clark, Steer, Haslam, Beck, & Brown, 1997) and suggests that they do not capture distinctive characteristics of most of the PDs. Besides being too broad-brush to differentiate the PDs, the dimensions clearly fail to embody a relational approach or to offer any sort of explanatory account of underlying processes, structures, and mechanisms.

A RELATIONAL THEORY OF PDS

This overview of interpersonal approaches to the PDs deliberately emphasizes some of their limitations. To summarize, established interpersonal accounts generally fail to address the distinctly relational dimension of interpersonal life, the ways in which it cannot be understood purely in terms of individuals' behavioral dispositions and their thinking about individuals' attributes. They do not consistently afford an explanation of the structures and processes that give rise to PDs, instead usually offering ways of redescribing PD symptomatology according to an organized schema. Some of these schemas also appear to lack empirical support as structural mod-

344

HASLAM

els. In addition, interpersonal accounts of PDs sometimes fail to discriminate between disorders that appear to have distinctive interpersonal characteristics, they are sometimes unparsimonious, and they sometimes rest on theoretical assumptions that lack broad appeal. An account of the PDs based on Relational Models (RMs) theory, first sketched by Fiske (1991) and subsequently developed further (Fiske & Haslam, 1997; Haslam, 1997), can perhaps escape some of these pitfalls and supply a complementary analysis to those reviewed earlier. This account starts from a simple argument about social coordination. For social relations to be well-coordinated, people who participate in them must share basic assumptions and rules about how interactions and relationships should proceed. Such coordinating assumptions and rules might be loosely described as one or more "grammars." If participants bring different assumptions and rules to an interaction—as often happens in intercultural encounters and in social settings with ambiguous demands—this discrepancy is likely to produce bilaterally violated expectations, disrupted interchanges, and possibly confusion and upset. Furthermore, if for whatever reason some people are unusually prone to bring discrepant assumptions and rules to bear on their interactions, they will harvest an unusually large quantity of such unpleasant interpersonal consequences. These consequences will tend to take characteristic forms if such people consistently bring a particular set of assumptions and rules to bear when their interactional partners expect a different set. Fleshing out this argument, RMs theory proposes a set of distinct relational grammars that people employ in their social lives, distinct sets of rules and parameters that govern how relationships are conducted, construed, motivated, and evaluated (see chap. 1, this volume, for a review of RMs theory). For social relations to be coordinated, participants should operate according to a common RM, and specifying what sorts of relationships and interactions are appropriately organized in terms of particular models is a significant part of shared culture. As these models are categorically distinct—a theoretical postulate backed up by empirical research (Haslam, 1994a, 1999)—it is always possible that interactants may organize their interactions according to a discrepant set of rules. Doing so should produce cognitive, behavioral, and emotional disturbances as well as moral recriminations, as a variety of studies suggests (Fiske & Tetlock, 1997; Tetlock, Kristel, Elson, Green, & Lerner, 2000). If certain people were apt to employ a particular RM more intensely or in a more extensive range of relationships than their culture or social milieu specified as appropriate, they would tend to experience more than their shares of such discord. (The same could be said of people who employ a model to a lesser extent than their milieu, although for simplicity I will focus here on overimplementation.) By the account that I am developing

12. PERSONALITY DISORDERS

345

here, PDs are attributed to many of these people. That is, at least some PDs represent dispositional tendencies to employ particular RMs to socioculturally inappropriate degrees or in socioculturally inappropriate ways. The traits that characterize PDs therefore represent attributions in a reifying shorthand of distinctive patterns of interpersonal disturbance that stem from particular cognitive or motivational aberrations. Dispositions such as manipulativeness and dependency are not so much attributes residing within individuals as perceived regularities in particular kinds of disrupted interactions, these disruptions explicable in terms of one interactant's relational abnormalities. Imagine, for instance, someone who approaches relationships with a strong and enduring disposition to implement the Authority Ranking (AR) model more and the Equality Matching (EM) model less than others. He is likely to fall foul of his interaction partners in predictable ways. By introducing elements of ranking, asymmetry, and hierarchy into relationships in which different relational understandings are culturally prescribed, and by failing to feel bound by reciprocal egalitarian obligations, he is likely to come across as grandiose, pretentious, arrogant, exploitative, entitled, and lacking in reciprocal interest. All of these poisonous descriptors can be found in the DSM-IV description of narcissistic PD. The same traits might well be attributed to someone who makes a single relational faux pas by his aggrieved interactant. That is, the trait attributions are primarily features of a particular kind of relational discrepancy, which may indicate a PD if the person is chronically liable to discrepancies of that kind. It is also worth noting that by this account the narcissist may not have difficulties in relationships that are by cultural consensus organized along AR lines; PDs should primarily reveal themselves in relational contexts in which their overemployed RM is inappropriate. Similar linkages can be drawn to many other PDs, I would argue, proposing specific forms of over- or underimplementation of particular RMs. Formulations of each PD are presented shortly. Each form of aberrant use of the RMs should be associated with somewhat distinctive forms of relational discrepancy and disturbance, and if these aberrations are chronic these disturbances can be represented by distinctive pathological traits. Each of the four RM-specific aberrations is presented in turn below. Someone who employs the Communal Sharing (CS) model when interactants expect a different set of relational assumptions and rules will tend to be perceived in particular ways that reflect the distinctive elements of the model. The model constitutes relationships as communal and close, categorically separating an undifferentiated in-group from an out-group and using need as a distributive principle. When someone mis-applies it—that is, applies it in a relational context in which it should not apply by implicit consensus—others may find the person to be inappropriately demanding, intru-

346

HASLAM

sive, needy, dependent, unwilling to do their fair (egalitarian) share, naive about social give and take, overly exclusionary in their dealings with third parties, and presuming an unwarranted intimacy. Attributions such as these might be made in any particular relationship in which there is a discrepancy of this kind, and they should be widespread and repetitive if they reflect the sort of enduring relational aberration that may underlie a PD. Someone who implements the Authority Ranking (AR) model in an aberrant fashion should incur a different pattern of social perceptions and trait attributions. The model constitutes relationships as ranked and asymmetrical. When this model is inappropriately applied in a relationship, other participants are apt to feel that the person is overly focused on status or position and too concerned with protocol. Their attributions may vary according to whether the person adopts a superior or inferior position with respect to them. In the former case the person may be seen as inappropriately domineering and arrogant, in the latter as inappropriately deferential, timid, stubborn, conforming, ingratiating, or passive. Someone who inappropriately implements the Equality Matching (EM) model by definition construes a relationship as egalitarian and governed by balanced reciprocity where others construe it differently. As a result, the person might be perceived as rigid and inflexible in their insistence on strict balance, tight or ungenerous for their unwillingness to offer aid without the expectation of reciprocation, and insubordinate or impudent for their leveling failure to heed social rank and status. Finally, those who overemploy the Market Pricing (MP) models risk having their tendency to understand relationships in terms of instrumental or calculative rationality perceived in a similarly moralized fashion. They are apt to be seen as cold, callous, Machiavellian, overly individualistic, fickle and lacking in basic social trust, in view of their preoccupation with whether relationships are yielding an adequate rate of return on their investment of time and effort. The present account is primarily a theory of relational discrepancies and of how dispositions might give rise to them in patterned ways, and it does not specify where these tendencies to employ particular models in culturally aberrant ways might originate. One strong cognitive form of explanation (Fiske, 1991) would argue that some people might simply lack a particular model, just as autistic individuals are claimed to lack a modular theory of mind (Baron-Cohen, 1997). A weaker cognitive explanation would simply propose that a particular RM is unusually salient or chronically activated, so that individuals with PDs are especially likely to construe and cognize relationships in terms of that model and to be sensitive to elements of it. Alternatively, people with PDs might simply have stronger motives to engage in particular kinds of relationship, so that they might seek to conduct relationships according to a favored model beyond the social domains in which

12. PERSONALITY DISORDERS

347

it is culturally prescribed. Whether the ultimate origins of such aberrant cognitive and motivational tendencies are to be sought in genetic, social learning, or psychodynamic factors is left open in the relational account, whose emphasis is on proximate structures and processes. Similarly, this account takes no position on whether the relational aberrations it proposes are the fundamental explanatory underpinnings of the PDs or whether they are instead simply intermediate-level explanations of the PDs' interpersonal aspects that are ultimately explained in other terms. Despite leaving etiological issues unexplored, I would argue that this relational account has several virtues compared to the interpersonal approaches outlined earlier, quite apart from its distinctively relational level of analysis. First, it is based on a structural theory that has solid empirical support, the defining characteristics, number, interrelations, and categorical status of the RMs having been consistently validated in many studies (e.g., Haslam, 1994a, 1999; Haslam & Fiske, 1999; Vodosek, 2000). Although the same can be said for the 1C, evidence is rather weaker for the structural validity of the SASB model and for the attachment styles. Second, the account is quite parsimonious, invoking only four basic structures. Even if aberrations in the employment of these structures are subdivided according to their cognitive (i.e., implementation, construal) and motivational bases, the number of explanatory variables equals the differentiation of the 1C and is much more economical than the three-circumplex SASB. Third, and most importantly, this account does not simply link PDs to RMs in a descriptive fashion, but offers a theory of discrepant relational implementations of particular psychological structures that explains why such a link might exist. That is, the relational account tells an explanatory story about the processes and structures involved in relational problems and how recurring patterns of such discrepancy-based problems constitute PDs. The 1C, by contrast, serves primarily as a schema for the interpersonal characterization of the PDs. PROPOSED RELATIONAL ABERRATIONS FOR SPECIFIC PDS These theoretical virtues of the relational account are abstract, and ultimately the validation of any theory of PDs rests on whether its formulation of them is plausible and empirically confirmed. In this section, before presenting a preliminary empirical test of this account, I present brief formulations of the DSM-IV PDs based on the discrepancy-based analysis laid out earlier. No formulation is offered for schizotypal PD, whose undoubted interpersonal deviancy I see as being best understood as the outcome of schizophrenia diathesis whose neural basis is not fundamentally interpersonal (see Meehl, 1990).

348

HASLAM

Antisocial PD People diagnosed with antisocial PD are noted for deceit, manipulation, aggression, irresponsibility, cynicism, and disregard for the rights of others. Fiske (1991) argues that they ("sociopaths") operate according to none of the models, simulating them for selfish ends but actually operating in a purely instrumental Asocial mode. To the extent that their relational aberration can be understood in terms of the four models, however, I hypothesize that their characteristic coldness, callousness, lack of empathy, and indifference to others' feelings indicate a somewhat specific lack of motivation for and cognitive implementation of CS relationships. These deficits suggest a lack of the basic emotional bonds and the sense of deeply rooted shared identity on which close communal relationships depend, and might be expected to be less disruptive of relationships operating under different models. Someone with these deficits would be less likely to violate the relational expectations of their interaction partners if these other models were prescriptively in force. Tentatively, then, antisocial PD should be associated with low levels of CS motivation and implementation. Avoidant PD

The hypersensitivity to negative evaluation and fears of disapproval characteristic of avoidant personalities suggest that the vertical dimension of status, hierarchy, and rank are unusually salient to them. Viewing the social world through an AR lens, in which one's standing is paramount, would lead them to see interactions in which the other person presumes a communal or egalitarian bond as tests of comparative worth or position, or as asymmetrical arrangements into which they must please or be judged by a superior. In addition, descriptions of avoidant personalities as lonely, isolated, and longing for intimate relationships which they only enter into when uncritical acceptance is assured, imply strong but thwarted desires for close CS relationships. In sum, avoidant PD should be associated with high levels ofAR and CS. Borderline PD

Borderline personalities are described as being extremely concerned about separation in close relationships, showing a preoccupation with the other person's degree of care and constancy. They tend to fall precipitously into intense intimate relationships marked by an expectation of total fusion and mutuality and an idealization that may suddenly shift into its opposite, a phenomenon often discussed as "splitting." All of these features suggest an unusually strong motivation for CS relationships and a tendency to cognitively implement them. Merged identity, intimacy, and categorical all-

12. PERSONALITY DISORDERS

349

or-nothing thinking is characteristic of this model, as are themes of inclusion, need, and nurture. People with such a strong and inflexible motives and cognitive tendencies would be expected to experience relational discrepancies of the sort typical of borderline PD. They would tend to implement a CS definition of relationships when their interactants do not, making interpersonal demands for care and intimacy of an unconditional and boundary-denying kind which the other believes to be excessive and premature. Needless to say, the other's predictable response of withdrawal or refusal to operate according to CS rules of engagement will be perceived as terrible betrayals by the person who is following these rules; hence, perhaps, the sudden souring of the borderline's close relationships and her repetitive and desperate pattern of relational disappointments. In sum, borderline PD should be associated with high levels of CS motivation and cognitive implementation. Dependent PD Dependent PD has several characteristic interpersonal features. These include a strong need to be taken care of, protected, and reassured; clinging and submissive behavior apparently designed to elicit nurture; fears of separation and abandonment; a need for others to assume responsibility; a willingness to tolerate imbalanced relationships to keep a valued bond; and a propensity to form quick and indiscriminate attachments following loss or separation. In RMs terms, these features might be understood as a combination of tendencies to desire and implement CS and AR relationships. The former is evident in the strong desire for close attachments and preoccupation with exclusion from or loss of them. The dependent person seeks a somewhat child-like communal merger in which care is given in an unqualified and need-based manner. At the same time, a tendency to construe the interpersonal world in terms of ranked asymmetries and a desire to embody subordinate roles within them is clearly apparent. According to RMs theory, such a tendency need not indicate an unwilling capitulation to a dominating power, but may be an intrinsically motivated urge that gives the person access to the benefits of the other's largesse and protection. In sum, dependent PD should be associated with high levels of CS and AR motivation and cognitive implementation. Histrionic PD The relational characteristics of histrionic PD appear to involve the CS model, although this link is perhaps less clearly evident than it is for some other PDs. Histrionic individuals appear to be strongly motivated to seek out close communal relationships, manifested in tendencies to enact greater intimacy in a relationship than others think appropriate, to engage

350

HASLAM

in romantic fantasy, and to act in a way that people find inappropriately seductive. The first and last of these are very readily interpreted in terms of the relational discrepancies discussed earlier in the chapter, representing disjunctions between what is culturally prescribed or normative and what the histrionic implements. The fact that the manifestations of this PD primarily focus on romantic and sexual relations rather than communal relationships in general, and the fact that the seductiveness is often disavowed by the histrionic, suggests that this analysis requires further qualification Nevertheless, histrionic PD should be associated with high levels ofCS motivation and cognitive implementation. Narcissistic PD

According to DSM-IV narcissistic personalities are characterized by grandiosity, a need for admiration, boastfulness, fantasies of unlimited power and success, and a belief in being superior or special and understandable only by high status others. With little difficulty these characteristics can be understood as evidence of a chronic tendency to over-use the AR model, to construe the social milieu in ranked and status-ridden terms and to desire, imagine, and project a place for oneself in the higher echelons. Other descriptions—the narcissist's sense of entitlement, lack of reciprocal interest, enviousness, and snobbishness—might be taken to indicate a tendency to employ the EM model less than is socially expected. The relational discrepancy in this case will occur in social relationships where egalitarian expectations are in force. The narcissist will tend to violate these by implementing a different model, probably AR, and thereby evoke charges such as those embodied in the descriptions. In sum, narcissistic PD should be associated with high levels of AR and low levels of EM. Obsessive-Compulsive PD

Several features of obsessive-compulsive PD are amenable to a relational analysis. People diagnosed with this PD are described as being preoccupied with interpersonal control, didactic, rigidly deferential to authority, insistent on literal compliance, stubborn or resistant when they do not respect another's authority, and attentive to their relative status. All of these features plainly point to an unusually strong and inflexible construal or relationships in AR terms in contexts where others either do not recognize the aptness of this model or implement it in a less rigid fashion. A case might also be made for an MP component, in view of obsessive-compulsive PD's characteristic competitiveness and devotion to work and productivity, but these might more parsimoniously be taken as preferred ways of establishing and safeguarding rank and status. In sum, obsessive-compulsive PD should be associated with high levels of AR.

12. PERSONALITY DISORDERS

351

Paranoid PD

Paranoid personalities are often described as being motivated by control, unwilling to tolerate perceived domination by others (i.e., stubbornness), arrogant, and prone to fantasies of power and rank. These features suggest a heightened tendency to construe social relations in AR terms where others find these terms to be misplaced, and correspondingly strong AR motives. Paranoid personalities are described as distrustful and overly concerned with being cheated and exploited, which imply a chronic implementation of the calculative MP model. Finally, they are often presented as cold, apparently lacking in tender feelings, preoccupied with personal autonomy and averse to close, intimate relationships, all of which are consistent with a lower than usual desire for or cognitive implementation of the CS model. In sum, paranoid PD should be associated with high levels ofAR and MP and low levels of CS. Schizoid PD

The thoroughgoing detachment, isolation, and aloofness of schizoid PD might appear to indicate a general deficit in social motivation, and the lack of social skill and poise suggests a broad social impairment. However, the schizoid's lack of desire for intimacy, close friendships, and sexual relations imply a more specific detachment from or aversion to CS relationships. That is, these attributions would tend to be made about someone who fails to seek CS relations where others expect them to. In sum, schizoid PD should be associated with low levels of CS.

AN EMPIRICAL STUDY

The theoretical account and formulations of individual PDs developed here have to date only been subject to one empirical study. In it (Haslam, Reichert, & Fiske, 2002), participants were recruited by newspaper advertisements soliciting "individuals with difficulties relating to others in their work or personal lives." Unlike the standard practice of sampling individuals with PDs from clinical settings, we hoped to collect a sample from the general community that might be more representative than clinical samples of actually occurring PD, and less confounded with the coexisting axis 1 disorders (e.g., depression, anxiety, schizophrenia) than normally bring people with PDs into treatment. Needless to say, this sampling policy is only appropriate if the severity of PD symptomatology reaches clinical levels in the obtained community sample, which it did to a very substantial extent. The 57 participants recruited in this manner completed a series of selfreport instruments, including a measure of PD symptomatology (the Per-

352

HASLAM

sonality Disorder Questionnaire-4; Hyler, 1994). To assess individual differences in the use of the RMs, participants completed two instruments. In one, used to assess model-specific relational difficulties and motivations, participants read paragraphs describing relationships of each type and rated the degree to which they found relationships of each type to be problematic as well as the strength of their motivation for or investment in such relationships. In the other instrument, the Modes of Relationship Questionnaire (MORQ; Haslam & Fiske, 1999), participants listed their personal acquaintances, representatively sampled 10 of them, and rated the 10 on a series of items assessing the degree to which they construed or cognitively implemented particular RMs in their relationship with each person. Indices were then derived to assess each participant's broad tendencies to implement each model across the sample of their relationships. For comparison purposes, the IC-based Inventory of Interpersonal Problems (IIP; Alden, Wiggins, & Pincus, 1990; Horowitz, Rosenberg, Baer, Ureno, & Villasenor, 1988), which yields scores on the eight circle octants, was also administered. The study's hypotheses were complicated, and followed from the relational formulation of the respective PDs presented in the previous section. As a simplifying step, it was assumed that patterns of abnormal motivation and implementation of a particular model covary, so that hypotheses predicted the same associations for both. Again, no predictions were made about schizotypal PD. The hypotheses are briefly summarized next: • high CS motivation/implementation: avoidant, borderline, dependent and histrionic PDs • low CS motivation/implementation: antisocial, paranoid and schizoid PDs • high AR motivation/implementation: dependent, narcissistic, obsessive-compulsive and paranoid PDs • low EM motivation/implementation: narcissistic PD • high MP motivation/implementation: paranoid PD The obtained associations between the RMs measures and the PD scales are summarized in Table 12.1. The table shows that many of the predicted associations were found, and that most of the PDs had distinctive relational patterns. Several conclusions can be drawn prior to describing individual patterns. First, every hypothesis was supported for at least one of the relevant measures—motivation or implementation or both—indicating that the hypotheses were broadly accurate. Second, in 10 of the 13 cases in which hypothesized associations were not significant, they were in the predicted direction, suggesting that some additional hypotheses would be supported

12. PERSONALITY DISORDERS

353

TABLE 12.1 Association Between RM Measures (M = motivation, I = implementation) and PD

CS Personality Disorder

M

Antisocial Avoidant Borderline Dependent Histrionic Narcissistic Obsessive-compulsive Paranoid Schizoid

4+ + + +

M

M

M

•f ft + t

• 4-

MP

EM

AR

-Jv

ft

ft

^

+

ft

4-

1^ U U

U t

'f hypothesized positive association ft unhypothesized positive association •^ hypothesized negative association U unhypothesized negative association • nonsignificant hypothesized association

given greater statistical power (i.e., a larger sample, more reliable measures, or both). (The three exceptions were all essentially zero correlations rather than contrary effects.) However, some of these null effects might equally reflect real disjunctions between relational implementation and motivation, such that one but not both is associated with a particular PD. Thi might in turn support a more differentiated and textured analysis of the relational aberration associated with some PDs. For instance, avoidant PD may truly be characterized by strong motives for close communal relationships without any accompanying tendency to construe relationships in these terms, and with a tendency to construe relationships in asymmetric or ranked terms without a correspondingly strong desire for such relationships. Indeed, this portrait of conflicted sociality seems quite plausible. Similarly, schizoid PD's link to low levels of CS motivation but not cognitive implementation may truly reflect the fact that it is primarily a disorder of absent desire. A third conclusion that can be drawn from Table 12.1 is that although seven significant but unhypothesized associations were obtained, five of these can be readily explained by two straightforward findings. Borderline PD was associated with nonspecifically strong relational motivations, and EM implementation was low for obsessive-compulsive and paranoid PDs because AR implementation was high for them, AR and EM implementations having a strong negative association (r = -0.78). As a fourth and final conclusion, it is clear that just as in other interpersonal approaches, the

354

HASLAM

RMs do not appear to discriminate some PDs with apparently distinctive interpersonal signatures. For example, according to the study antisocial and schizoid PDs are more or less relationally equivalent, contrary to appearances. In short, the relational account presented here probably stands in need of complementary accounts to distinguish PDs such as these. Integrating the findings presented in Table 12.1 with the exploratory findings on relational difficulties yields the following thumbnail sketches: AvoidantPD was associated with difficulties in asymmetrical and egalitarian relationships, with a relatively strong motivation for communal relationships, and with a tendency to construe relationships in terms of status and hierarchy. Borderline PD was strongly and generally associated with relational difficulties, especially in egalitarian relationships, and also with high investment in relationships of all kinds. Its one distinctive association was with the tendency to construe relationships in asymmetrical terms. Dependent PD was associated with high levels of nonspecific relational difficulties, but especially difficulties in close communal relationships. It was also associated with high motivation for such relationships; such excessive communal motivation may cause interpersonal conflict. Histrionic PD was linked to difficulties in authority- and equality-based relationships and to relatively strong motivation for communal relationships. Narcissistic PD was not strongly associated with self-reported relational difficulties, but was linked to strong motivation for authority relationships and a tendency to construe relationships in these terms. Narcissists also tended to construe relationships in egalitarian terms less than others. Obsessive-compulsive PD was nonspecifically associated with relational difficulties, and was linked to the relative underimplementation of egalitarian relationships and overimplementation of authority relationships. Paranoid PD was associated with high motivation for "rational" instrumental relationships, the tendency to construe relationships in terms of dominance and rank, and the tendency not to construe them in communal or egalitarian terms. Schizoid PD was associated with low investment in close communal relationships, with the tendency not to construe relationships in egalitarian terms, and with the absence of self-reported relational difficulties. In addition to these distinctive relational patterns associated with the respective PDs, a few other patterns emerged in the findings that suggest interesting broader generalizations about PD. First, two clear associations

12. PERSONALITY DISORDERS

355

emerged between the DSM's PD clusters and the RMs measures, with odd cluster symptomatology strongly associated with relatively strong motivation for MP relationships and anxious cluster symptomatology with strong CS motivation. One interpretation of this finding is that odd cluster PDs are associated with a strong preference for or investment in more uninvolving and autonomous relationships, perhaps grounded in a defensive withdrawal from closer and more emotionally engaging forms of relationship. Second, PD symptomatology as a whole was associated with the tendency to construe and implement the AR model in participants' personal relationships, signaling perhaps that this represents a somewhat nonspecific relational aberration in personality pathology that is not captured by the 1C. Psychological disturbance may involve exaggerated concerns with dominance and submission, superiority and inferiority, and power and prestige, all of which bear on the AR model. An important additional point to note about the study's findings is that they supported the substantial independence of the RMs characterization of individual differences and the 1C characterization, a vital step toward demonstrating that a relational account can complement more familiar interpersonal approaches. Significant associations between the RMs measures and the IIP were evident, not surprisingly, but they were relatively sparse and weak. In particular, measures relating to the EM and AR models were essentially unrelated to the IIP scales, suggesting broad interpersonal domains which the 1C does not adequately capture. The significant correlations between the RMs and 1C measures were readily intelligible. For example, participants who tend to construe relationships in CS terms more than others tended to have problems with self-assertion and being exploited and to have few problems with excessive vindictiveness, and those with strong CS motivation tended to be excessively nurturant. Model-specific relational difficulties, however, were largely independent of the 1C, adding to a picture of substantial nonredundancy between the two frameworks. The correlations between the RMs measures with the PD symptom scales were comparable in strength to those shown by the 1C measure, and several of these correlations captured interpersonal components of the PDs that are not captured by the 1C. For example, it was noted earlier that 1C analyses typically fail to distinguish avoidant and schizoid PDs on the one hand, and narcissistic and paranoid PDs on the other. As expected, the former pair were sharply distinguished in relational terms, with avoidant PD marked by high levels of CS motivation and schizoid PD by low levels. Similarly, although the AR model is highly salient for both narcissistic and paranoid PDs, these disorders are distinguished by the latter's relatively strong investment in MP relationships and its corresponding underimplementation of close CS relationships. In addition, obsessive-compulsive and borderline PDs, both of which normally fail to yield reliable 1C projections,

356

HASLAM

revealed somewhat distinctive relational patterns. Borderline PD also had the strongest association with self-reported relational difficulties, implying that the IC's failure to capture it is a significant omission. Needless to say, one study cannot establish the validity of a new relational account of PDs. This study in particular suffers from the use of a single measure of PDs, the symptomatology of which is notoriously difficult to assess. In addition, the relatively low statistical power of the study, given the probable unreliability of some measures, probably limited its ability to uncover associations between the RMs variables and the PD scales. Finally, a replication with clinical samples would provide an important test of the generality of the study's findings. Nevertheless, these findings suggest strongly that the relational account has promise as a theoretically parsimonious complement to existing accounts. Related Issues

This chapter has focused on a relational account of the PDs, but the general approach that I have sketched has other implications for related issues. Two examples spring readily to mind. First, there are no obvious reasons why aberrant motivations for and implementations of particular RMs could not underlie normal as well as abnormal personality variation, offering a novel account of interpersonal dispositions. Second, discrepant implementations of RMs in dyads, whether due to the trait-like aberrations of a personality disordered individual or not, would seem to offer an interesting perspective on psychotherapy and on what is often referred to in psychotherapeutic circles as transference. Normal Personality

As noted at the beginning of this chapter, PDs are commonly understood as relatively extreme personality variants that are on quantitative continua with normal variation. Although there is some evidence against this assumption of continuity in the cases of schizotypal, antisocial and borderline PDs (Haslam, 2003), it is probably sound for many others. If this is the case, then milder personality characteristics that fall below diagnostic thresholds may be amenable to the sort of relational account presented earlier, with their relational aberrations presumed to be correspondingly milder. Although there has been little empirical research on this question to date (Caralis & Haslam, 2002), this would seem to represent a worthwhile avenue to explore. For one thing, it may be useful to determine whether personality dispositions other than those on an apparent continuum with PDs but with a clear interpersonal dimension are amenable to a relational analysis. For example, we might expect warmth to be associated with relatively strong tendencies

12. PERSONALITY DISORDERS

357

to implement CS, conformity with AR, cooperativeness with EM, and Machiavellianism with MP. Although these associations have the dull ring of obviousness about them, they might be valuable nonetheless. First, demonstrating them would ground the traits in an account of the cognitive and motivational processes and structures that underpin them, the absence of which is a common criticism of trait psychology. Second, efforts to test these associations would allow us to determine the extent to which interpersonal traits are rooted in relational motives and schemas. Multivariate analyses might allow relational and other contributors to interpersonal traits to be disentangled. For instance, if Machiavellianism were found to be associated with MP implementation, this relational variable might simply mediate other, more basic influences on it such as selfishness or autonomy, or its influence on Machiavellianism might be fundamental and mediated by these other, more derivative variables. Finding links between implementation of RMs and self-evidently interpersonal traits would not be especially surprising, even if it yielded the sorts of theoretical benefits laid out before. More interesting, perhaps, would be demonstrations that traits whose interpersonal aspects are unrecognized or regarded as incidental are associated with characteristic relational patterns. If such an association were found it might not only modify our understanding of the trait but also expand the interpersonal domain. One possibility raised by Haslam et al.'s (2002) study is that Neuroticism, a pre-eminent personality factors in contemporary trait psychology that is normally conceptualized in strictly intrapersonal terms (i.e., emotional liability or proneness to negative affect), may be associated with a tendency to construe relationships in AR terms. This tendency was fairly nonspecifically associated with the PDs, as is Neuroticism, and was strongly associated with self-reported relational distress. This hypothesized link between Neuroticism and AR implementation was supported by Caralis and Haslam (2002), and suggests that Neuroticism should be understood in part as a sensitivity to or preoccupation with rank and status, a conclusion that might be congenial to evolutionary psychologists and followers of Alfred Adler and Erich Fromm, psychoanalytic pioneers who did not scant the interpersonal realm. Similar revelations might be in store for other personality dimensions whose interpersonal aspects are equally obscure. Psychotherapy

It is a truism that psychotherapy is an unusually fraught kind of relationship: At least one party is identifiably distressed or damaged, the emotional temperature is often high and, according to some schools of thought, the nature of the relationship itself calls forth transferential distortions and misunderstandings. The relational analysis developed here may offer an

358

HASLAM

original viewpoint on these matters. It suggests that the psychotherapeutic relationship is unusually complex, and that difficulties in it may be understood in terms of the relational discrepancies discussed earlier in regard to PDs. Furthermore, these difficulties are apt to reveal the client's relational aberrations, and the complexity of the psychotherapeutic relationship usefully encourages these revelations. Few relationships are likely to be entirely "pure" in RM terms, governed exclusively by one model, but psychotherapy would appear to be unusually complex. Indeed, a case can be made that every model is implicitly woven into the relationship. The intimacy and personal disclosure of the therapeutic relationship are typically associated with CS relationships such as those that exist within families and romantic pairs. An AR dimension is implied because the therapist is presumed to command superior knowledge and skills for understanding and treating psychological suffering, and to have the authority to direct the treatment according to professional judgment. The EM presence is perhaps weakest, but therapy is often framed, to clients and fellow therapists alike, as a collaborative exercise governed by an implicitly or explicitly egalitarian "working alliance." Finally, an MP element to the therapeutic relationship is evident in the fact that money is changing hands to procure the service and in the concern that therapeutic progress be as speedy and efficient as possible, a concern heightened by developments in managed care and the proliferation of outcome research. Psychotherapy is an unusually complex relational field.1 Whenever there is complexity of this sort, there is also likely to be ambiguity. Coordinated social relationships require that people hold shared understandings of how relationships of a certain kind are to be organized, and where these understandings are multifaceted—and also in flux as therapeutic fashions change—many opportunities arise for jarring misunderstandings based on discrepant expectations. This same point has been made by Fiske and Tetlock (1997), who argue that coordination is apt to be most difficult in modern societies where the understandings that govern relationships are relatively loose, changing, and provisional. We should not be at all surprised that clients and therapists experience an unusually high level of relational problems within the 50-minute hour, during which the rules in place may be unclear, complicated, and differently un-

'This complexity of the psychotherapeutic relationship is reflected in the ways in which clinicians refer to those whom they treat. "Patient" carries an AR connotation of someone acted upon by a quasi-medical authority. "Client" implies a more egalitarian, collaborative relationship with joint responsibility for therapeutic progress. "Consumer" evokes a full-blown MP orientation in which the treated are understood to have selected a provider in the mental health marketplace and are entitled to efficient service delivery. The continuing co-existence of these terms in clinicians' discourse attests to the ambiguity of the therapeutic relationship.

12. PERSONALITY DISORDERS

359

derstood by the two parties. If a client construes her side of the relationship primarily in CS terms and her therapist does not, predictable sorts of disjunction will occur. The factors contributing to such disjunctions may be numerous. Some can, of course, be ascribed to the client. A large fraction of what brings people into therapy is interpersonal problems (Horowitz & Vitkus, 1986), and their propensity for interpersonal difficulties is likely to follow them into the consulting room. To the degree that such propensities follow the schema presented earlier in this chapter, they may take predictable forms. For instance, people who are predisposed to overimplement AR may tend to understand the therapeutic relationship in ways that disagree with their therapists' less asymmetrical expectations in ways that engender frustration and impasse on both sides. Where the clients may see their role as being compliant recipients of authoritative directions, therapists may see a show of resistant passivity and disengagement, and clients may in turn see their therapists' apparent unwillingness to tell them how to overcome their problems as a product of incompetence or disinterest. Similar inferences can be made about other forms of aberrant implementation of particular models, each of which probably leads to its own characteristic pattern of misapprehensions. By implication, each of the PDs should have a predictable pattern of "transferential" difficulties, difficulties that represent traitlike tendencies to implement relationships in particular ways rather than re-evocations of key individuals from the person's early history as is often assumed. In cases such as the one described in the previous paragraph, where a dispositional tendency of the client is projected into the therapeutic relationship, the therapist is warranted in making potentially important "transferential" inferences about the client's interpersonal difficulties. The very ambiguity of the therapeutic relationship, its incorporation of so many distinct relational elements, helps to make the relationship a particularly sensitive instrument for deriving these inferences, affording many distinct ways in which trait-like relational abnormalities can be revealed. The classical psychoanalysts strove to maximize the sensitivity of this instrument by systematically stripping from the relationship all traces of ordinariness, but by my argument it is instead the multiple affordances of the therapeutic relationship, its having elements of all of the RMs in ambiguous proportions, that gives it its diagnostic power. However, inferences about the client's interpersonal style cannot be straightforwardly drawn whenever there is a mismatch between the expectations of therapist and client. Some mismatches may be quite unsystematic and some may spring from the client's general tendency to have poorly calibrated relational understandings, rather than understandings that are consistently and inflexibly channeled along particular relational lines. More

360

HASLAM

interestingly, perhaps, some mismatches might be attributed at least in part to therapists' misapprehensions of the therapeutic relationship. Some of these may be directly akin to those of the client (i.e., countertransference). Others might be due to therapists failing to recognize particular dimensions of the therapeutic relationship. Therapeutic orientations commonly represent the therapeutic relationship in a partial fashion that fails to acknowledge or disavows some of its elements, and their proponents may therefore make inappropriate inferences about clients who act upon some of this unacknowledged relational complexity. For example, treatments of a cognitive-behavioral sort tend to operate in a didactic or advisory AR mode more than others, and therapists of these persuasions might consequently be more prone to misunderstand or fail to recognize clients' behavior that stems from a different relational grammar. Similarly, modalities that foreground different elements of the therapeutic relationship may be prone to misunderstand a client's AR-based expectations about some aspect of the relationship as a distortion of the relationship that manifests the client's pathology. An analysis of relational discrepancies might help to clarify misunderstandings of this sort, and other complications of the therapeutic relationship.

CONCLUSION

Investigation of individual differences from the perspective of RM theory is in its early stages, both in regard to PDs and to personality more broadly. However, it is original enough in its basic assumptions about interpersonal phenomena, based strongly enough on a body of social psychological research, and sufficiently complementary to existing approaches in its focus, to warrant further development. The PDs are notoriously elusive and complicated phenomena, and no theory can hope to offer a total account. However, the relational theory presented in this chapter seems likely to expand the reach and richness of interpersonal accounts. The potential for this same discrepancy-based theory to illuminate normal personality variation and relational discrepancies arising in the clinic and elsewhere, may be a significant added benefit.

REFERENCES Alden, L. E., Wiggins, J. S., &Pincus, A. L. (1990). Construction of circumplex scales for the Inventory of Interpersonal Problems. Journal of Personality Assessment, 55, 521-536. American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author.

12. PERSONALITY DISORDERS

361

Baron-Cohen, S. (1997). Mindblindness: An essay on autism and theory of mind. Cambridge, MA: MIT Press. Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles among young adults: A test of a four-category model. Journal of Personality and Social Psychology, 61, 226-244. Beck, A. T. (1983). Cognitive therapy of depression: New perspectives. In P. J. Clayton & J. E. Barrett (Eds.), Treatment of depression: Old controversies and new perspectives (pp. 265-290). New York: Raven. Benjamin, L. S. (1996). Interpersonal diagnosis and treatment of personality disorders (2nd ed.). New York: Guilford. Birtchnell, J. (1996). How humans relate: A new interpersonal theory. Hove, England: Psychology Press. Birtchnell, J., & Shine, J. (2000). Personality disorders and the interpersonal octagon. British Journal of Medical Psychology, 73, 433-448. Blatt, S. J., &Zuroff, D. (1992). Interpersonal relatedness and self-definition: Two prototypes for depression. Clinical Psychology Review, 12, 527-562. Caralis, D., & Haslam, N. (2002). Personality correlates of the relational models. Manuscript submitted for publication. Cassidy, J., & Shaver, P. R. (Eds.). (1999). Handbook of attachment: Theory, research, and clinical applications. New York: Guilford. Clark, D. A., Steer, R. A., Haslam, N., Beck, A. T., & Brown, G. K. (1997). Personality vulnerability, psychiatric diagnoses, and symptoms: Cluster analyses of the Sociotropy-Autonomy subscales. Cognitive Therapy and Research, 21, 267-283. Clarkin, J. F., & Lenzenweger, M. F. (Eds.). (1996). Major theories of personality disorder. New York: Guilford. Depue, R. (1996). A neurobiological framework for the structure of personality and emotion: Implications for personality disorders. In J. F. Clarkin & M. F. Lenzenweger (Eds.), Major theories of personality disorder (pp. 347-390). New York: Guilford. Fiske, A. P. (1991). Structures of social life: The four elementary forms of human relations. New York: Free Press. Fiske, A. P., & Haslam, N. (1996). Social cognition is thinking about relationships. Current Directions in Psychological Science, 5, 143-148. Fiske, A. P., & Haslam, N. (1997). Prerequisites for satisfactory relationships. In L. Meyer, M. Grenot-Scheyer, B. Harry, H.-S. Park, & I. Schwartz (Eds.), Making friends: Understanding the social worlds of children and youth. Baltimore: Paul H. Brooks. Fiske, A. P., & Tetlock, P. E. (1997). Taboo tradeoffs: Reactions to transactions that transgress spheres of exchange. Political Psychology, 18, 255-297. Fonagy, P., Leigh, T., Steele, M., Steele, H, Kennedy, R., Mattoon, G., Target, M., & Gerber, A. (1996). The relationship of attachment status, psychiatric classification, and response to psychotherapy. Journal of Consulting and Clinical Psychology, 64, 22-31. Fraley, R. C., & Waller, N. G. (1998). Adult attachment patterns: A test of the typological model. In J. A. Simpson & W. S. Rholes (Eds.), Attachment theory and close relationships (pp. 77-114). New York: Guilford. Gurtman, M. B. (1991). Evaluating the interpersonalness of personality scales. Personality and Social Psychology Bulletin, 17, 670-677. Haslam, N. (1994a). Categories of social relationship. Cognition, 53, 59-90. Haslam, N. (1994b). Mental representation of social relationships: Dimensions, laws, or categories? Journal of Personality and Social Psychology, 67, 575-584. Haslam, N. (1997). Personality disorders as social categories. Transcultural Psychiatry, 34, 473-479. Haslam, N. (1999). Taxometric and related methods in relationships research. Personal Relationships, 6, 519-534.

362

HASLAM

Haslam, N. (2003). The dimensional view of personality disorders: A review of the taxometric evidence. Clinical Psychology Review, 23, 75-93. Haslam, N., & Fiske, A. P. (1999). Relational models theory: A confirmatory factor analysis. Personal Relationships, 6, 241-250. Haslam, N., Reichert, T., & Fiske, A. P. (2002). Aberrant social relations in the personality disorders. Psychology and Psychotherapy: Theory, Research and Practice, 75, 19-31. Horowitz, L. M., Rosenberg, S. E., Baer, B. A., Ureno, G., & Villasenor, V. S. (1988). The Inventory of Interpersonal Problems: Psychometric properties and clinical applications. Journal of Consulting and Clinical Psychology, 56, 885-892. Horowitz, L. M., & Vitkus, J. (1986). The interpersonal basis of psychiatric symptoms. Clinical Psychology Review, 6, 443-469. Hyler, S. E. (1994). PDQ-4 manual. New York: NiJo Publications. Kernberg, 0. F. (1996). A psychoanalytic theory of personality disorders. In J. F. Clarkin & M. F. Lenzenweger (Eds.), Major theories of personality disorder (pp. 106-140). New York: Guilford. Kiesler, D. J. (1983). The 1982 interpersonal circle: A taxonomy for complementarity in human transactions. Psychological Review, 90, 185-214. Meehl, P. E. (1990). Toward an integrated theory of schizotaxia, schizotypy, and schizophrenia. Journal of Personality Disorders, 4, 1-99. Orford, J. (1986). The rules of interpersonal complementarity: Does hostility beget hostility and dominance, submission? Psychological Review, 93, 365-377. Pincus, A. L., & Ansell, E. B. (in press). Interpersonal foundations for an integrative theory of personality. In T. Millon & M. J. Lerner (Eds.), Comprehensive handbook of psychology, Vol. 5: Personality and social. New York: Wiley. Pincus, A. L., Gurtman, M. B., & Ruiz, M. A. (1998). Structural Analysis of Social Behavior (SASB): Circumplex analyses and structural relations with the Interpersonal Circle and the fivefactor model of personality. Journal of Personality and Social Psychology, 74, 1629-1645. Pretzer, J. L., & Beck, A. T. (1996). A cognitive theory of personality disorders. In J. F. Clarkin & M. F. Lenzenweger (Eds.), Major theories of personality disorder (pp. 36-105). New York: Guilford. Sim, J. P., & Romney, D. M. (1990). The relationship between a circumplex model of interpersonal behaviors and personality disorders. Journal of Personality Disorders, 4, 329-341. Soldz, S., Budman, S., Demby, A., & Merry, J. (1993). Representation of personality disorders in circumplex and Five-Factor space: Explorations with a clinical sample. Psychological Assessment, 5, 41-52. Tetlock, P. E., Kristel, O. V., Elson, S. E., Green, M. C., & Lerner, J. S. (2000). The psychology of the unthinkable: Taboo trade-offs, forbidden base rates, and heretical counterfactuals. Journal of Personality and Social Psychology, 78, 853-870. Vodosek, M. (2000, August). Relational models and the outcomes of work groups: A confirmatory factor analysis. Paper presented at the Academy of Management Meetings, Toronto, Canada. Widiger, T. A., & Hagemoser, S. (1997). Personality disorders and the interpersonal circumplex. In R. Plutchik & H. R. Conte (Eds.), Circumplex models of personality and emotions (pp. 299-325). Washington, DC: American Psychological Association. Wiggins, J. S. (1982). Circumplex models of interpersonal behavior in clinical psychology. In P. C. Kendall & J. N. Butcher (Eds.), Handbook of research methods in clinical psychology (pp. 183-221). New York: Wiley. Wiggins, J. S., & Pincus, A. L. (1989). Conceptions of personality disorders and dimensions of personality. Psychological Assessment, 1, 305-316.

Author Index

A Abrams, R., 106, 136 Abramson, L., 314, 323, 328, 334 Adams, J., 211-212, 219, 236, 242 Agha, A., 100, 130, 136 Alden, L, 352, 360 Alfano, M., 329, 333 Alfieri, T., 193, 195 Allan, S., 311, 330 Allen, N., 45, 54, 313, 326-327, 331 Allport, G., 265, 283 Almagor, U., 87, 136 Altorki, S., 75, 136 Andersen, B., 110, 136 Anderson, B., 4, 24 Ansell, E., 338, 362 Anthony, A., 87, 145 Ardner, S., 110, 136 Asch, S., 263, 283 Asquith, P., 182, 196 Austin, J., 129, 136 Avis, J., 152, 162

B Backett, C., 185, 195 Badcock, P., 313, 326, 331

Baer, B., 352, 362 Baker, W., 265, 284 Baldwin, J., 84, 94, 136 Baldwin, M., 36, 54, 150, 162 Balkin, D, 287, 305 Banks, C., 82, 136 Bardi, A., 266, 285 Barkow, J., 229, 242, 311, 314, 331, 332 Barlow, D., 313, 331 Barnea, M., 268, 283 Barnett, W., 110, 136 Baron-Cohen, S., 148, 158, 162, 321, 331, 346, 361 Barth, F., 74, 76, 86, 136 Bartholomew, K., 342, 361 Bateson, G., 130, 136 Batson, D., 264, 283 Baumeister, R., 319, 331 Baxter, P., 87, 136 Beach, S., 323, 328, 333, 334 Beck, A., 310, 331, 337, 343, 361, 362 Beck, J., 89-90, 136 Behan, A., 328, 334 Beidelman, T., 76, 80, 136 Beker, W. E., 265, 284 Belkin, A., 259, 261 Bern, D., 269, 283 Bern, S., 291, 304

363

364 Benjamin, L., 340, 361 Bennett, N., 15, 24 Bergson, H., 71, 114, 136 Berlin, I., 261 Berry, D., 310, 331 Berscheid, E., 214, 220, 228, 242, 289, 304 Bianchi, G., 268, 280 Bies, R., 223, 236, 242 Biglan, A., 310, 327-329, 331, 333 Birtchnell, J., 342-343, 361 Blalock, J., 309, 333 Blatt, S., 343, 361 Blau, P., 236, 242 Blumberg, S., 310, 333 Blurton Jones, N., 92, 140 Bobbit, P., 252, 261 Boehm, C., 235-236, 240-241, 242, 312, 331 Boehnke, K., 268, 285 Bogdan, R., 158, 160, 162 Bohannan, P., 107-108, 136 Bolton, H., 106, 136 Boster, F., 289, 306 Bouhuys, A., 327, 332 Bourdieu, P., 71, 136 Bourne, A., 151, 163 Bowes, J., 171, 174-177, 180, 182, 184-185, 193-194, 195 Bowlby, J., 310, 312, 326, 331 Boyd, R., 13, 24, 67, 143 Bozionelos, N., 99, 142 Brain, R., 80, 136 Bravo, G., 86, 89-90, 136 Brewer, M., 287, 306 Bromiley, P., 287, 305 Brothers, L., 314, 331 Brown, G. K., 343, 361 Brown, G. W., 310, 314, 331 Brown, L., 77, 136 Brown, P., 82, 84, 100, 137 Brown, R., 100, 137 Bruner, J., 72, 137 Buckner, R., 72, 144 Budman, S., 338, 362 Burchill, S., 328, 331 Burgess, R., 289, 305 Burton, R., 87, 137 Busch, L., 329, 333 Buss, D., 274, 283, 315, 319, 331 Butler, J., 292, 305 Buttram, R., 38, 55 Byrne, R., 152, 162, 314, 317, 331

AUTHOR INDEX

C Caillouet, L., 292, 305 Calabrese, G., 252, 261 Call, J., 154, 163 Campbell, K., 116, 145 Campos, J., 230, 243 Cangemi, J., 292, 305 Capps, H., 292, 305 Caralis, D., 46, 54, 356-357, 361 Garment, D., 99, 140 Carr, L., 93, 137 Carter, C., 88, 137 Carver, V., 327, 331 Cassidy, J., 342, 361 Cayley, W., 268, 284 Charmley, J., 260, 261 Chase, J., 106, 137 Chen, M., 182, 195 Cheney, D., 155, 162 Chrisman, K., 171, 195 Clark, D., 343, 361 Clark, L., 309-310, 331, 334 Clark, M., 32, 34, 56, 171-172, 184, 195, 216, 219, 327-328, 331, 334 Clarkin, J., 336, 361 Clutton-Brock, T., 316, 331 Codere, H., 114, 137 Coiro, M., 310, 332 Collis, K., 116, 145 Connelley, D., 42-43, 54, 201, 213, 219 Connerton, P., 84, 137 Cooper, R., 99, 140 Corcoran, R., 158, 162 Cosmides, L., 14, 25, 229, 243, 314, 332 Cox, T., 197, 220 Coyne, J., 309, 328, 332, 333 Crawley, A., 74-75, 137 Cropanzano, R., 221, 243 Cruse, H., 320, 324, 332 Cummings, L., 287, 305 Cvetkovitch, G., 49, 54

D Dalton, G., 112, 137 Damasio, A, 229, 242, 309, 332 D'Andrade, R., 18, 119, 137 Dannenmaier, W., 99, 137 Barley, J., 264, 283

365

AUTHOR INDEX Darwin, C., 83, 91, 98, 137, 230, 242 Dasgupta, P., 292, 305 Davidson, R., 324, 332 Davila, J., 310, 332 Davis, C., 84, 138 Davis, J., 75, 137 Dawes, R., 316, 334 Deacon, T., 116, 122, 137 de Bono, M., 88, 137 de Botton, A., 147, 162 Decety, J., 93, 142 de Dampierre, E., 79, 137 Delaney, S., 168, 174, 180, 195 Demby, A., 338, 362 Depue, R., 337, 361 Dettwyler, K., 75, 137 Deutsch, M., 31, 57, 212-213, 216, 219, 287, 289, 305 DeVries, A., 88, 137 de Waal, F., 91-92, 111, 137, 155, 162, 221, 242 Dias, S., 278-279, 284 Didow, M., 84, 138 Donald, M., 123, 138 Douglas, M., 78, 81, 108, 138 Downs, D., 318-319, 333 Dubeau, M., 93, 137 Dugatin, L, 14, 24, 91, 138 Dulaney, S., 12, 24, 86-87, 138 Dumont, L., 82, 138 Dunbar, R., 91, 138 Dunn, J., 178, 195 Duranti, A., 130, 138 Durkheim, E., 18, 71, 74, 81, 84-85, 87, 111, 115, 134, 138, 198, 219

E Earle, T., 49, 54 Earley, P., 51, 54, 213, 219 Eckerman, C., 84, 138 Eibl-Eibesfeldt, L, 83, 138 Einzig, P., 112, 138 Eisenhardt, K., 287, 305 Eisenstadt, S., 80, 138 Ekman, P., 324, 332 Eliason, B., 268, 284 Ellis, L., 316, 332 Elson, S., 41-42, 57, 249, 262, 344, 362 Empson, S., 116, 138 Evans-Pritchard, E., 69, 79, 108, 138

F Fajens, J., 75, 138 Fama, E., 287, 305 Faulkes, C., 15, 24 Feather, N., 265, 268, 283 Fildes, V., 75, 138 Firth, R., 96, 108, 110, 130, 138, 139 Fiske, A., 4, 9-13, 17, 19, 21, 24, 25, 27-38, 41-42, 44, 46, 48-49, 51, 53, 54, 55, 62, 66-68, 78, 84, 86-87, 93-94, 108, 119-120, 122, 125, 129, 138, 139, 142, 150-154, 162, 167-171, 173-174, 177-179, 181-188, 190, 192-193, 195, 198-200, 203, 208, 212, 214, 216, 219, 221, 223-224, 229, 232-233, 237, 240, 242, 243, 248, 260, 261, 263, 268, 270, 274, 276-277, 280, 282, 283, 288, 293, 305, 312, 332, 340, 344, 346-348, 351-352, 358, 361, 362 Fiske, S., 34, 55, 247, 293, 305 Flanagan, C., 194, 195 Fletcher, G., 160, 162 Foa, E., 29, 31-32, 34, 55, 315, 318, 332 Foa, U., 29, 31-32, 34, 55, 315, 318, 332 Fogassi, L., 93, 143 Folger, R., 38, 55, 221-222, 226, 236, 243 Fonagy, P., 342, 361 Ford, D., 320, 324, 332 Forsyth, D., 49, 53, 55 Fortes, M., 73-74, 78, 139 Fraley, R., 342, 362 Frank, R., 229, 243 Fraser, S., 269, 284 Frazer, J., 70, 95, 139 Freedman, J., 269, 283 Freud, S., 78, 80, 139 Frith, C., 158, 162 Fruth, B., 92, 140

G Gabarro, J., 287, 292, 305 Gallagher, H., 160, 162 Gallese, V., 93, 139, 143 Gallistel, C., 15, 25, 115, 124, 139 Gansmann, H., 114, 139 Gardner, R., 311, 334 Garvin, P., 97, 100, 139 Gee, J., 226, 243

366 Gelfand, M., 50-51, 57 Getz, L, 88, 137 Gewirth, A., 232, 243 Gibbon, J., 15, 25 Gibson, C, 213, 219 Gibson, J., 63, 139 Giffin, K., 292, 305 Gilbert, P., 45-46, 54, 311, 313, 315, 317-318, 322, 330, 332, 334 Gillison, G., 78, 139 Gilman, A, 100, 137 Gil-White, F., 234, 239, 243 Gingrich, B., 88, 146 Girard, R., 86, 139 Glantz, K., 312, 326, 332 Goffman, E., 67, 78, 81, 83, 97, 99, 103, 139 Goldstein, W., 83, 144 Gomez-Mejia, L., 287, 305 Good, D., 292, 305 Goodenough, W., 130, 139 Goodnow, J., 38-39, 55, 168-171, 173-177, 179-180, 184-187, 189-190, 192, 195, 196 Goodrick, E, 213, 219 Gotlib, I., 309, 327, 332 Gottesman, I., 310, 332 Grafen, A., 325, 332 Green, M., 41-42, 57, 249, 262, 344, 362 Greer, G., 89-90, 139 Grierson, P., 112, 139 Grunert, S., 268, 284 Gurin, B., 213, 220 Gurin, G., 213, 220 Gurtman, M., 338, 341, 361, 362 Gutierrez, C., 77, 140

H Hagemoser, S., 338, 362 Haidt, J., 231-232, 243, 278-279, 281, 284 Hale, W., 327, 332 Haley, J., 130, 136 Hallahan, M., 151, 163 Hamaker, S., 310, 334 Hamilton, D., 40, 56 Hamilton, W., 314, 332 Hammen, C., 309, 332 Hanley, S., 99, 140 Hansen, E., 91, 140

AUTHOR INDEX Hansen, J., 310, 331 Harcourt, A., 155, 162 Harlow, H., 91, 140 Harlow, M., 91, 140 Harris, 0., 109-110, 140 Harris, P., 152, 162 Harris, T., 310, 331 Hart, K., 292, 305 Hartland, E., 73-75, 80, 140 Haslam, N., 12, 14, 19, 21, 25, 29-36, 44, 46-48, 51, 54, 55, 86-87, 139, 150, 153, 155, 162, 188, 196, 268, 284, 293, 305, 327, 331, 340, 343-344, 347, 351-352, 356-357, 361-362 Hawkes, K., 92, 140 Hedeen, C., 328, 332 Heiser, M., 93, 140 Henrich, J., 234, 239, 243 Henry, J., 90, 140 Henry, S., 88, 140 Hensley, W., 99, 140 Hepper, P., 88, 140 Hepworth, C., 310, 331 Herdt, G., 69, 86-87, 140 Herzog, T., 151, 163 Hewes, G., 98, 140 Higgins, E., 193, 195 Higham, P., 99, 140 Hinchcliffe, M., 327, 332 Hirschfeld, L., 69, 75, 140 Hirshleifer, D., 316, 333 Hoben, A., 110, 140 Hobson, R., 158, 162 Hofstede, G., 265, 284 Hohmann, G., 92, 140 Hokanson, J., 310, 327-329, 332, 334 Holland, J., 85-86, 89-90, 140 Hollis, A., 76, 140 Holmes, J., 287, 306 Horn, H., 231, 243 Hoofdakker, R., 327, 332 Hooper, D., 327, 332 Hops, H., 328-329, 331, 333 Horowitz, L., 342, 352, 359, 361, 362 Howes, M., 328, 332 Huismans, S., 266, 285 Humphrey, N., 314, 333 Hunting, R., 105, 111, 132, 140 Huston, T., 289, 292, 305, 306 Huttenlocher, J., 116, 142 Hyler, S., 352, 362

367

AUTHOR INDEX

I lacoboni, M., 36, 56, 93, 137, 140 Inglehart, R., 265, 284 Ingram, R., 312, 333 Insel, T., 88-89, 133, 140, 141, 146 Izard, C., 309, 333

J

Jackendoff, R., 37, 56, 150, 163 Jackson, D., 130, 136 James, W., 71, 74, 114, 141 Jansen, J., 327, 332 Jaymous, R., 76, 141 Jemmer, J., 327, 332 Jensen, L, 83, 144 Jensen, M., 287, 305 Johnson, M., 95, 141 Joiner, T., 309, 323, 327, 329, 333 Jonsson, T., 200, 220 Juhl, H., 268, 284

K Kagel, J., 247, 261 Kahneman, D., 225, 243, 256, 258, 261, 262 Kakridis, H., 76, 141 Kaplan, S., 31, 57 Karantzas, G., 168, 196 Karp, D., 268, 284 Katz, J., 323, 333 Katz, L., 229, 243 Keating, E., 97, 130, 141 Kelley, H., 289-290, 306 Kenrick, D., 274, 283 Kernberg, 0., 337, 362 Khatib-Chadidi, J., 76, 141 Kiesler, D., 33, 56, 340, 362 King, K., 168, 196 Kitayama, S., 151, 162 Klam, M., 89-90, 141 Klein, D. N., 309, 310, 333 Klin, A., 158, 163 Klinger, E., 310, 333 Kluckhohn, C., 265, 284 Kluckhohn, R., 87, 145 Knetsch, L., 225, 243, 256, 261

Knowlton, B., 36, 56, 72, 144 Koehler, J., 258, 261 Koller, S., 278-279, 284 Kowalski, R., 311, 333 Kramer, R., 287, 306 Krauss, R., 287, 305 Kristel, 0., 41-42, 57, 249, 257, 261, 262, 344, 362

L Lakoff, G., 95, 141 Landrum, G., 310, 333 Laponce, J., 100, 141 Larzelere, R., 292, 306 Lavelle, J., 226, 243 Lawrence, J., 168, 175-176, 190, 195, 196 Laye, C., 80, 141 Leach, E., 94, 97, 141 Leary, M., 311, 318-319, 331, 333 Lee, A., 158, 162 Lee, F., 151, 163 Lee, S., 78, 141 Lenzenweger, M., 336, 361 Lenzi, G., 93, 137 Leonard, R., 193, 196 Lerner, J., 41-42, 57, 249, 254, 262, 344, 362 Leslie, A., 321, 333 Lester, R., 82, 141 Leventhal, G., 216, 220 Levine, D., 17 Levine, S., 116, 142 LeVine, R., 16-17 Levinger, G., 289, 306 Levinson, S., 100, 137 Levi-Strauss, C., 108, 141 Lewicki, R., 215, 220 Lewinsohn, P., 309-310, 327, 333, 334 Li, Q., 182, 195 Li, Y., 182, 195 Libert, J., 327, 333 Lickel, B., 40, 56 Lieberman, M. D., 36, 56 Lierberman, J., 292, 306 Lillard, A., 151-152, 163 Lim, M., 88, 146 Lind, E., 109, 145, 190, 196 Lindblom, G., 76, 141 Lindenbaum, S., 78, 141

368 Loewenstein, D., 328, 332 Lovejoy, M, 329, 333 Luce, R. D., 287, 306 Lucy, J., 120, 141 Lutz, C., 75, 141

M MacCrimmon, K., 32, 56 MacGregor, 89, 146 Machotka, P., 98-99, 144 MacLean, P., 323, 333 Maeda, F., 93, 140 Mahapatra, M, 278, 285 Maio, G., 268, 284 Malinowski, B., 96, 108, 141 Manson, J., 15, 25, 92, 141 March, J., 256, 261 Marcus, J., 93, 140 Marett, R., 71, 114, 141 Markus, H., 151, 162 Mathews, A., 310, 333 Matthiassen, A., 89, 141 Mauss, M., 71, 141 Maynard Smith, J., 13, 25, 314, 316, 333 Maziotta, J., 93, 137, 140 McCauley, C., 275, 284 McDougal, W., 84, 142 McFall, L, 292, 306 McGraw, A., 42, 57, 252, 257, 261 McGuire, M., 311, 322, 332 Mclntyre, C., 310, 334 McNeill, W., 85, 142 Meckling, W., 287, 305 Meehl, P., 20, 31, 56, 347, 362 Mehrabian, A., 83, 100, 142 Meillassoux, C., 108, 142 Melamed, T., 99, 142 Meltzoff, A., 93, 142 Merleau-Ponty, M., 71, 114, 142 Merry, J., 338, 362 Messick, D., 32, 56 Metalsky, G., 323, 329, 333 Mikula, G., 39, 56, 172, 196, 216, 220 Milgram, S., 264, 284 Millberg, S., 328, 331 Miller, G., 289, 306 Miller, J., 151, 163, 278, 285 Miller, L., 78, 142 Mills, J., 32, 34, 56, 171, 195, 216, 219

AUTHOR INDEX Minton, J., 215, 220 Miranda, J., 312, 333 Mishra, A., 292, 306 Mitchell, P., 148, 163 Mitchell, W., 107, 142 Mithen, S., 116, 142 Mix, K., 116, 142 Molnar-Szakacs, 36, 53 Monroe, S., 310, 333 Montagu, A., 83, 142 Moritz, M., 36, 56 Morris, D., 83, 97, 100, 142 Morris, M., 151, 163 Morrison, B., 213, 220 Much, N., 278, 285 Murnighan, J., 287, 306 Mussell, K., 77, 136

N Nelson, G., 328, 334 Nemeroff, C., 70, 81, 84, 143 Nesse, R., 119, 142, 310-311, 334 Nilsen, R., 89, 146 Nisbett, R., 134, 142, 151, 162, 163 Nissen, E., 89, 141 Nkomo, S., 197, 220

O Oatley, K., 309, 334 O'Connell, J., 92, 140 O'Connell, S., 148, 163 Olausson, H., 89, 142 Olaveson, T., 85, 90, 142, 144 Olson, J., 256, 261, 264, 268, 284, 285 Orbell, J., 316, 334 Orford, J., 340, 362 Ottenberg, S., 86, 142 Ouchi, W., 10, 25 Ouellette, R., 328, 331 Ouston, J., 158, 162 Ozonoff, S., 159, 163

P Packer, C., 316-317, 334

369

AUTHOR INDEX Panksepp, J., 89, 142, 230, 241, 243, 310, 313, 334 Parish, A., 92, 141 Park, L, 278, 285 Parker, G., 316, 331 Parsons, T., 32, 56, 112-113, 142 Pataki, S., 327, 331 Pearce, J., 312, 326, 332 Pegalis, L., 328, 334 Peirce, C., 63, 70, 103, 120, 142 Peng, K., 151, 163 Pennington, B., 159, 163 Perry, A., 92, 141, 142 Peterson, C., 329, 334 Peterson, R., 41, 57, 249, 254, 262 Piaget, J., 17, 65, 72, 103-104, 116, 143 Pincus, A., 338, 341, 352, 360, 362 Pitt-Rivers, J., 69, 76, 143 Plutchik, R., 48, 56 Polanyi, K., 18, 111-112, 114, 143 Poole, F., 69, 143 Poulson, B., 42, 56 Povinelli, D, 154, 163 Powell, E., 98, 130, 144 Powell, M., 328, 331 Premack, D., 147, 163 Pretzer, J., 337, 362 Preuss, T., 154, 163 Price, G., 13, 25, 316, 333 Price, J., 311-312, 326, 334 Puohiniemi, M., 268, 284 Pusey, A., 316-317, 334

Q,R Quiatt, D., 155, 163 Rabain, J., 75, 143 Raiffa, H., 287, 306 Randel, A., 213, 219 Rankin, L., 106, 136 Ransjo-Arvidson, A., 89, 141 Rasmusen, E., 316, 332 Rawls, J., 118, 129, 143, 235, 243 Reece, S., 76, 143 Reichert, T., 46, 55, 351, 362 Rella, J., 90, 140 Rempel, J., 287, 302, 306 Reynolds, S., 85, 89-90, 143, 163 Reynolds, V., 155, 163 Rheingold, H., 178, 196

Rhode, P., 310, 311, 333, 334 Riesenberg, S., 97, 100, 139 Richards, A., 76, 121, 143 Richerson, P., 13, 24, 67, 143 Ridgeway, V., 310, 333 Ring, P., 287, 306 Rizzolatti, G., 93, 143 Roberts, F., 327, 332 Roberts, R., 88, 137 Robinson, L., 327, 332 Roccas, S., 268, 284 Rogers, S., 159, 163 Rokeach, M., 265, 284 Romney, D., 338, 362 Rosaldo, R., 98, 143 Rosenbaum, M., 89-90, 136 Rosenberg, E., 231, 243 Rosenberg, S., 352, 362 Rosenblatt, J., 89, 143 Rosova, V., 268, 283 Roth, A., 247, 261 Rothlind, J., 329, 331 Rothschild-Whitt, J., 88, 143 Rozin, P., 70, 73, 78, 81, 84, 142, 143, 278-279, 282, 284 Rubin, P., 234, 239, 243 Ruble, D., 193, 195 Ruiz, M., 341, 362 Rushkoff, D., 85, 90, 143 Russell, B., 71, 114, 143 Russell, J., 159, 163 Ryan, J., 168, 196 Ryle, G., 71, 114, 143

S Sabini, J., 269-270, 284 Sacco, W., 310, 332 Sagiv, L., 266, 268, 276, 284, 285 Sahlins, M., 77, 143 Salter, F., 321, 334 Samuelson, W., 256, 261 Sapir, E., 71, 143 Scanlon, T., 235, 243 Schacter, D., 72, 144 Schieffelin, E., 76, 144 Schneider, D., 69, 144 Schoener, T., 325, 334 Schubot, D., 268, 284 Schulkin, J., 158, 163 Schutz, A., 71, 144

370

AUTHOR INDEX

Schwartz, B., 98-99, 130, 144 Schwartz, S., 265-268, 276, 283, 284, 285 Schwinger, T., 216, 220 Seeley, J., 309-310, 333 Segal, Z., 312, 333 Segrin, C., 314, 323, 328, 334 Seligman, C., 262, 285 Seligman, M., 310, 323, 334 Semedar, A., 45, 56, 327, 331 Seyfarth, R., 155, 162 Sharpley, C., 105, 111, 132, 140 Shaver, P., 342, 361 Sheppard, B., 38, 43-44, 50, 55, 56, 215, 220, 287-288, 292, 302, 304, 306 Sherman, D., 50, 56, 287-288, 292, 304, 306 Sherman, L., 329, 331, 333 Sherman, S., 40, 56 Sherrod, A., 70, 81, 143 Shils, E., 32, 56 Shine, J., 343, 361 Shure, M., 178, 196 Shweder, R., 16, 83, 144, 151, 163, 278-280, 282, 285

Sieroszewski, W., 110, 144 Sim, J., 338, 362 Simmel, G., Ill, 113, 144 Simon, H., 10, 25 Simson, P., 323, 334 Skarlicki, D., 236, 243 Sloman, L, 311,334 Smetana, J., 182-183, 196 Smith, A., 232, 243 Smith, J., 192, 196 Smith, P., 265, 285 Smith, W., 69, 72-73, 75, 79, 144 Sobal, J., 77, 144 Soldz, S., 338, 362 Solomon, L., 292, 306 Sondak, H.( 38-39, 56, 220 Sparrow, S., 158, 163 Spencer, H., 97, 102, 144 Spencer, P., 87, 144 Sperber, D., 122, 144 Spiegel, J., 98-99, 144 Squire, L., 72, 144 Steer, R., 343, 361 Stenberg, C., 230, 243 Stephens, R., 329, 334 Stiles, W., 328, 331 Strathern, A., 78, 144 Strickland, L., 292, 306 Strong, R., 99, 132, 144

T Tajfel, H., 87, 144 Takahashi, M., 85, 144 Tambor, E., 318, 333 Taraban, C., 328, 331 Tarde, G., 84, 144 Taylor, A. 194, 195 Taylor, S., 247, 261 Taymans, S., 88, 137 Tegnaeus, H., 80, 145 Teitelbaum, 0., 159, 163 Teitelbaum, P., 159, 163 TenHouten, W., 48-49, 56 Terdal, S., 318, 333 Tesser, A., 98, 130, 144 Testart, A., 92, 145 Tetlock, P., 41, 42, 55, 57, 190, 195, 199-200, 208, 219, 247-250, 252-254, 257-259, 261, 262, 344, 358, 361, 362 Thaler, R., 225, 243, 256, 261, 262 Thibaut, J., 289-290, 306 Throop, C. J., 36, 56 Thumin, F., 99, 137 Tice, D., 319, 331 Toates, F., 323, 334 Tolbert, P., 203, 220 Tolbert, R., 89-90, 139 Tomasello, M., 149, 154, 163 Tonnies, F., 198, 220 Tooby, J., 14, 25, 229, 243, 314, 332 Toren, C., 97, 101, 145 Tornblom, K., 200, 220 Triandis, H., 50-51, 57 Trivers, R., 316-317, 334 Trumbell, H., 80, 145 Tuchinsky, M., 43-44, 56, 289, 302, 306 Tulving, E., 72, 145 Turillo, C., 226, 228, 232, 238, 243 Turner, J., 20, 25 Turner, T., 83, 145 Turner, V., 85, 145 Tversky, A., 256, 258, 261, 262 Tyler, T., 109, 145, 190, 196 Tylor, E., 70, 95, 145

u Udy, S., 10, 18, 25 Umphress, E., 226, 243

371

AUTHOR INDEX Ureno, G., 352, 362 Uvnas-Moberg, K., 88, 89, 141, 145

V van de Yen, A., 287, 306 van Gennep, A., 74-75, 83, 85, 145 Villasenor, V., 352, 362 Vitkus, J., 359, 362 Vodosek, M., 30, 40, 43, 50, 57, 347, 362 Volkmar, F., 158, 163 Von Hoist, D., 323, 334 Vygotsky, L, 67, 145

w Wagner, A., 72, 144 Waller, N., 342, 361 Walster, E., 214, 220, 228, 242 Walster, G., 214, 220 Wane, M., 70, 81, 143 Wang, Z., 89, 141 Warton, P., 169, 173, 179-180, 186-187, 189, 196 Watson, D., 310, 334 Watson, J. L, 76, 145 Watson, J. M., 116, 145 Waymire, K., 89, 146 Weakland, J., 130, 136 Weber, M., 17, 198, 220 Weiss, J., 323, 334 Welker, R., 329, 334 Wellman, H., 148, 163 Wells, R., 80, 145 Wetzel, P., 99, 145

White, G., 96, 145 Whitehead, H., 20, 25, 37-38, 57, 80, 82, 145, 199, 208, 213-214, 220 Whitehouse, H., 123, 145 Whiten, A., 148, 152, 162, 163, 314, 331 Whiting, J., 16, 74, 87, 92, 137, 145 Widiger, T., 338, 362 Wiggins, J., 338, 340, 352, 360, 362 Williams, G., 314, 334 Williams, J., 88, 137 Williams, K., 319, 334 Williamson, D., 310, 333 Williamson, G., 328, 334 Williamson, O., 10, 25, 287, 306 Wiggins, J., 33, 46, 57 Wish, M., 31, 57 Woodburn, J., 107, 145 Woodruff, G., 147, 163 Woods, K., 168, 196 Wrangham, R., 15, 25

Y Yamagata, N., 76, 146 Yamey, B., 110, 139 Young, F., 86, 146 Young, L., 88-89, 133, 140, 141, 146 Youngren, M., 327, 334

Z Zadro, L., 319, 334 Zanna, M., 262, 285, 287, 306 Zeckhauser, R., 256, 261 Zelizer, V., 114, 146 Zuroff, D., 343, 361

This page intentionally left blank

Subject Index

A Africa, 17-18, 37, 68, 75, 79-80, 85-86 Animal models and homologies, 14-16, 88, 91-92, 102-103, 111, 115, 154-155, 240, 311, 323 Anthropology, 16-17, 19, 95, 131, 151 Attachment, 310, 326, 342 Autism, 157-159

B,C Bodies, 69-94, 96 Child development, 9, 17, 49, 53, 70, 75, 84, 92, 104, 111, 116, 121, 170-171, 178-179, 181-183, 185, 187, 192-193 Cognitive dissonance, 269-270 Communal vs. exchange distinction, 32, 34-35, 171, 184, 216 Conflict, 20, 39-43, 102, 177, 180-181, 185, 187, 190, 197, 200, 203-211, 219, 222, 231-233, 277, 280-281, 344 Conformation systems, 68, 116-118, 124-127 Conformity, 263-264 Culture, 3, 7, 11, 22, 35, 50-51, 53, 67, 97-99, 120-123, 128, 129, 134, 151-152, 157, 193-194, 242 Cybernetics, 320-325

D Decision-making, 6, 190, 258 Depression, 45, 309-319, 323, 326-330 Diversity, 201-211

E Ecstasy, 89-90 Emotions, 9-10, 41, 48-49, 65, 177, 190, 229-231, 238-240, 277-278, 281-283, 309-310, 344 Evolution, 9, 12-14, 37, 47-48, 94, 153, 224, 229-231, 238, 310-311, 314-316, 323 Experimental games, 225-228, 268

F Families, 38, 39, 168, 169, 174-176 Folk psychology, 151, 160 Food, 71-82 Fundamentalism, 259, 260

373

374

SUBJECT INDEX

G,H Groups, 39, 40 Hospitality, 76

I

Individualism/collectivism, 36, 38, 50-51, 150, 214, 219, 342, 346 Initiation, 86-87 Institutions, 10, 269 Interdependence, 289-292, 294-301 Interpersonal circle, 33, 46, 337-340, 343, 347, 352, 355-356

J Justice, 38, 39, 109, 200, 211-218, 221-223, 234-236

Personality, 44, 46-47, 276-277, 281, 343, 356-357 Personality disorders, 21, 45-46, 133, 335-356, 360 Power, 223, 228, 234, 236, 242, 274 Preos, 4-7, 9, 19, 23, 30, 62, 67, 120 Psychoanalysis, 49, 357, 359 Psychopathology, 21, 44-46, 133, 157-159, 310, 312, 335-356, 360 Psychotherapy, 357-360

R Raves, 85-86, 89 Relational taxonomies, 28-33 Resource exchange theory, 29, 31-32, 34-35, 315, 318 Risk, 291-292, 300-302, 313, 316, 319, 325 Ritual, 84-87, 93

S M Meta-relational models, 12, 21, 53 Mods, 3-7, 9, 62, 67 Money, 112-114, 249-253 Morality, 17, 66, 222-223, 228-238, 248, 251, 278-281, 344 Moose, 18, 37-38, 68, 99, 105, 107, 109, 111, 121, 129, 242

SASB model, 340-342, 347 Semiotics, 19, 63, 67, 119-120 Sex, 80-81, 91-92 Social cognition, 34-37, 148-155 Social errors, 19, 34-35, 190-192 Symbolism, 111-116, 129

T N Nationalism, 272 Neural bases, 12-13, 15, 36-37, 88-89, 93, 133, 337 Null relationship, 11, 274-275

O Oceania, 96-97, 102 Organizations, 42-43, 54, 197, 201-211, 221-223, 234-236, 297-298, 302-304 Ostracism, 312-313, 316-319

P Papua New Guinea, 38, 69, 76, 82, 86, 96

Taboos, 41-42, 80-82, 248-252, 254-256, 260 Theory of mind, 48, 147-160, 317-318, 321, 346 Time, 7, 48-49, 101, 103 Transference, 359-360 Trust, 49-50, 287-304

u,v Utilitarianism, 247, 251-252, 255-256, 258 Values, 41, 247, 263-277

w Well-being, 276-277 Work, 6-7, 11, 18, 168-172, 179-180, 184