Social Structure and Emotion

Citation preview

Preface

Sociologists have long recognized the importance of emotion in the organization of social life. Recent advancements in the field have brought theories of inequality, justice, exchange, and status to bear on the study of emotion. At the same time, some traditional sociological theories of emotion have developed an increasingly structural focus. The result is a rich and growing literature that incorporates these research traditions. Social Structure and Emotion brings together many of the prominent scholars responsible for these innovations. The result is a unique collection of new theoretical and empirical contributions that adds to our understanding of the relationship between social structure and emotion. An introductory chapter sets the intellectual context for the contributions contained in the rest of the pages. The chapters are organized into six sections, each of which is framed by a commentary that provides synthesis and perspective. The first section focuses on the relationship between emotions and status and power processes. Identity work is the theme of the second section. The papers in the third section highlight the role of emotions in social exchange, while the fourth section features two papers on justice and moral emotions. In section five are two papers that bring new insights to the frequently studied topic of emotional labor. The book ends with two papers on negative emotions, one theoretical and one empirical, as well as a section commentary. Together these sections result in a book that is both coherent and diverse and that, we hope, will stimulate further research in the area of social structure and emotion. This book would not have been possible without a State of the Art Conference Grant from the University of Georgia’s Office of the Provost, as well as financial support from the UGA Department of Sociology and Institute of Behavioral Research. This funding allowed us convene a group of eminent scholars to a three day meeting at the University of Georgia in April of 2006 to discuss cutting edge research and theory on the topic of social structure and emotion. This conference brought together established scholars in two of the most active traditions in xiii

PRE-P374095.indd xiii

5/27/2008 7:15:57 PM

xiv

preface

the sociology of emotions together with scholars from the more structural group processes tradition. These are scholars who are all actively making new bridges between the study of emotion and the study of social structure. Our invited scholars comprised the intellectual elite in the sociology of emotion – some senior scholars and some ‘rising stars’ – all of whom are actively involved in the production of new knowledge in the area. Early drafts of the papers contained in this volume were initially presented at this conference and later refined based upon feedback from other conference participants and the volume editors. We are indebted to a number of UGA students who assisted with various organizational and/or editorial duties. First among these is Tiffani Everett, who assisted with both conference organization and management of manuscripts for the volume. Christopher D. Moore, David Johnson, and Callie H. Burt also provided invaluable assistance. We also acknowledge the support of two grants from the National Science Foundation awarded to the volume editors (NSF SES0729396 and NSF BCS-0519969). Finally, we thank the chapter authors for their contributions, as well as for their patience in the editorial process. We are honored to be associated with this impressive group of scholars.

PRE-P374095.indd xiv

5/27/2008 7:15:57 PM

Acknowledgments

This book would not have been possible without a State of the Art Conference Grant from the University of Georgia’s Office of the Provost, as well as financial support from the UGA Department of Sociology and Institute of Behavioral Research. This funding allowed us convene a group of eminent scholars to a three day meeting at the University of Georgia in April of 2006 to discuss cutting edge research and theory on the topic of social structure and emotion. This conference brought together established scholars in two of the most active traditions in the sociology of emotions together with scholars from the more structural group processes tradition. These are scholars who are all actively making new bridges between the study of emotion and the study of social structure. Our invited scholars comprised the intellectual elite in the sociology of emotion – some senior scholars and some “rising stars” – all of whom are actively involved in the production of new knowledge in the area. Early drafts of the papers contained in this book were initially presented at this conference and later refined based upon feedback from other conference participants and the book editors. We are indebted to a number of UGA students who assisted with various organizational and/or editorial duties. First among these is Tiffani Everett, who assisted with both conference organization and management of manuscripts for the book. Christopher D. Moore, David Johnson, and Callie H. Burt also provided invaluable assistance. We also acknowledge the support of two grants from the National Science Foundation awarded to the volume editors (NSF# 0729396 and NSF#0519969). Finally, we thank the chapter authors for their contributions, as well as for their patience in the editorial process. We are honored to be associated with this impressive group of scholars.

xv

ACK-P374095.indd xv

5/26/2008 5:21:23 PM

List of Contributors

Seth Abrutyn University of California, Riverside, CA, USA James W. Balkwell University of Georgia, Athens, GA, USA Candyce S. Berger Stony Brook University, Stony Brook, NY, USA Peter J. Burke University of California, Riverside, CA, USA Michael J. Carter University of California, Riverside, CA, USA Christine Cerven University of California, Riverside, CA, USA Jody Clay-Warner University of Georgia, Athens, GA, USA Karen S. Cook Stanford University, Stanford, CA, USA Rebecca J. Erickson University of Akron, Akron, OH, USA Tiffani Everett University of Georgia, Athens, GA, USA Linda E. Francis University of Akron, Akron, OH, USA Clare Francis University of North Dakota, Grand Forks, ND, USA Natasha M. Ganem Savannah College of Art and Design, Savannah, GA, USA Alexandra Gerbasi Stanford University, Stanford, CA, USA Ellen M. Granberg Clemson University, Clemson, SC, USA Michael M. Harrod University of California, Riverside, CA, USA Karen A. Hegtvedt Emory University, Atlanta, GA, USA David R. Heise Indiana University, Bloomington, IN, USA Cathryn Johnson Emory University, Atlanta, GA, USA Christopher P. Kelley University of Iowa, iowa City, IA, USA Karina Kim Stony Brook University, Stony Brook, NY, USA Edward J. Lawler Cornell University, Ithaca, NY, USA Kathryn J. Lively Dartmouth College, Hanover, NH, USA

xvii

CTR-P374095.indd xvii

5/26/2008 4:58:20 PM

xviii

List of Contributors

Michael J. Lovaglia University of Iowa, iowa City, IA, USA Jeffrey W. Lucas University of Maryland, College Park, MD, USA Linda D. Molm University of Arizona, Tucson, AZ, USA Christopher D. Moore Lakeland College, Sheboygan, WI, USA Dawn T. Robinson University of Georgia, Athens, GA, USA Christabel L. Rogalin Purdue University North Central, Westville, IN, USA Lynn Smith-Lovin Duke university, Durham, NC USA Shane D. Soboroff University of Iowa, iowa City, IA, USA Jan E. Stets University of California, Riverside, CA, USA Shane R. Thye University of South Carolina, Columbia, SC, USA Lisa Troyer University of Connecticut, Storrs, CT, USA Jonathan H. Turner University of California, Riverside, CA, USA Jeongkoo Yoon Ajou University, South Korea

CTR-P374095.indd xviii

5/26/2008 4:58:20 PM

Introduction Dawn T. Robinson, Jody Clay-Warner and Tiffani Everett University of Georgia, Athens, GA, USA

Key Words: Group Processes, Structural Symbolic Interaction

The study of human emotions has long interested scholars in various disciplines. Early sociologists gave emotion and sentiment a central theoretical role in key works. Durkheim [1912] (2001) described the way that tribal communities experienced intense, shared feelings – collective effervescence – as a result of religious rituals that distinguished the sacred from the profane. Durkheim argued that this form of religious ecstasy was a social fact rather than simply an individual experience. For Marx (1983), the emotional lives of the laboring class were important features of class struggle. Marx described the way that a certain separation from one’s feelings – alienation – could result in a failure to recognize class interest and an inability to resist oppression. Contrastingly, religious fervor could serve to support repressive class systems. According to Marx only when workers learn to identify with their authentic emotions – that reflect and signal their repressed status – can they effectively energize class action. Weber (1947) described how rational bureaucracies develop to purge emotional experiences from institutional spheres and how, in turn, charismatic leaders occasionally rise to fill the affective voids created by highly developed bureaucracies. Despite the central role emotions played in classic sociological theory, modern sociology largely ignored emotion until about 30 years ago. Within contemporary sociology, however, some of the most interesting theoretical innovations now are occurring at the intersection of social structure and emotion. Scholars are asking how power and status relations affect our emotional lives and how these feelings, in turn, contribute to the development, organization, and stability of social arrangements. Sociologists studying these questions agree on the central role that emotions play in shaping social structure, as well as the importance

Social Structure and Emotion

ITR-P374095.indd 1

1

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/27/2008 7:20:06 PM

2

Social Structure and Emotion

of social structure in shaping emotional lives. The mechanisms for these processes, on the other hand, are the focus of much recent theoretical innovation and debate. In this volume, we bring together the diverse group of scholars responsible for many of these innovations. In the chapters that follow, prominent scholars in the two main traditions that launched the sociology of emotions, along with scholars in the group processes area who are responsible for recent innovative integrations of emotion theory into existing theories of social structure offer new theoretical and empirical contributions to this dialogue. These original works are grouped together and prefaced by commentaries from other leading scholars in the field who provide some additional background and syntheses for the individual chapters. Consequently, we will not provide a chapter by chapter overview as part of this introduction. Rather, here we briefly review the history of research in the sociology of emotions and in the more structurally oriented social psychology traditions. We then describe how the trajectories of these research traditions have led us to the present watershed in research relating social structure and emotion.

SOCIOLOGY OF EMOTION In the latter half of the 1970s sociologists began to rediscover the importance of emotions in understanding social process. This emergent focus manifest in several new books, articles in the discipline’s top research journals, and a new Section of the American Sociological Association dedicated to the Sociology of Emotions. Early work in the sociology of emotions progressed along two divergent paths. The first, mostly qualitative and descriptive, began to develop accounts for the relationship between emotions and social structure by documenting the process of emotion norms and management in a variety of different social domains (e.g., Hochschild 1979; 1983; Simon 1992; Thoits 1985). The second, largely quantitative and predictive, began to develop formal models of the relationship between emotions and social structure (e.g., Averett and Heise 1987; Heise and Thomas 1989; Kemper 1978, 1979, 1991; MacKinnon and Keating 1989; Smith-Lovin 1990; Stets 2003, 2005). The first tradition developed a base of empirical understandings that provided a foundation from which scholars across the discipline (and beyond) could begin to see the central role that emotions play in shaping routine interactions. The second tradition yielded a rich accumulation of empirical tests and theoretical elaboration of those early emotions models. The developing prominence of emotion scholarship was due in large part to the work of Arlie Hochschild. Arlie Hochschild’s (1983) study of Delta flight attendants, The Managed Heart, was one of the most influential of these early works that spearheaded the modern tradition. Prior to Hochschild’s landmark study, sociologists had already begun to study the role of emotions in linking micro and macro processes, specific emotions, and the interactive nature of emotions (Collins 1975, 1981; Hochschild 1979; Kemper 1978, 1979; Scheff 1979).

ITR-P374095.indd 2

5/27/2008 7:20:06 PM

Introduction

3

What distinguished Hochschild’s study from previous work was its ability to generate new questions and research in the sociology of emotions. Hochschild identified four central concepts – feeling rules, emotional labor, surface acting, and deep acting – that defined the modern sociology of emotion for most non-specialists and are featured in most sociological research on emotions. These ideas proved to be incredibly fertile inspiring a cascade of studies examining the management of emotion in various work and institutional contexts. Sociologists quickly embraced Hochschild’s insight that emotions were governed by feeling rules and controlled through culturally guided management. Hochschild called attention to situated character of emotion cultures, such as the emotion culture of flight attendants, and soon there were a variety of studies examining other emotion cultures such as those required of bill collectors (Sutton 1991), litigators (Pierce 1995), fast food employees and insurance salespersons (Leidner 1993), and medical school students (Smith and Kleinman 1989). The normative nature of emotion cultures calls into question issues of emotional deviance. Consequently, more recent examinations of the sociology of emotion have emphasized emotional deviance, crises, and vivid emotional experiences. Thoits (1985, 1990, 2005) and Pugliesi (1987) examined mental health issues as forms of emotional deviance. While Carolyn Ellis examined the ways emotions are experienced and enacted during death and grief (1995). Similarly, other work examined the role of emotions during periods of coping (Francis 1997; Thoits 1986, 1994). Thoits’ work also examined how possessing multiple identities impacts emotional well being (Thoits 1991, 1992). While emotion scholars ran with these early ideas about situated emotion cultures and process and consequences of managing emotions to adhere to these cultures, the field largely ignored the Marxist foundations – and implications – of Hochschild’s theory. Hochschild’s original argument had clear structural roots that were mostly neglected over the next 25 years of emotion-management research. Rebecca Erickson (this volume) addresses this squarely and offers an approach for rectifying this oversight. Other early emotion theorizing did pay more explicit attention to the role of structure in shaping emotional experience. Kemper (1978) proposed a structural theory of emotions based on social exchange and attribution principles. He argued that two dimensions of social relationships – status and power – are universal. Relative positions on these dimensions characterize relationships and determine their emotional character (Kemper and Collins 1990). In Kemper’s social interactional theory, the emotion generated by status and power changes depends on the perceived source of the change and, in some cases, on whether or not the other person in the interaction was liked or disliked. For example, he predicted that status loss would result in anger if the loss appeared to be remediable; such anger would be functional in that it motivated action to regain the lost status. If the loss was irredeemable, however, it would lead to sadness and depression, saving energy and acclimating the individual to his or her new lowered state of resources. Status loss by another, if caused by oneself, led to guilt

ITR-P374095.indd 3

5/27/2008 7:20:06 PM

4

Social Structure and Emotion

if the other was liked (facilitating group survival by preventing in-group insult). If the other was not liked, his or her status loss would cause happiness. Kemper (1991) tested these ideas, comparing predictions from his social interactional theory a set of 162 emotions reported by respondents in the eight nation study of emotions by Scherer, Wallbott, and Summerfield (1986) and found support for the hypotheses about four emotions – fear, anger, sadness, and joy. While explicitly structural, and highly accurate in the few existing tests, Kemper’s theory has not as yet inspired the volume of research that was stimulated by Hochschild’s arguments. Smith-Lovin (2000) argues that this might, in part, be due to the “backward-looking” nature of this theory. Kemper’s theory exists largely as a taxonomy of emotion predictions under various structural conditions. It is not a generative theory; it has no underlying logic or calculus by which one could generate new predictions about previously unconsidered conditions.

STRUCTURAL SYMBOLIC INTERACTION The other main focus of research in the sociology of emotion has been fostered by the structural symbolic interaction tradition. The two major theories of emotion in this more quantitative, structural branch of symbolic interaction are affect control theory (Heise 1979; MacKinnon 1994; Smith-Lovin and Heise 1988) and identity control theory (Burke 1991; Burke and Reitzes 1991). These theories emphasize the central role of identity meanings and the ability of social interactions to sustain them in producing emotional responses. Both of these theories argue that emotion results from a control process by which social actors attempt to maintain cultural meanings during interaction (Wisecup, Robinson, and Smith-Lovin 2006). Affect control theory views emotions as signals that indicate the distance and location of the current impressions in the situation from the stable, culturally given identity meanings. In identity control theory, reflected appraisals (self-relevant social actions from others in a social situation) are compared to identity meanings. When reflected appraisals do not support identity meanings, actors are expected to experience stress and to generate new actions to restore meanings. Both of these theories view emotions as signals that reveal to an actor how well the situation is supporting his or her identity. Emotions occur both when identities are being maintained and when they are disturbed, but their character and valence differs depending both on the situation and on the degree of identity maintenance. While the theories differ in their specific predictions (Robinson and Smith-Lovin 2006), the signaling value of emotions is central to this symbolic interactionist view (Smith-Lovin 1990). Notice that this signal value allows both the emoting actor and the observer (to the extent that the emotions are revealed) to use emotions as an indicator of how interactions are supporting or failing to support the

ITR-P374095.indd 4

5/27/2008 7:20:06 PM

5

Introduction

situation’s definition. Displays of emotion can therefore be manipulated to support a self-presentation – to give off signals that are designed to create a certain impression about who one is and how events are being interpreted. Researchers in these traditions have studied the way that emotional displays impact judgments made about criminal defendants (Robinson, Smith-Lovin, and Tsoudis 1994; Tsoudis 2000a,b; Tsoudis and Smith-Lovin 1998, 2001), how identity meanings can help us understand emotional responses to injustice (Stets 2003, 2005), and how emotions arise from confirming and disconfirming interactions with marital partners (e.g., Stets and Burke 2005).

GROUP PROCESSES RESEARCH The group processes tradition in sociology has generated one of the most systematic and cumulative bodies of theoretical and empirical knowledge in the discipline. This work focuses on the structural forces that organize face-to-face behavior. Research in this area takes place largely within highly standardized paradigms that emphasize control, comparison, and cumulative development of knowledge. Research in expectations states family of theories has amassed an impressive body of knowledge about the way that power and prestige orders develop, evolve, and are legitimated. One of the most prolific variants in this family of theories, status characteristics theory, models the way that we import external statuses into local task interactions, thereby reflecting and perpetuating existing structural inequalities (Berger, Fiskek, Norman, and Zelditch 1977). Researchers in this area have also developed theory based research into how the effects of external status can be mitigated (Cohen and Lotan 1995; Markovsky, Smith, and Berger 1984). Reward expectations states theory describes how referential beliefs about status structures and distributive justice lead to expectations about the distribution of material rewards (e.g., Wagner 1995). Status legitimation theory (Berger, Ridgeway, Fisek and Norman 1998; Ridgeway and Berger 1986) describes the process by which power and prestige hierarchies come to be seen as legitimate. Status construction theory (Ridgeway 1991) models the process by which a social difference can come to have status value in the first place. Another core tradition in the structurally oriented study of group process is the exchange theory tradition (Blau 1964; Cook and Emerson 1978; Molm 1991; Thibault and Kelly 1959). Exchange theory tradition assumes that individual actors are motivated by self-interest. Research in this tradition emphasizes the structural roots of power and models the way that relationships (e.g., Cook et al. 1983), resources (e.g., Molm, Peterson, and Takahashi 1999) and previous exchanges (e.g., Thye 2000) shape social interactions. Power-dependence theory research has yielded insights about how the relationship between structural power and power use is affected by the reciprocal or negotiated nature of exchange (Molm 2003). Network exchange theory has informed us about how structural

ITR-P374095.indd 5

5/27/2008 7:20:06 PM

6

Social Structure and Emotion

power arises from the configuration of entire structures of relations, and not just local dyadic relations (e.g., Markovsky, Willer and Patton 1988). For decades, theoretical progress in group process traditions proceeded by paying only occasional attention to emotion (e.g., Molm 1991; Ridgeway and Johnson 1990). In recent years, however, group processes research has expanded in a number of directions. The traditional theories have begun to integrate theories of justice, evolutionary models, artificial intelligence, models from psychophysiology and neuropsychology, and organizational ecology. More relevantly, in the past several years, creative new variants of group processes theories have begun to build affective processes explicitly into existing structural theories of social interaction. Group process researchers are now investigating the way that emotions shape the character and development of power and prestige hierarchies. Lovaglia and Houser (1996) conducted a study in which workers exchanged gifts (purchased with their own money) before working together on a task. Workers were either made happy – by a partner who gave them a beautiful flower in a vase – or angry – by a partner who sent a card that said: “NO GIFT. I came here to make money, not friends.” People who were angry were much more difficult for the partner to influence than those who received the gift. A second study demonstrated that angry people are harder to influence – even when the person trying exert the influence is not the one who made them angry. Robert Shelley and colleagues (e.g., Shelley 1993, Shelley and Webster 1997) are investigating the way that liking leads to more opportunities for influence in task group interactions. One implication of Lovaglia and Houser’s research is that when high status members feel positively toward low status members it should moderate the status hierarchy. Shelley and Webster, however, found that when the positive regard is mutual and high status folks are in the minority, the status hierarchy can actually be amplified. Drawing from dependence and justice theories, Johnson, Ford, and Kaufman (2000), developed arguments about the relationship between legitimacy, dependence and emotion. In a vignette based experiment, they found that when faced with a supervisor behaving inappropriately, respondents anticipated feeling and expressing more anger and resentment when they were less dependent, and when the supervisor had less legitimacy. Recent theoretical innovation in the social exchange theory similarly focuses on the “order producing” effects of emotions in social exchange (Lawler and Thye 2006, p. 296). Lawler, Thye and Yoon (Lawler 2001; Lawler and Thye 1999; Lawler and Yoon 1996) have led this charge developing new insights and evidence about the way that emotions arise from completed social exchanges, how they affect group cohesion and how cohesion, in turn affects future groupdirected behaviors, as well as how task structure, type of exchange and group identity come into play. This new wave of theoretical development in group processes research meaningfully incorporates emotion dynamics into the very process of social organization. At the same time, the pendulum of research in the sociology of emotion

ITR-P374095.indd 6

5/27/2008 7:20:06 PM

7

Introduction

seems to be swinging from a heavy focus on explorations of situated emotion cultures to include more structurally oriented investigations. Researchers from each of these traditions have much to learn from one another – and much to contribute to our emerging knowledge about the relationship between social structure and emotion. The chapters that follow start us down this path.

REFERENCES Averett, Christine P. and David R. Heise. 1987. “Modified Social Identities: Amalgamations, Attributions, and Emotions.” Journal of Mathematical Sociology 13:103–132. Berger, Joseph, M. Hamit Fisek, Robert Z. Norman, and Morris Zelditch. 1977. Status Characteristics and Social Interaction. New York: Elsevier. Berger, Joseph, Cecilia Ridgeway, M. Hamit Fisek, and Robert Z. Norman. 1998. “The Legitimation and De-legitimation of Power and Prestige Orders.” American Sociological Review 63:379–405. Blau, Peter M. 1964. Exchange and Power in Social Life. New York: Wiley. Burke, Peter J. 1991. “Identity Processes and Social Stress.” American Sociological Review 56:836–849. Burke, Peter J. and Donald C. Reitzes. 1991. “An Identity Approach to Commitment.” Social Psychology Quarterly 54:239–251. Cohen, Elizabeth G. and Rachel A. Lotan. 1995. “Producing Equal Status Interaction in the Heterogeneous Classroom.” American Educational Research Journal 32:99–120. Collins, Randall. 1975. Conflict Sociology: Toward an Explanatory Science. New York: Academic Press. Collins, Randall. 1981. “On the Micro-foundations of Macro-sociology.” American Journal of Sociology 86:984–1014. Cook, Karen S. and Richard M. Emerson. 1978. “Power, Equity, and Commitment in Exchange Networks.” American Sociological Review 43:721–739. Cook, Karen S., Richard Emerson, Mary R. Gillmore, and Toshio Yamagishi. 1983. “The Distribution of Power in Exchange Networks: Theory and Experimental Results.” American Journal of Sociology 89:275–305. Durkheim, Émile. [1912] 2001. The Elementary Forms of the Religious Life. Carol Cosman (Trans.) and Mark S. Cladis (Ed.). London, England: Oxford University Press. Ellis, Carolyn. 1995. Final Negotiations: A Story of Love, Loss, and Chronic Illness. Philadelphia, PA: Temple University Press. Francis, Linda E. 1997. “Ideology and Interpersonal Emotion Management: Redefining Identity in Two Support Groups.” Social Psychology Quarterly 60(2):153–171. Heise, David R. 1979. Understanding Events. Cambridge University Press. Heise, David R. and Lisa Thomas. 1989. “Predicting Impressions Created by Combinations of Emotion and Social Identity.” Social Psychology Quarterly 52:141–148. Hochschild, Arlie R. 1979. “Emotion Work, Feeling Rules, and Social Structure.” American Journal of Sociology 85:551–575. Hochschild, Arlie R. 1983. The Managed Heart: Commercialization of Human Feeling. Berkeley, CA: University of California Press. Johnson, Cathryn, Rebecca Ford, and Joanne Kaufman. 2000. “Emotional Reactions to Conflict: Do Dependence and Legitimacy Matter?.” Social Forces 79:107–137. Kemper, Theodore D. 1978. A Social Interactional Theory of Emotions. New York: Wiley. Kemper, Theodore D. 1979. “Emotion Work, Feeling Rules and Social Structure.” American Journal of Sociology 85:551–575. Kemper, Theodore D. 1991. “Predicting Emotions from Social Relations.” Social Psychology Quarterly 54:330–342.

ITR-P374095.indd 7

5/27/2008 7:20:06 PM

8

Social Structure and Emotion

Kemper, Theodore and Randall Collins. 1990. “Dimensions of Microinteraction.” American Journal of Sociology 96:32–68. Lawler, Edward J. 2001. “An Affect Theory of Social Exchange.” American Journal of Sociology 107:321–352. Lawler, Edward J. and Shane R. Thye. 1999. “Bringing Emotions into Social Exchange Theory.” Annual Review of Sociology 25:217–244. Lawler, Edward J. and Shane R. Thye. 2006. “Social Exchange Theory of Emotions.” In: Jan E. Stets and Jon H. Turner (ed.), Handbook of the Sociology of Emotions, pp. 295–320. New York: Springer. Lawler, Edward J., Shane R. Thye, and Jeongkoo Yoon. 2000. “Emotion and Group Cohesion in Productive Exchange.” American Journal of Sociology 106:616–657. Lawler, Edward J. and Jeongkoo Yoon. 1996. “Commitment in Exchange Relations: Test of a Theory of Relational Cohesion.” American Sociological Review 61:89–108. Leidner, Robin. 1993. Fast Food, Fast Talk: Service Work and the Routinization of Everyday Life. Berkeley: University of California Press. Lovaglia, Michael and Jeffrey A. Houser. 1996. “Emotional Reactions and Status in Groups.” American Sociological Review 61:864–880. MacKinnon, Neil J. 1994. Symbolic Interaction as Affect Control. Albany, NY: SUNY Press. MacKinnon, Neil J. and Leo Keating. 1989. “The Structure of Emotion: A Review of the Problem and a Cross-Cultural Analysis.” Social Psychology Quarterly 52:70–83. Markovsky, Barry, John Skvoretz, David Willer, Michael J. Lovaglia, and Jeffery Erger. 1993. “The Seeds of Weak Power: An Extension of Network Exchange Theory.” American Sociological Review 58:197–209. Markovsky, Barry, LeRoy F. Smith, and Joseph Berger. 1984. “Do Status Interventions Persist?.” American Sociological Review 49:373–382. Marx, Karl and E. Kamenka (ed.). 1983. The Portable Karl Marx. New York: Penguin Books. Molm, Linda D. 1991. “Affect and Social Exchange: Satisfaction in Power-Dependence Relations.” American Sociological Review 56:475–493. Molm, Linda D. 2003. “Theoretical Comparisons of Forms of Exchange.” Sociological Theory 21:1–17. Molm, Linda D., Gretchen Peterson, and Nobuyuki Takahashi. 1999. “Power in Negotiated and Reciprocal Exchange.” American Sociological Review 64:876–890. Pierce, Jennifer L. 1995. Gender Trials: Emotional Lives in Contemporary Law Firms. Berkeley: University of California Press. Pugliesi, Karen. 1987. “Deviation in Emotion and the Labeling of Mental Illness.” Deviant Behavior 8:79–102. Ridgeway, Cecilia. 1991. “The Social Construction of Status Value: Gender and Other Nominal Characteristics.” Social Forces 70:367–386. Ridgeway, Cecilia L. and Joseph M. Berger. 1986. “Expectations, Legitimation, and Dominance Behavior in Task Groups.” American Sociological Review 51:603–617. Ridgeway, Cecilia and Cathryn Johnson. 1990. “What Is the Relationship Between Socioemotional Behavior and Status in Task Groups?.” American Journal of Sociology 95:1189–1212. Robinson, Dawn T. and Lynn Smith-Lovin. 2006. “Affect Control Theory.” In: Peter J. Burke (ed.), Theories in Social Psychology. Stanford University Press. Robinson, Dawn T., Lynn Smith-Lovin, and Olga Tsoudis. 1994. “Heinous Crime or Unfortunate Accident? The Effects of Remorse on Responses to Mock Criminal Confessions.” Social Forces 73:175–190. Scheff, Thomas J. 1979. Catharsis in Healing, Ritual and Drama. Berkeley: University of California Press. Scherer, Klaus R., Harald G. Wallbott, and Angela B. Summerfield. 1986. Experiencing Emotion. New York: Cambridge University Press. Shelley, Robert K. 1993. “How Emotions Organize Interaction.” Advances in Group Processes 10:113–132.

ITR-P374095.indd 8

5/27/2008 7:20:06 PM

Introduction

9

Shelley, Robert K. and Murray Webster, Jr.. 1997. “How Formal Status, Liking, and Ability Status Structure Interaction.” Sociological Perspectives 40:81–107. Smith, Allen C. III and Sherryl Kleinman. 1989. “Managing Emotions in Medical School: Students’ Contact with the Living and the Dead.” Social Psychology Quarterly 52(1):56–69. Smith-Lovin, Lynn. 1990. “Emotion as the Confirmation and Disconfirmation of Identity: An Affect Control Model.” In: T. D. Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 238–270. Albany, NY: SUNY Press. Smith-Lovin, Lynn. 2000. “Simplicity, Uncertainty, and the Power of Generative Theories.” Contemporary Sociology 29:300–306. Smith-Lovin, Lynn and David R. Heise. 1988. Analyzing Social Interaction: Advances in Affect Control Theory. New York: Gordon and Breach Science Publishers. Stets, Jan E. 2003. Justice, Emotion, and Identity Theory. Advances in Identity Theory and Research. Kluwer Academic/Plenum Publishers. Stets, Jan E. 2005. “Examining Emotion in Identity Theory.” Social Psychology Quarterly 68:39–74. Stets, Jan E. and Peter J. Burke. 2005. “New Directions in Identity Control Theory.” Advances in Group Processes 22:43–64. Sutton, Robert L. 1991. “Maintaining Norms about Expressed Emotions: The Case of Bill Collectors.” Administrative Science Quarterly 36:245–268. Thibaut, J. W. and Harold H. Kelley. 1959. The Social Psychology of Groups. New York: Wiley. Thoits, Peggy A. 1985. “Self-Labeling Processes in Mental Illness: The Role of Emotional Deviance.” American Journal of Sociology 91:221–249. Thoits, Peggy A. 1986. “Multiple Identities: Examining Gender and Marital Status Differences in Distress.” American Sociological Review 51:259–272. Thoits, Peggy A. 1990. “Emotional Deviance: Research Agendas.” In: Theodore D. Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 180–203. Albany, NY: SUNY Press. Thoits, Peggy A. 1991. “On Merging Identity Theory and Stress Research.” Social Psychology Quarterly 54:101–112. Thoits, Peggy A. 1992. “Identity Structures and Psychological Well-Being: Gender and Marital Status Comparisons.” Social Psychology Quarterly 55:236–256. Thoits, Peggy A. 1994. “Stressors and Problem-Solving: The Individual as Psychological Activist.” Journal of Health and Social Behavior 35:143–159. Thoits, Peggy A. 2005. “Differential Labeling of Mental Illness by Social Status: A New Look at an Old Problem.” Journal of Health and Social Behavior 46(1):102–119. Thye, Shane R. 2000. “A Status Value Theory of Power in Exchange Relations.” American Sociological Review 65(3):407–432. Tsoudis, Olga. 2000a. “Relation of Affect Control Theory to the Sentencing of Criminals.” Journal of Social Psychology 140:473–486. Tsoudis, Olga. 2000b. “The Likelihood of Victim Restitution in Mock Cases: Are the “Rules of the Game” Different from Prison and Probation?.” Social Behavior and Personality 28:483–500. Tsoudis, Olga and Lynn Smith-Lovin. 1998. “How Bad Was It? The Effects of Victim and Perpetrator Emotion on Responses to Criminal Court Vignettes.” Social Forces 77:695–722. Tsoudis, Olga and Lynn Smith-Lovin. 2001. “Criminal Identity: The Key to Situational Construals in Mock Criminal Court Cases.” Sociological Spectrum 21:3–31. Wagner, David. 1995. “Gender Differences in Reward Preference.” Small Group Research 26:353–371. Weber, Max. 1947. Max Weber: The Theory of Social and Economic Organization. A. M. Henderson and Talcott Parsons (Eds.). New York: The Free Press. Wisecup, Allison, Dawn T. Robinson, and Lynn Smith-Lovin. 2006. “The Sociology of Emotions.” In: Clifton D. Bryant and Dennis L. Peck (ed.), 21st Century Sociology: A Reference Handbook. Thousand Oaks, CA: Sage Publications.

ITR-P374095.indd 9

5/27/2008 7:20:07 PM

PA R T

1 Status, Power, Emotion

CH01-P374095.indd 11

5/26/2008 2:25:21 PM

1 Commentary Lynn Smith-Lovin Duke University, Durham, NC USA

Status and power are the core of the sociological study of emotion. Our discipline’s central contribution to the study of emotion is that these relational features affect emotional response. Encounters that occur in dyadic relationships or small groups evoke emotions that depend on the groups’ status and power structures. Actors experience emotions that are typical of their structural (status/power) positions. Those emotions vary when there is loss or gain in status or power.

EARLY WORK ON STATUS AND POWER IN THE SOCIOLOGY OF EMOTIONS The scholars that founded the modern sociology of emotions in the 1970s focused on these issues. They directed our attention to the interactional imbeddedness of emotional experience and attempted to describe the structural patterning of the interactional environment. These theorists considered how social structural positions, the cultural meanings of those positions, and interactional situations that evoke them influence the personal experience of emotion. They also asked how emotional responses lead individuals to support or change social structures. Kemper’s (1978, 1990) social interactional theory of emotion firmly established status and power as the two central features of the micro-interactional order. Later work with Randall Collins, a more macro-level theorist, argued that these two dimensions were fundamental to social processes and universal in their impact (Kemper and Collins 1990). In Kemper’s theory, status refers to deference that is given voluntarily, while power refers to compliance gained through coercion. Relative positions on these two dimensions defined the key aspects of a Social Structure and Emotion

CH01-P374095.indd 13

13

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 2:25:21 PM

14

Social Structure and Emotion

relationship and determined its emotional character. Changes in status and power led to specific emotional outcomes. For example, loss of power resulted in fear or anxiety; guilt resulted from excess power. Kemper’s theory (1978, 1990) analytically distinguished three types of emotions that came from status–power relationships. Structural emotions were characteristic of a relationship by virtue of its constellation of relative status and power. These structural emotions correspond roughly to what others have called sentiments in that they are stable affective associations with the relationship; one might feel secure and happy in a relationship where one experiences adequate status and power. Situational emotions resulted from changes in status and power as a result of interaction. Finally, anticipatory emotions resulted from the contemplation of future interaction. Kemper (1978, 1990) also included attributional elements in his theory. He argued that emotional outcomes from status and power changes would depend on the perceived source of the change and, in some cases, to whether or not the other person in the interaction was liked or disliked. For example, he said that status loss would result in anger if the loss appeared to be remediable and the result of another’s actions; such anger would be functional in that it motivated action to regain the lost status. If the loss was self-generated or irredeemable, however, it would lead to sadness, saving energy and acclimating the individual to his or her new lowered state of resources. Status loss by another, if caused by oneself, led to guilt if the other was liked (facilitating group survival by preventing in-group insult). If the other was not liked, his or her status loss would cause happiness. Kemper (1991) tested 10 predictions from his theory using undergraduates’ narratives about situations in which they experienced an emotion (happy, sad, angry, and afraid). Coders trained in status–power judgments and his emotion theory accurately predicted the emotion used to elicit the narrative in 65–75% of the situations. The other foundational theory in the early sociology of emotions used status and power in a more subtle way. Hochschild’s (1983) book, The Managed Heart, took a more cultural perspective. It developed the concept of feeling rules, cultural norms about how one should react emotionally to a situation. But Hochschild (1983) embraced the concept of power through her Marxist analysis of workplace control. Hochschild’s feeling rules may have been cultural, but they were not formed on a level playing field devoid of status/power structures. Feeling rules are often dictated by those in positions of power, especially capitalists who control workers in the occupational system. When emotion management is done for a wage, Hochschild called it emotional labor. In her empirical work, Hochschild (1983) saw an airline employer set explicit standards for how flight attendants would react emotionally to customers. The attendants were trained to manage their own emotions to conform to those work rules, using techniques that Hochschild called surface and deep acting. They used positive emotional expressions to manage the emotions of passengers, often eliciting genuinely positive interactions with this surface acting. Other times, they redefined the situation as

CH01-P374095.indd 14

5/26/2008 2:25:21 PM

15

Commentary

one of dealing with a frightened child-like figure, rather than thinking of the abusive customer as a jerk. Such deep acting allowed them to actually elicit the emotions that were required by their work rules. Hochschild linked the prevalence of emotion work to class position, arguing that middle-class service jobs often involve managing one’s own feelings to benefit capitalists by making clients feel good. Professional socialization trained workers explicitly to regulate emotions in dealing with often-demeaning, unreasonable consumers. Hochschild hypothesized that alienation from authentic feelings resulted when emotion management becomes a pervasive part of occupational life. Erickson’s chapter in this book, focusing on burnout among nurses, further develops Hochschild’s structural insights.

STATUS AND POWER IN THE GROUP PROCESS TRADITION Given how central the concepts of status and power were to the early sociology of emotion, it is not surprising that the central micro-sociological theories focusing on each of these relational domains – expectation states theory and exchange theory – soon developed integrative ideas. Scholars in these theoretical frameworks asked how emotion was intertwined with their phenomenon of interest. Not surprisingly, they often found that emotional reactions mediated, resulted from, or qualified the group processes of interest to them. Early in the small group tradition, Bales (1950, 1953) noted that task groups needed socioemotional management to maintain solidarity while discussing disagreements about the work at hand. Bales and Slater (1955) found that groups separated the task and socioemotional functions in different people, but later research indicated that one person usually dominated both task and socioemotional interactions (Anderson and Blanchard 1982). Ridgeway and Johnson (1990) offered a new analysis of the task/socioemotional division, based on expectation states theory and Kemper’s (1978) social interactional theory of emotions. Ridgeway and Johnson proposed that when one group member disagreed with another, the emotional response to the disagreement would vary depending on the target’s position in the group status structure. If someone disagreed with a low-status person, that person would attribute responsibility for the disagreement to his own low-quality task contribution and would respond emotionally with depression (further suppressing task contributions). A high-status person would respond to disagreement by attributing fault to the other group member (“he doesn’t know what he’s talking about”) and respond emotionally with annoyance or anger. This emotional response would likely lead to a sanctioning behavior, negatively reinforcing the disagreement response. Positive responses to task behaviors also caused different emotional reactions, depending on the causal attribution (with internal attribution leading to pride and external attribution leading to gratitude), but both types of response lead to positive behaviors, which reinforced

CH01-P374095.indd 15

5/26/2008 2:25:21 PM

16

Social Structure and Emotion

the positive response. The outcome prediction was that both high- and low-status people were likely to show positive socioemotional behaviors, while only high-status people would display negative socioemotional acts. Ridgeway and Johnson cited a wide body of evidence that agreed with these predictions (see review in Anderson and Blanchard 1982). Status researchers later noted that structural inequality in groups could create negative emotions in the disadvantaged (Lovaglia and Houser 1996). These negative emotions reduced the influence that structurally advantaged people had in collaborative task groups by generating resistance to their ideas. On the other hand, positive feelings evoked by occupying a privileged position could increase the receptivity to ideas from those in lower positions. Kemper’s structural emotions therefore could act to decrease inequality in groups, encouraging group cohesion and productivity. The second major advance in group processes research on emotion comes from the exchange tradition. Lawler, Thye, and Yoon have developed a series of theories that use emotions as primary mediators in response to exchange relations, producing stable social orders that are difficult to explain in purely resource-maximizing terms (see Lawler and Thye 1999 for an excellent review of this work). Lawler (Lawler 1992; Lawler and Yoon 1993, 1996) began this line of development with his argument that the positive emotions evoked by repeated exchanges could lead to relational commitment. Later, the theoretical reasoning was expanded to show how clusters of such positive, committed exchange relationships might lead to the perception of (and commitment to) group boundaries (Lawler 2001; Lawler and Yoon 1998). Thye has argued that an exchange partner’s status might actually affect the value of his or her resources, influencing the extent to which another experiences positive emotion upon their receipt (and increasing the dependence on the alter for their provision) (Thye 1999, 2000). Again, we see that even the most structured of interactions evokes emotional responses that complicate and enrich group processes.

STATUS, POWER, AND EMOTION IN SOCIAL ORDER In Chapters 2 and 3, dealing with status and power, Lovaglia et al. and Thye et al. ask how emotions are involved in maintaining social order when people occupy different positions within the stratification system. First, we deal with a classic problem – how leaders can use their power to direct a group’s activities without evoking negative emotion and resistance. Second, we ask why both high- and lowpower people might feel positive emotion after an unequal exchange. In both cases, the questions and their answers are complicated by long-term structural positions that the actors hold and by the cultural meanings that those positions carry. Lovaglia et al. suggest that integrative humor can be useful to mitigate the effects of leaders’ power use, producing positive emotions and group cohesion.

CH01-P374095.indd 16

5/26/2008 2:25:21 PM

Commentary

17

Non-traditional leaders – those who historically have occupied low-power and low-status positions – may gain more from this technique than more traditional leaders. Unfortunately, Robinson and Smith-Lovin (2001) provide data showing that humor is risky business for such low-status actors: their jokes are liable to fall flat. But if a non-traditional leader (in this case, a woman) can use integrative humor effectively, her status within the group will be enhanced. It will create positive emotion on the part of subordinates, generate group cohesion, and communicate group-oriented motivation. This will be a tough hypothesis to test. In less-controlled group observations, non-traditional leaders who are effective at humor use may be effective because they have other, subtle status characteristics that increase the legitimacy of their leadership position. So, an adequate test of the hypothesis that Lovaglia et al. propose will require manipulation of the extent to which male and female leaders use effective integrative humor. Given that no one has been able to specify exactly what makes a joke funny, this experiment will be hard to design! It will also be tricky to discover whether emotional mediators create the hypothesized effects on leadership effectiveness or whether some other mechanism is operating. Group members may respond with positive emotion to a funny, integrative leader. Alternatively, successful joking may act as a status indicator itself, creating a more legitimate leader without any necessary emotional response. So, if we want to know the mechanism, our experiment just got more complex: we might need confederates to laugh, as well as leaders who joke. In fact, that would probably be the easiest way to create “effective” humor in the lab. One thing we do know about humor is that it is contagious (Robinson and Smith-Lovin 2001). In Chapter 3 on status and power, Thye et al. also deal with the interaction of power and status. The unexpected results of an exchange between Anglo men and Black women are a complex function of status and power. But here, as in the Lovaglia et al. research program, isolating the theoretical mechanisms may be challenging. Thye et al. make an argument based on the positive affect generated by frequent exchange, together with the legitimation of inequality of outcomes through status processes. An alternative explanation may be that identity processes are at work. Longterm occupation of extreme status positions (like Anglo men and Black women at a traditional Southern university) may teach people identity meanings and other cultural patterns that complicate our purely structural understanding of these exchange outcomes. The fact that substantial outcome inequalities between Anglo men and Black women result even when the research participants think they are interacting with status equals means that they have incorporated something of their long-term stratification history into their identity meanings or behavioral routines. The fact that both categories of actors enjoy the exchange interaction in spite of dramatically differing resource outcomes indicates that rewards are often judged relative to expectations, and these expectations may be based on personal history in the stratification system. That which makes a pauper feel rich would be beneath

CH01-P374095.indd 17

5/26/2008 2:25:21 PM

18

Social Structure and Emotion

notice for a prince. These identity- and culturally based interpretations will be difficult to isolate from the legitimacy and affect/relational cohesion arguments that Thye et al. make. Indeed, since legitimacy and group boundary designations are cultural features themselves, there may not be as much difference between the interpretations as the theoretical language would imply. In any case, both of these chapters alert us to the rich interactions between status and power. Since status leads to power advantage and the exercise of power creates status, the two are intricately intertwined. Furthermore, emotional reactions to status and power are both structural and situational. Disentangling the sources of emotional response in complex relationships will require our best theoretical thinking and experimental design.

REFERENCES Bales, Robert F. 1950. Interaction Process Analysis: A Method for the Study of Small Groups. Cambridge, MA: Addison-Wesley. Bales, Robert F. 1953. “The Equilibrium Problem in Small Groups.” In: Talcott Parsons, Robert F. Bales, and Edward A. Shils (eds), Working Papers in the Theory of Action, pp. 111–161. Glencoe, IL: Free Press. Bales, Robert F. and Phillip E. Slater. 1955. “Role Differentiation in Small Decision-making Groups.” In: Talcott Parsons and Phillip E. Slater (eds), The Family, Socialization and Interaction Processes, pp. 259–306. Glencoe, IL: Free Press. Haas, Jack. 1977. “Learning Real Feelings: A Study of High Steel Ironworkers’ Reactions to Fear and Danger.” Sociology of Work and Occupations 4(2):147–170. Hegtvedt, Karen A. 1990. “The Effects of Relationship Structure on Emotional Responses to Inequity.” Social Psychology Quarterly 53(3):214–228. Hochschild, Arlie R. 1983. The Managed Heart: Commercialization of Human Feeling. Berkeley: University of California Press. Kemper, Theodore D. 1978. A Social Interactional Theory of Emotions. New York: Wiley. Kemper, T. D. 1990. “Social Relations and Emotions: A Structural Approach.” In: T. D. Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 207–237. Albany, NY: SUNY Press. Kemper, Theodore D. 1991. “Predicting Emotions from Social Relations.” Social Psychology Quarterly 54(4):330–342. Kemper, Theodore D. and Randall Collins. 1990. “Dimensions of Microinteraction.” American Journal of Sociology 96(1):32–68. Lawler, Edward J. 1992. “Affective Attachments to Nested Groups: A Choice-Process Theory.” American Sociological Review 57(3):327–339. Lawler, Edward J. 2001. “An Affect Theory of Social Exchange.” American Journal of Sociology 107:321–352. Lawler, Edward J. and Shane R. Thye. 1999. “Bringing Emotions into Social Exchange Theory.” Annual Review of Sociology 25:217–244. Lawler, Edward J., Shane R. Thye, and Jeongkoo Yoon. 2000. “Emotion and Group Cohesion in Productive Exchange.” American Journal of Sociology 106(3):616–657. Lawler, Edward J. and Jeongkoo Yoon. 1993. “Power and the Emergence of Commitment Behavior in Negotiated Exchange.” American Sociological Review 58:465–481. Lawler, Edward J. and Jeongkoo Yoon. 1996. “Commitment in Exchange Relations: Test of a Theory of Relational Cohesion.” American Sociological Review 61:89–108.

CH01-P374095.indd 18

5/26/2008 2:25:21 PM

Commentary

19

Lawler, Edward J. and Jeongkoo Yoon. 1998. “Network Structure and Emotion in Exchange Relations.” American Sociological Review 63:871–894. Lovaglia, Michael J. and Jeffrey Houser. 1996. “Emotional Reactions and Status in Groups.” American Sociological Review 61:867–883. Ridgeway, Cecilia and Cathryn Johnson. 1990. “What is the Relationship between Socioemotional Behavior and Status in Task Groups?” American Journal of Sociology 95:1189–1212. Robinson, D. T. and L. Smith-Lovin. 2001. “Getting a Laugh: Gender, Status and Humor in Task Discussions.” Social Forces 80:123–158. Scherer, Klaus R., Harold G. Wallbott, and Angela B. Summerfield. 1986. Experiencing Emotion: A Cross-cultural Study. Cambridge: Cambridge University Press. Thye, Shane R. 1999. “Status Influence and Status Value.” In: David Willer (ed.), Network Exchange Theory, pp. 248–255. Westport, CT: Praeger. Thye, Shane R. 2000. “A Status Value Theory of Power in Exchange Relations.” American Sociological Review 65(3):407–432.

CH01-P374095.indd 19

5/26/2008 2:25:21 PM

2 Humor and the Effectiveness of Diverse Leaders Michael J. Lovaglia1, Christabel L. Rogalin2, Shane D. Soboroff1, Christopher P. Kelley1 and Jeffrey W. Lucas3 1

University of Iowa, Iowa City, IA, USA Purdue University North Central, Westville, IN, USA 3 University of Maryland, College Park, MD, USA

2

ABSTRACT The appropriate use of humor may increase the effectiveness of leaders, especially non-traditional leaders, women, and minority group members. Leaders are often in positions of formal power that they may be required to use. Power, however, creates resistance, more so for non-traditional leaders. Furthermore, because their leadership defies expectations, non-traditional leaders may need to use power more than white male leaders who can more easily rely on the influence conferred by their high status. Effective leaders convey unpleasant information in ways that reduce group members’ resistance to it. Humor has been shown to increase the acceptance of unpleasant information. Thus it may not be coincidence that effective leaders use group-oriented humor more than do less effective leaders. But non-traditional leaders may avoid using humor for fear of being trivialized. If humor, however, operates similarly to self-presentation as group-motivated in group processes, then it will have a greater positive effect on the performance of non-traditional than traditional leaders. Key Words: Humor, Leaders, Gender, Status, Power, Leadership

Social Structure and Emotion

CH02-P374095.indd 21

21

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 2:26:40 PM

22

Social Structure and Emotion

INTRODUCTION Effective leaders use humor in a variety of ways to influence others, damage opponents, and protect themselves from criticism. Examples include the flashing wit of John Kennedy, the biting sarcasm of former US Senator Bob Dole, the homespun warmth of Ronald Reagan, and the parables of Abraham Lincoln carried in amusing stories. These examples also suggest that for leaders some forms of humor may be more effective than others. Bob Dole, a successful leader as a senator, struggled in national politics. The media cited his negative humor as one cause of his poor showing as the Republican Presidential candidate defeated by Bill Clinton, himself known for warm humor. Leadership is one of the most gender and ethnically stratified facets of human society. When we ask students to name a leader they admire, the resulting list is almost all men. African-American students often name Martin Luther King but the list is also mostly white. Women sometimes name their mothers. Women who have led countries do not make the list. Ellen Johnson-Sirleaf was elected President of Liberia in 2006. The media framed the event as the first woman head of state in Africa, seemingly in contrast to more advanced democracies in other parts of the world. North America’s first and only woman head of state, Kim Campbell, was elected Prime Minister of Canada in late June 1993. She held that position until early November the same year, barely 4 months. Women rarely lead countries. In the United States, women have also made little progress leading states and cities. At this writing, eight of our current governors are women and only 14 mayors of the 100 largest US cities are women. The public image of prominent women leaders is also curiously devoid of humor. Think of Golda Meir of Israel, Indira Gandhi of India, Margaret Thatcher of the United Kingdom, or Hillary Clinton. It is difficult to imagine them telling a joke. Although Hillary is working hard to warm up her image as she runs for President, humor is not what she is known for. Indira Gandhi apparently had no personal friends. These are serious people. Governor Ann Richards of Texas was an exception, using humor often and well. When George H.W. Bush was running for president in 1988, she famously said of his propensity for verbal gaffs, “Poor George can’t help himself…. He was born with a silver foot in his mouth.” Widely popular, Governor Richards was unexpectedly upset in her bid for re-election by George W. Bush, some say as payback for her disrespectful comment about George’s father when he lost the 1992 Presidential election, “Don’t let the door hit your ass on the way out” (quoted online in Wikepedia). The perception is that women leaders pay a political price for using humor that white men need not pay. Male members of minority groups who rise to leadership positions also seem reticent about humor, at least in public. Martin Luther King, Jr, for example, had a brilliant sense of humor but he carefully kept it private. After the “I have a dream” speech, King and President Kennedy met to strategize how best to promote integration and civil rights legislation. Also in the meeting were

CH02-P374095.indd 22

5/26/2008 2:26:40 PM

Humor and the Effectiveness of Diverse Leaders

23

other religious leaders promoting civil rights legislation, including Reverend Blake, whose congregation included former president Eisenhower. President Kennedy was unwilling to commit to actively backing civil rights legislation. Worried about political problems for the Democrats, President Kennedy complained about the lack of help from Republican and business leaders. He argued that civil rights groups should push for a bipartisan consensus rather than a Democratic crusade. King stayed quiet as the argument grew heated. Near the end of the meeting he asked whether a private moral appeal to former President Eisenhower might help. Kennedy quickly snapped “No it won’t. No it won’t.” To defuse any concern that Kennedy might have over King appealing directly to Republicans through Eisenhower, King said he meant that groups other than his own might do it, as Ike “happens to be in the other denomination.” “There must be a way,” King said. “Isn’t he a Democrat when he goes to church?” The resulting laughter broke the tension, and a plan was made to quietly approach the former President (Branch 1988:885–886). Despite King’s command of the strategic use of humor, he avoided it in speeches or with the media present. Humor appears to be an effective leadership tool when used by white men while non-traditional leaders usually avoid its use. After briefly reviewing relevant research on leadership, gender, and humor, we develop a theory that humor appropriately used helps leaders by increasing group solidarity, stabilizing the status hierarchy of the group, and increasing the group’s acceptance of necessary but unpalatable information. Counter to the concern of women and minority group members in leadership positions about using humor, the theory suggests that non-traditional leaders will benefit more from the appropriate use of humor than will white men.

GENDER, LEADERSHIP, AND HUMOR As we use the term, a leader holds a high-status position in a task group, a position that receives deference from other group members and confers legitimate means to use power to accomplish the goals of the group. Following status characteristics theory (Berger, Cohen, and Zelditch 1972), a leader’s high status in terms of race, gender, age, and education creates expectations that leaders will make competent contributions to group goals. A leader’s high status increases his influence, defined as willing compliance to the leader’s suggestions by group members that share those goals (Lovaglia et al. 2006). We contrast the leader’s influence based on status with a leader’s power based on her or his position. Leaders usually hold a position that allows and sometimes requires the use of power over group members. Bosses hire and fire workers, for example. Following Willer, Lovaglia, and Markovsky (1997), we define power as a leadership position’s structural capacity to reward group members for following a leader’s directives or to punish them for resistance.

CH02-P374095.indd 23

5/26/2008 2:26:40 PM

24

Social Structure and Emotion

This study focuses on the experiences of women who lead and how the problems they face differ from men in leadership roles. Leaders that are minority group members share the status disadvantage of women, and our developing theory applies to them as well. We focus on women because of the extensive research that has been done on the problems faced by women leaders and because the “glass ceiling” effect, by which women are increasingly disadvantaged as they rise in a leadership hierarchy, has been less thoroughly researched for minority men (Lovaglia et al. 2006). The literature review below covers three areas: (1) gender and its role in leadership, (2) humor and gender, and (3) humor and leadership. WOMEN IN LEADERSHIP POSITIONS

Men and women are employed at nearly equal rates, yet women are visibly absent from higher-power leadership positions in politics and management. Research has found that women are less likely to emerge as leaders compared to men (Eagly and Karau 1991, 2002). Women in high-status positions, including leadership positions, are often viewed as illegitimate (Ridgeway 2001). Furthermore, when women are in leadership positions, the leadership techniques that have been found to be effective for men are not always effective for women (Yoder 2001). For example, women who use dominance are more negatively evaluated because of their lower status (Ridgeway, Diekema, and Johnson 1995). Men, on the other hand, are advantaged when using non-verbal dominance activities. Men who use a high level of visual dominance witness an increase in their influence, while the same behavior decreases women’s influence and their likeability as well (Carli 2001). The evidence suggests a paradox: Women may attain a leadership position legitimately while being perceived as illegitimate users of the power that comes with the position. Findings of previous research show that women who display dominance behaviors are indeed evaluated more negatively than men in the same roles. This difference is more pronounced when the role is traditionally dominated by males and when the type of dominance behavior is classified as directive or autocratic (Eagly, Makhijani, and Klonsky 1992). It has been argued by others that power use, not gender, impacts the evaluations given to leaders by subordinates. Ragins (1991) found that both male and female subordinates did not significantly differ in their evaluations of male and female leaders. Their interpretation of their findings was that subordinates react to leader power in many senses of this term, not gender. Other research has made a similar argument (Podsakoff and Schriesheim 1985). Counter to this, research focusing on the responses of subordinates to female and male leaders found that female leaders received more negative affective responses than did males (Butler and Geis 1990). Furrowed brows, nods of disagreement, and tightening of the mouth accompanied interactions between female leaders and subordinates of both genders at a higher rate than between male leaders and subordinates of either

CH02-P374095.indd 24

5/26/2008 2:26:40 PM

Humor and the Effectiveness of Diverse Leaders

25

gender. Female leaders also were given fewer positive affective responses by male and female subordinates, though no difference in competence was reported between male and female leaders. Similar research has come to the same conclusions (Koch 2005): Female leaders receive more negative evaluations than male leaders. Apparently, then, the leadership of women faces resistance. One explanation supported by research is that the expectations for the orientation of a woman leader to those around her are different than those for men. Women are expected to act in a communal manner (e.g. nurturing and supportive of others), and others react negatively to women who do not act in these ways. If a woman is selfpromoting, those around her will find her behavior unacceptable (Wosinska et al. 1996). Self-promoting women are liked less and are less influential than women perceived to be modest (Rudman 1998). Hence, only women acting in communal ways can expect to gain influence (Carli 2001; Rudman and Glick 2001). In order to be influential, women more than men need to be likable and appear group-oriented (Carli 2001). WOMEN AND HUMOR

Humor research has been criticized for perpetuating the myth that women are not funny or that they lack a sense of humor (Crawford 1995). Humor often depicts women or other disadvantaged groups in ways that perpetuate existing negative stereotypes and advances humor that positively portrays white men. Perhaps the study of humor has been confined to humor that is funny to men, while ignoring unique aspects of female humor construction and appreciation. Humor may also deny status to women or minorities. When a high-status individual tells a joke, the reaction is more likely to be positive. Not laughing when a lower-status individuals attempts to be humorous may reinforce the status hierarchy of the group (Crawford 1995). Humor increases the acceptance of extreme, disapproved, or unpleasant information. Humor that portrays status disadvantaged groups in stereotypical ways has been shown to increase negative social perceptions of members of those groups (Ford 1997). Mannell (1977) found greater acceptance of violent behavior presented in a humorous way. Men who are sexist and hostile toward women are more likely to find disparaging jokes aimed at women funny and to repeat them (Thomas and Esses 2004). Women are less likely to engage in this activity. Exposure to sexist humor can lead to a perceived norm of sexism for men who are hostile toward women. Their behavior may come to reflect this perception. Notably the use of neutral, or nonsexist, humor was not found to have this effect (Ford, Wentzil, and Lorion 2001). Despite the negative uses of humor, women may be able to use humor to their advantage. Humor is a tool of gender construction, and it may be used to deconstruct negative associations between genders and their stereotypes (Crawford 2002). By understanding that we use humor in ways that build stereotypes

CH02-P374095.indd 25

5/26/2008 2:26:40 PM

26

Social Structure and Emotion

and negative associations, we may also use it to build positive associations, weaken negative associations, and redefine the concepts of gender in more meaningful ways that are conducive to an equitable and fair assessment of each individual outside the limiting context of gender stereotypes (Crawford 2002). Humor also has the potential to bolster a woman’s performance by reducing anxiety responses to stereotype threat (Ford and Ferguson 2004; Ford et al. 2004). Coping humor may be an effective mechanism for female leaders to improve performance and reduce stress. Humor has also been associated with increased influence over individuals who consider themselves to be highly masculine (Conway and Dube 2002). That is they scored highly on the masculinity dimension of a personality inventory and so can be either men or women. Humorous relative to non-humorous appeals on threatening topics are effective for high-masculinity individuals because they seem particularly adverse to experiencing distress (sadness or fear). High-masculinity men and women have been shown to have a stress-avoidant orientation. Thus humorous appeals with their playfulness and diminishment are more effective than no-humor appeals in promoting preventative behaviors because the humor appeals can match the manner in which these individuals typically respond to threat. High-masculinity men and women exhibit greater intent to adopt preventative behavior in response to humorous appeals as compared to non-humorous appeals. Therefore the use of humorous appeals in threatening situations should be more effective when dealing with high-masculinity men and women and may be a useful strategy for leaders to employ. Thus, humor with groups composed of highly masculine men would be amenable to influence by a humorous female leader. Labeling statements as humor also provides a lower risk alternative to volunteering contributions to the group. Women may employ the protection from negative evaluation that humor affords by the use of disassociating language, such as saying “I’m kidding” after putting forth a suggestion, or claiming that they were simply joking. In this fashion, a female leader could make suggestions or float ideas that may be either acceptable or widely unpopular in a way that can be quickly disavowed, thus minimizing damage to her status as leader. The ability of a leader to test the waters with an idea and whose dominance may be perceived as illegitimate could prove both useful and empowering (Wendt 1998). This technique’s effect on a leader’s influence, however, has not been demonstrated. Research addressing the appropriateness of humor by professionals in the workplace suggests that the use of positive humor – humor that enhances positive perceptions of subordinates’ task behavior (e.g., having a good sense of humor, communicating with humor, enjoying jokes, telling jokes, and using nonoffensive humor) – can increase status, while the use of negative humor – humor that denigrates subordinates – decreases influence. Although the use of positive and negative humor appears to have the same kind of effect when employed by any leader, the effects appear to be greater when employed by female leaders than when used by males (Decker and Rotondo 2001). This research suggests

CH02-P374095.indd 26

5/26/2008 2:26:40 PM

Humor and the Effectiveness of Diverse Leaders

27

that the overall effect of the use of positive humor may offer a greater advantage to women than it does for their male counterparts in the same position. Conversely the use of negative humor may prove more costly for women than for men. If we establish that women gain a greater advantage than do men through the use of positive humor, then using humor well might help women who lead defuse resistance to their leadership. HUMOR AND LEADERSHIP

Leaders use their influence and power to accomplish their work. The problem for leaders, then, is to use power and influence others without alienating group members or endangering the goals of the group. Humor is one strategy for accomplishing this, but not all types of humor are equally effective. Transformational leaders, those who are able to motivate and influence group members without resorting to incentive-based arguments or power use threats, can enhance the positive effects of their leadership through humor use (Avolio, Howell, and Sosik 1999). Leaders rated as “good” rather than “bad” in the US Military Academy were perceived to use good natured humor, while those who were considered bad leaders were more often thought of as using demeaning or mean-spirited humor (Priest and Swain 2002). Leaders rated as good were also perceived to have a sense of humor about themselves and able to laugh about their own failures. Humor can have a positive impact on the performance of groups. The use of humor can be used in place of domination as a method for conflict management. Groups may also come to identify themselves partly on the basis of what is commonly considered funny (Robinson and Smith-Lovin 2001; Seckman and Couch 1989). Humor can release tension in groups (O’Quin and Aronoff 1981). It has been shown to moderate the effects of stress and to increase satisfaction at work (Martin and Lefcourt 1983; O’Quin and Aronoff 1981). Humor can become an end in itself, however, sometimes distracting from group goals (Avolio, Howell, and Sosik 1999; Martin and Lefcourt 1983). Humor has been shown to increase group cohesion (Duncan 1982; Robinson and Smith-Lovin 2001) and to increase the motivation of group members (Crawford 1994). Humor may also be one method for clarifying the social norms and values of a group (Greatbatch and Clark 2003) and establishing shared social identity among group members (Robinson and Smith-Lovin 2001). IMPLICATIONS FOR A THEORY OF GENDER, LEADERSHIP, AND HUMOR

Humor has been intimately linked with status processes in task groups (Robinson and Smith-Lovin 2001). Our gloss of relevant research suggests that humor has several functions in group processes. Appropriately used, humor

CH02-P374095.indd 27

5/26/2008 2:26:40 PM

28

Social Structure and Emotion

builds group solidarity and increases motivation; it stabilizes the status hierarchy of the group by specifying members’ positions in it; and humor can increase the acceptance of unpalatable information by group members. These functions of humor are closely related to the main objectives of effective leaders: building group solidarity, motivating individual sacrifice to achieve group goals, maintaining acceptance of the status hierarchy, and increasing the influence of the leader. The literature also suggests that women in leadership positions face resistance to their leadership. Perhaps not coincidentally, women who use humor risk increasing resistance to status achievement and reducing the legitimacy of their leadership. One study (Decker and Rotondo 2001), however suggests that in at least one context, one type of humor might increase the effectiveness of female leaders more than it increases the effectiveness of male leaders. The following section attempts to explain why.

THEORETICAL DEVELOPMENT Robinson and Smith-Lovin (2001) identify several functions of humor and develop propositions in a theory of humor in task groups. Because they emphasize the implications of humor for status hierarchies and gender stratification, their work is particularly relevant to understanding the relationship between humor and leadership. From previous literature, they identify the functions of humor: meaning making, hierarchy building, cohesion building, and tension relief. They propose that: (a) high-status group members use humor more and more successfully than do low-status group members, (b) successful humor will enhance evaluations of a group member, (c) humor strengthens affective ties among members of task groups, and (d) low-status members will use more cohesion-building humor and less-differentiating humor than will high-status members. We build on Robinson and Smith-Lovin’s (2001) theory and the literature reviewed above to identify ways in which the effects of humor intersect with the requirements of leadership. Recall that we defined leadership as holding a status position in a task group that receives deference from other group members and holds legitimate means for the use of power to accomplish the goals of the group. It follows that an effective leader: 1. holds a high-status position in a group that must be maintained to ensure continued deference by group members, 2. promotes group solidarity and a shared identity to focus members’ contributions on accomplishing group goals. 3. motivates group members to perform difficult tasks to accomplish group goals, tasks they otherwise would avoid, and 4. uses power when needed to accomplish group goals while arousing the least amount of resistance among group members and maintaining influence over them.

CH02-P374095.indd 28

5/26/2008 2:26:40 PM

Humor and the Effectiveness of Diverse Leaders

29

1. Holding a Leadership Position. In task groups, individual members have at least two goals: (1) successful completion of group tasks and (2) status attainment. Members want to move up the status hierarchy, which can motivate them to contribute to group goals. This may be especially true where a formal leadership position exists that produces not only the benefits of high status but also the ability to wield power. One goal of leaders, some would say their primary goal, is to maintain power. Humor can be used to stabilize the status hierarchy in at least two ways. First, high-status members can tell jokes at the expense of those below them who are trying to rise in the hierarchy, signaling the group to belittle a low-status member’s attempts to attain higher status, thus keeping members in their place and stabilizing the hierarchy. Second, responses to humor attempts can be a status indicator. We propose that the more members laugh at another’s humor attempts, the more status that member is likely to have. Notice that this mechanism can explain why the jokes told by leaders and other high-status group members are often dramatically less funny than those told by low-status members. As a status indicator, robust laughter in response to a bad joke is much better than equal laughter in response to a truly funny one. Moreover, low-status group members must work hard to become very good comedians to successfully wrest laughter from those above them. Paradoxically, women and minority leaders may be penalized for being too funny. They are vulnerable to being marginalized as mere comedians. 2. Promoting group solidarity and a shared identity. Solidarity occurs to the extent that individual members have an emotional investment in the group as a thing in itself, as more than the mere sum of its members. Another way to frame solidarity is the formation of a group identity. Shared humor especially of the “us versus them” variety can be especially effective. Remember how Ann Richards brought democrats together when she told the elder George Bush not to let the door hit his ass on the way out of the White House. 3. Motivating members to accomplish group goals. An important aspect of a leader’s job is to induce members to undertake tasks they ordinarily would not consider. Think of a lieutenant leading soldiers in a war: “You three charge up that hill and lob a grenade onto that machine gun before it shoots you 500 times. Any questions? You didn’t want to live forever, did you?” Humor can reduce the resistance to unpalatable information. Robert Reich, President Clinton’s Secretary of Labor, had the difficult task of addressing a conference of labor union presidents after Clinton won passage of NAFTA, the North American Free Trade Agreement that jeopardized union jobs. Union leaders, unwilling to consider the advantages of increased trade, were infuriated by the administration’s betrayal. One was quoted anonymously as saying that the administration knew nothing about working people, and that Bob Reich having spent his whole life in college, did not even know what a screwdriver looks like. Reich began his talk with confrontation. Boos and catcalls greeted his demand to know who had made the screwdriver remark. Bill Bywater, a particularly pugnacious union leader, rose to his feet to respond, “Yeah, I said it. And I’ll say it again. You’ve never even seen a screwdriver.”

CH02-P374095.indd 29

5/26/2008 2:26:40 PM

30

Social Structure and Emotion

Reich stared back at him, “You’re wrong Bill.” The tension increased as they both glared in silence. Then Reich said, “To prove it, Bill, I want to present you with this screwdriver … which I brought here especially for you.” Reich then pulled a large pipe wrench from under the table and handed it to him. The room exploded with laughter. Then Bywater started laughing too. “Mr. Secretary, you’re really something,” he said. Reich responded, “Mr. Bywater, the feeling is mutual.” Bill and Bob shook hands. The meeting continued with a productive conversation on how best to create good jobs in a global economy (Reich 1997:151–152). 4. Using power when needed while minimizing resistance to it. Power use creates resistance that has the potential to decrease a leader’s prestige and influence in the group. To accomplish their goals, leaders whose high status is questionable, non-traditional leaders for example, must use power more often that do leaders with legitimate high status. A self-fulfilling cycle results in which members resist the influence of non-traditional leaders, who then feel required to use power more often, which creates more resistance (Lovaglia et al. 2006). Thus power use is addictive; using power increases the likelihood of its continued use. Effective leaders can use less power because group members are motivated to contribute to the group without the carrot and stick of power. Because the legitimacy of women and minority group leaders is questionable, they face more resistance to their leadership and especially to their power use than do white men. It is precisely this expectation deficit that suggests a counter-intuitive prediction: Women and minority group leaders that use humor effectively will decrease the resistance to their leadership and especially to their power use more than it will for white male leaders who use humor similarly. Because humor carries status expectations, we propose that the effects of humor will operate similarly to other status relevant information. Status characteristics theory and research has demonstrated that the larger the expectation deficit, the greater the increase that results from a single piece of status information disconfirming an individual’s low status – the augmentation and attenuation principles (Berger, Cohen, and Zelditch 1972; Berger, Norman, Balkwell, and Smith 1992; Berger, Rosenholtz, and Zelditch 1980). For example, an unemployed, African-American woman who graduates with a Master’s Degree from the University of Iowa will see a larger increase in expectations for her contributions to the group than will a white male engineer who has gone back to the University of Iowa to get a Master’s Degree. Similarly because the resistance to women and minority group leaders is higher than the resistance to white men, we expect the resistance-lowering benefits of using humor to be greater for nontraditional leaders than for white male leaders.

GENDER, HUMOR, AND LEADER INFLUENCE: A RESEARCH AGENDA Not only do women who lead risk marginalization by using humor, but researchers of humor also risk marginalization. Serious research on humor is rare.

CH02-P374095.indd 30

5/26/2008 2:26:40 PM

Humor and the Effectiveness of Diverse Leaders

31

The typical reaction when we tell people that we are studying humor is a smile and sometimes laughter. “Studying humor? You can’t be serious.” Even the response of emotions researchers is careful. When the lead author mentioned the need to explain why something is funny and what makes a good joke, the response was “That is the least interesting aspect of humor.” Humor is a fundamental aspect of human social life. When Groucho Marx was asked about his famous ability to improvise, how he seemed instantly able to fashion a comedy routine from any situation, Groucho said that improvisational comedy was too important to be left to the last minute. Humor is serious. We are unlikely to gain much understanding of human interaction without understanding humor. A systematic approach to studying humor by leaders would vary the type of leader, the leadership context, and type of humor to determine whether leaders can use humor effectively to enhance performance. Such a program of research requires a number of studies investigating different aspects of the problem including various combinations of status characteristics possessed by leaders and followers: gender, race, age, education, and various relevant task abilities are examples. An initial study might vary gender of leader and type of humor used in a single but important group context: a leader delivering negative information to the group while attempting to motivate them to enhance their performance on an unpleasant task. Hypothesis 1:

Leaders who use inclusive humor that draws members together will be more effective and have more influence over group members than will leaders who use no humor or exclusive humor that favors some group members at the expense of others.

We define inclusive humor as that which draws group members together, that makes them feel part of a solidary group. Note that inclusive humor could be negative in tone, disparaging an opposing group, for example, is a good way to make group members feel collectively superior. Recall Ann Richards rallying the Democrats by joking about Republican George Bush the elder’s wealth and verbal gaffs. “Poor George can’t help himself. … He was born with a silver foot in his mouth.” We contrast inclusive humor with exclusive humor. That is humor that makes some group member feel excluded or less a part of the group than other group members. This type of humor may be used by leaders to stabilize the status hierarchy, to keep rivals in their place, or designate a scapegoat. For example, one group member, Leslie, has recently turned in work late and is out of favor with the boss, causing rumors to circulate in the group. In announcing a critical new project to the group, the boss might say: “This important project is mission critical and has an absolute drop-deadline at close of business Friday, so Leslie has the rest of the week … off.” Hypothesis 2:

Using inclusive humor will benefit female leaders more than it benefits male leaders in terms of effectiveness and influence.

One study exists that suggests the counter-intuitive prediction of Hypothesis 2 might be correct. Decker and Rotondo (2001) proposed that female leaders

CH02-P374095.indd 31

5/26/2008 2:26:40 PM

32

Social Structure and Emotion

avoid humor use because they might be judged negatively for it. In their survey, they asked for respondents’ impressions of their supervisors’ use of two types of humor: positive and negative. Positive humor corresponds roughly to our conception of inclusive humor. That is humor that brings group members together, although humor that disparages an out-group would not be considered positive. Negative humor, while perhaps including some kinds of put-downs that we would classify as exclusive humor, consists of politically incorrect humor, racial jokes or jokes that demean women for example. Decker and Rotondo (2001) hypothesized that leaders’ use of positive humor would increase ratings of leader effectiveness and that negative humor would decrease those ratings. They further hypothesized – as common sense would suggest – that using positive humor would benefit male leaders more than the female leaders. That is, Decker and Rotondo predicted that the difference between subordinate ratings of manager effectiveness for male managers who used positive humor and male managers who use no humor would be greater than the difference between female managers who use humor and female managers who use no humor. They also predicted an overall negative effect on leadership effectiveness ratings for managers who used negative humor. They further predicted that the negative effect would be greater for female than for male managers. Supporting our analysis of leader humor using the principles of attenuation and augmentation, Decker and Rotondo’s (2001) results were opposite to their predictions for their key hypothesis: Female leaders who used positive humor were much more highly rated than female leaders who used no humor, while male leaders who used positive humor were only a little more highly rated than male leaders who used no humor. In contrast, their prediction for the negative relationship between gender and leader use of negative humor was supported. We have avoided the temptation to generalize these results. Decker and Rotondo (2001) found a correlation between the use of positive humor by women managers and the effectiveness ratings they receive from subordinates. That correlation does not imply a causal relationship between female leaders’ humor use and their effectiveness. It might mean, for example, that women leaders who are confident and competent in their humor use are also more confident and competent in general. Their general competence may then have produced higher ratings from subordinates. The use of humor by itself might not increase their effectiveness at all or even decrease effectiveness. That is, it might be that confident and competent female leaders can afford to use humor even though it costs them a little influence and prestige. Because Decker and Rotondo’s (2001) result was unexpected, however, it does justify further investigation of our theoretical mechanism for female leaders to reduce the resistance to their leadership. That mechanism is similar to the way that self-presentation as group-motivated has been shown to reduce resistance to the influence of women in mixed-sex work groups (Ridgeway 1982; Shackelford, Wood, and Worchel 1996). Because expectations for female leaders’ use of humor are lower than for male leaders, the augmentation and

CH02-P374095.indd 32

5/26/2008 2:26:40 PM

Humor and the Effectiveness of Diverse Leaders

33

attenuation principles suggest that the benefits of effective humor use will be greater for women than for men.

CONCLUSION We have begun the systematic investigation of humor used by leaders, first identifying a number of situations in which humor could enhance leader performance in critical areas: (1) stabilizing the status hierarchy and maintaining leadership, (2) promoting group solidarity and shared identity, (3) motivating performance and sacrifice for group goals, and (4) minimizing resistance to the leader’s legitimate power use. We made the counter-intuitive proposal that the effective use of inclusive humor will enhance the performance of non-traditional leaders (women and minority group members) more than it will that of white-male leaders – counter intuitive because non-traditional leaders often avoid the use of humor for fear it will impair rather than enhance their effectiveness. Should our proposal prove valid, it will facilitate the development of techniques applicable to work groups that enhance the performance of non-traditional leaders while decreasing resistance to their leadership. Our analysis suggests several additional studies, for example, investigating whether humor use by minority men can increase their leadership effectiveness. While women and minority men share a status disadvantage in work groups, they differ on important social dimensions. Women, for example, may avoid humor use because it may imply a personal familiarity with male coworkers that could cause problems. Women are more careful than men about inadvertently triggering sexual approaches by coworkers of the opposite gender. Research could discover whether women in leadership positions avoid the use of humor out of concern that it will be perceived by coworkers as signaling openness to initiating sexual relationships and whether certain types of humor use do create such impressions. A third avenue of investigation relates self-deprecating humor use to selfhandicapping, a phenomenon relevant to the behavior of many men but few women. Self-handicapping refers to a strategy of behavior that makes it more difficult for persons to succeed at a task they consider important. Going out drinking instead of studying the night before an important exam is classic selfhandicapping behavior. Men’s self-handicapping has been related to status concerns while women seldom self-handicap regardless of status (Lucas and Lovaglia 2005; Lucas et al. 2007). Self-deprecating humor can be seen as a form of self-handicapping that increases the effectiveness of male leaders. When a leader disparages his own performance in some non-essential area in a humorous way, it shows that he is a team player, that he does not take himself too seriously, and that he is confident in his abilities and the abilities of team members. Women in leadership positions, however, may avoid self-deprecating humor for fear it

CH02-P374095.indd 33

5/26/2008 2:26:40 PM

34

Social Structure and Emotion

will be used against them. For example, a male leader might make fun of his own inability to operate the new technical equipment his company makes, knowing that his reputation for technical innovation protects him. A woman making a similar remark, however, might worry that her attempted humor would be taken seriously, reinforcing perceptions that she is not qualified to lead. Research could discover whether female leaders avoid the use of self-deprecating humor and whether its effective use can benefit women in leadership positions as much or more than it does men. These are but a few ideas for serious research on humor.

REFERENCES Avolio, B. J., J. M. Howell, and J. J. Sosik. 1999. “A Funny Thing Happened on the Way to the Bottom Line: Humor as a Moderator of Leadership Style Effects.” The Academy of Management Journal 42:219–227. Berger, J., B. P. Cohen, and M. Zelditch, Jr.. 1972. “Status Characteristics and Social Interaction.” American Sociological Review 37:241–255. Berger, J., R. Z. Norman, J. Balkwell, and R. F. Smith. 1992. “Status Inconsistency in Task Situations: A Test of Four Status Processing Principles.” American Sociological Review 57:843–855. Berger, J., S. J. Rosenholtz, and Morris Zelditch, Jr.. 1980. “Status Organizing Processes.” Annual Review of Sociology 6:479–508. Branch, T. 1988. Parting the Waters. New York: Simon and Schuster. Butler, D. and F. L. Geis. 1990. “Nonverbal Affect Responses to Male and Female Leaders: Implications for Leadership Evaluations.” Journal of Personality and Social Psychology 58:48–49. Carli, L. L. 2001. “Gender and Social Influence.” Journal of Social Issues 57:725–741. Conway, M. and L. Dube. 2002. “Humor in Persuasion on Threatening Topics: Effectiveness is a Function of Audience Sex Role Orientation.” Personality and Social Psychology Bulletin 28(7):863–873. Crawford, C. B. 1994. “Theory and Implications Regarding the Utilization of Strategic Humor by Leaders.” Journal of Leadership Studies 1:53–67. Crawford, M. 1995. Talking Difference: On Gender and Language. London: Sage Publications. Crawford, M. 2002. “Gender and Humor in Social Context.” Journal of Pragmatics 35:1413–1430. Decker, W. H. and D. M. Rotondo. 2001. “Relationships Among Gender, Type of Humor, and Perceived Leader Effectiveness.” Journal of Managerial Issues 13:450–465. Duncan, W. J. 1982. “Humor in Management: Prospects for Administrative Practice and Research.” The Academy of Management Review 7:136–142. Eagly, A. H. and S. J. Karau. 1991. “Gender and the Emergence of Leaders: A Meta-analysis.” Journal of Personality and Social Psychology 60:685–710. Eagly, A. H. and S. J. Karau. 2002. “Role Congruity Theory of Prejudice Toward Female Leaders.” Psychological Review 109:573–598. Eagly, A. H., M. G. Makhijani, and B. G. Klonsky. 1992. “Gender and the Evaluation of Leaders: A Meta-analysis.” Psychological Bulletin 111:3–22. Ford, T. E. 1997. “Effects of Stereotypical Television Portrayals of African-Americans on Person Perception.” Social Psychology Quarterly 60:266–277. Ford, T. E. and M. A. Ferguson. 2004. “Social Consequences of Disparagement Humor: A Prejudiced Norm Theory.” Personality and Social Psychology Review 8:79–94. Ford, T., E. R. Wentzil, and J. Lorion. 2001. “Effects of Exposure to Sexist Humor on Perceptions of Normative Tolerance of Sexism.” European Journal of Social Psychology 31:677–691. Ford, T. E., M. A. Ferguson, J. L. Brooks, and K. M. Hagadone. 2004. “Coping Sense of Humor Reduces Effects of Stereotype Threat on Women’s Math Performance.” Personality and Social Psychology Bulletin 30:643–653.

CH02-P374095.indd 34

5/26/2008 2:26:41 PM

Humor and the Effectiveness of Diverse Leaders

35

Greatbatch, D. and T. Clark. 2003. “Displaying Group Cohesiveness: Humour and Laughter in the Public Lectures of Management Gurus.” Human Relations 56:1515–1544. Koch, S. 2005. “Evaluative Affect Display Toward Male and Female Leaders of Task-oriented Groups.” Small Group Research 36:678–703. Lovaglia, M. J., J. W. Lucas, C. L. Rogalin, and A. Darwin. 2006. “Power, Status, and Leadership in Diverse Organizations: From Basic Research to Program Development.” In: S. R. Thye and E. J. Lawler (eds), Advances in Group Processes, Vol. 23. New York: Elsevier. Lucas, J. W. and M. J. Lovaglia. 2005. “Self-handicapping, Gender, Race, and Status.” Current Research in Social Psychology 10:234–249. Lucas, J. W., H. Ridolfo, R. Youngreen, C. L. Rogalin, S. D. Soboroff, L. Navarre-Jackson, and M. J. Lovaglia. 2007. “Status Processes and Gender Differences in Self-handicapping.” In: S. R. Thye and E. J. Lawler (eds), Advances in Group Processes, Vol. 24. New York: Elsevier. Mannell, R. C. 1977. “Vicarious Superiority, Injustice, and Aggression in Humor: The Role of the Playful Judgmental Set.” In: J. A. Goldstein and H. C. Foot (eds), It’s a Funny Thing, Humor. Elmsford, NY: Pergamon. Martin, R. A. and H. M. Lefcourt. 1983. “Sense of Humor as a Moderator of the Relation Between Stressors and Moods.” Journal of Personality and Social Psychology 45:1313–1324. O’Quin, K. and J. Aronoff. 1981. “Humor as a Technique of Social Influence.” Social Psychology Quarterly 44:349–357. Podsakoff, P. M. and C. A. Schriesheim. 1985. “Field Studies of French and Raven’s Bases of Power: Critique, Reanalysis, and Suggestions for Future Research.” Psychological Bulletin 97: 387–411. Priest, R. F. and J. Swain. 2002. “Humor and Its Implications for Leadership Effectiveness.” Humor 15:169–189. Ragins, B. R. 1991. “Gender Effects in Subordinate Evaluations: Real or Artifact?” Journal of Organizational Behavior 12:259–268. Reich, R. B. 1997. Locked in the Cabinet. New York: Vintage. Ridgeway, C. L. 1982. “Status in Groups: The Importance of Motivation.” American Sociological Review 47:76–88. Ridgeway, C. L. 2001. “Gender, Status, and Leadership.” Journal of Social Issues 57(4):637–655. Ridgeway, C. L., D. Diekema, and C. Johnson. 1995. “Legitimacy, Compliance, and Gender in Peer Groups.” Social Psychology Quarterly 58:298–311. Robinson, D. T. and L. Smith-Lovin. 2001. “Getting a Laugh: Gender, Status and Humor in Task Discussions.” Social Forces 80:123–158. Rudman, L. A. 1998. “Self-promotion as a Risk Factor for Women: The Costs and Benefits of Counterstereotypical Impression Management.” Journal of Personality and Social Psychology 74:629–645. Rudman, L. A. and P. Glick. 2001. “Prescriptive Gender Stereotypes and Backlash Toward Agentic Women.” Journal of Social Issues 57:743–762. Seckman, M. A. and C. J. Couch. 1989. “Jocularity, Sarcasm, and Social Relationships.” Journal of Contemporary Ethnography 18:327–344. Shackelford, S., W. Wood, and S. Worchel. 1996. “Behavioral Styles and the Influence of Women in Mixed-Sex Groups.” Social Psychology Quarterly 59:284–293. Thomas, C. A. and V. M. Esses. 2004. “Individual Differences in Reactions to Sexist Humor.” Group Processes and Intergroup Relations 7:89–100. Wendt, R. F. 1998. “The Sound of One Hand Clapping: Counterintuitive Lessons Extracted from Paradoxes and Double Binds in a Participative Organization.” Management Communication Quarterly 11:232–255. Willer, D., M. J. Lovaglia, and B. Markovsky. 1997. “Power and Influence: A Theoretical Bridge.” Social Forces 76:571–603. Wosinska, W., A. J. Dabul, R. Whetstone-Dion, and R. B. Cialdini. 1996. “Self-presentational Responses to Success in the Organization: The Costs and Benefits of Modesty.” Basic and Applied Social Psychology 18:229–242. Yoder, J. D. 2001. “Making Leadership Work More Effectively for Women.” Journal of Social Issues 57:815–828.

CH02-P374095.indd 35

5/26/2008 2:26:41 PM

3 Social Exchange and the Maintenance of Order in StatusStratifed Systems Shane R. Thye1, Edward J. Lawler2 and Jeongkoo Yoon3 1

University of South Carolina, Columbia, SC, USA 2 Cornell University, Ithaca, NY, USA 3 Ajou University, South Korea

ABSTRACT This paper examines the role of social exchange in the construction of microorder within status-differentiated relations. How order is constructed and maintained in the context of social inequality is a classic sociological problem. We use a serendipitous finding from a recent experiment as a stimulus for theorizing an important feature of this larger problem of order. The finding is that, in an experiment where African-American females negotiated with white males, the white males received much larger payoffs than the African-American females. Yet, despite substantial power and profit differentiation advantaging white males, both individuals reported positive feelings (pleasure/satisfaction and interest/excitement) to the same degree, which contradicts most research on emotional responses to power. We argue that these similar emotional responses, in the context of substantial payoff inequalities, are due to parallel, joint effects of (a) status processes that create and legitimate initial profit differences and (b) exchange processes that make salient a relationship between the actors during repeated exchange. This explanation integrates notions of status value, referential structure, and legitimacy from status theories with notions of relational cohesion Social Structure and Emotion

CH03-P374095.indd 37

37

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 2:30:17 PM

38

Social Structure and Emotion

and shared responsibility from exchange theories. Broadly, the paper proposes some ways to productively interweave ideas from status and exchange theories to explain the emergence or maintenance of enduring social inequalities. Key Words: Status, Power, Emotions, Status Value, Relational Cohesion, Legitimation

INTRODUCTION This paper deals with the general question of whether and how social-exchange processes generate order and stability in the context of social stratification. The question of how social order is maintained given stratification is as old as sociology itself. Structural-functional accounts emphasize how different strata perform different “functions” that are important for society to operate smoothly (Davis and Moore 1945; Parsons 1949, 1951, 1953). Conflict theorists note that stratification emerges from the class struggle over scarce resources (Dahrendorf 1959; Lenski 1966; Pareto 1935). More radical accounts claim that the very nature of social organization inherently produces power and class inequality (Habermas 1973; Marx and Engels 1959). At issue are the mechanisms whereby various kinds of inequalities come to be accepted and seen as legitimate (Della Fave 1980). We detail a number of mechanisms that mitigate problems of order at the microlevel, using status and social-exchange theories. We examine a context in which persons with different status characteristics – in terms of gender and race – have the opportunity to exchange. They have an incentive to exchange because any transaction will generate profit, but status differences may complicate their task. Will differences in gender and race activate processes that bear on how the negotiations transpire? Given such status differences, will higher status individuals earn more profit from exchange with lower status others? If so, how will these individuals respond emotionally to such power differences? Do higher status actors experience positive emotions while lower status actors experience negative emotions? This chapter theoretically and empirically investigates these issues. Our aim is to understand how social-exchange processes affirm or produce stable relations between status unequals. To stimulate our theorizing, we present a serendipitous finding from a recent experiment that examines the impact of status in exchange (Thye and Witkowski 2003). Consistent with previous research (Thye 1999, 2000; Thye, Willer, and Markovsky 2006), this study found that very large power differences emerge wherein white males earned significantly more profit in exchange with African-American females. At the same time, however, both males and females reported positive feelings of pleasure/satisfaction and interest/excitement, despite the inequality. The latter finding is surprising, especially given the literature indicating that low power or status individuals typically experience negative emotions when interacting with higher power or status others (Kemper 1978, 1984; Kemper and Collins 1990; Lawler and Thye 1999; Lovaglia and Houser 1996; Willer, Lovaglia, and Markovsky 1997).

CH03-P374095.indd 38

5/26/2008 2:30:17 PM

Social Exchange and the Maintenance of Order

39

We use this unusual result to explore how status-stratified systems become stable or orderly. To date, we do not know of a theory that can satisfactorily account for the combination of these two findings above.1 Thye’s (Thye 1999, 2000; Thye, Willer, and Markovsky 2006) status value theory of power can account for the first finding linking status-to-power differences, but not for the second; whereas relational cohesion theory can account for the second finding linking exchangeto-emotion, but not for the first (Lawler and Thye 2006; Lawler and Yoon 1993, 1996; Thye, Yoon, and Lawler 2002). The confluence of these two findings is a provocative and unexpected result that serves as the backdrop for this chapter. We (i) use structural social psychology (Lawler, Ridgeway, and Markovsky 1993) as an orienting framework for the problem of order, (ii) briefly review the experimental context and results described above, (iii) evaluate a variety of status- and exchange-based explanations for the serendipitous finding, and (iv) use the results to develop theoretical connections between status, exchange, emotion, and microorder.

THE RESEARCH PROBLEM AND ORIENTATION The larger theoretical problem addressed here can be traced to classic sociological questions regarding the sources of social order. That is, how can stable social order be constructed and maintained in the context of significant inequalities? Weber ([1918] 1968) distinguished inequality along the dimensions of power (i.e., the ability to realize your own will at the expense of others), status (i.e., differences in esteem or honor), and class (differences in wealth or sources of income). Inequalities along these dimensions create issues of domination/ consent, coordination/conflict, justice/injustice, and commitment/alienation. These, in turn, weaken the capacity of a group, organization, or social system to mobilize action on behalf of collective goals and interests (Durkheim [1933] 1964; Parsons 1951; Weber [1916] 1946). Kemper and Collins (1990) argue persuasively that power and status are fundamental dimensions at the microlevel that bridge to the macrorealms. We agree that the bases of social inequality are socially or institutionally defined at the macrostructural/cultural level (e.g., gender, race, class), but also recognize that power and status are “translated” into local encounters or situations (Collins 1981; Kemper and Collins 1990; Lawler, Ridgeway, and Markovsky 1993; Ridgeway 1991; Ridgeway and Erikson 2000; Thye 2000). In this sense, macrodimensions of inequality play out and become “real” to individuals as they 1

Here we focus explicitly on exchange theories of power and expectation state theories of status. Alternatively, the dimensions of male/female and white/African-American could be viewed as important identities that people hold. For example, one could draw insights from sociological theories of identity such as affect control theory (Smith Lovin and Heise 1988) or identity control theory (Burke and Stets 1999). We briefly consider how these theories might apply in the discussion section.

CH03-P374095.indd 39

5/26/2008 2:30:17 PM

40

Social Structure and Emotion

interact with others to achieve individual or collective goals in local situations. Structural social psychology is a broad theoretical frame for understanding this role of microprocesses and structures in creating, sustaining, or changing stable orders (Lawler, Ridgeway, and Markovsky 1993). Much of the social psychological work that falls within the rubric of structural social psychology can be construed as dealing with local contexts to which larger (external) inequalities are imported, and then acted upon. These local, immediate situations involve encounters between purposive and responsive individuals who are oriented to individual or collective goals (Lawler, Ridgeway and Markovsky 1993). Encounters are essentially arenas of social interaction that create, affirm, reproduce, and sometimes alter microstructures and macrostructures. The following excerpt captures this orienting strategy: …structural social psychology, as we conceptualize it, combines the following: a minimal conceptualization of the individual; a focus on encounters, in which multiple individuals interact; and a distinction between micro- and macrostructures as conditions affecting behavior in encounters. Macrostructures frame purposive encounters in which individuals relate to one another; within these encounters, microstructures emerge and organize further encounters. Microstructures forge the link between encounters and macrostructures, and also are the mechanisms by which macrostructures evolve and change … (Lawler, Ridgeway, and Markovsky 1993:270).

Structural social psychology addresses power and status processes with reference to a small set of assumptions about people, their interactions, and larger social structures. Applied to the problem at hand, if people in exchange (an encounter) act on status inequalities originating in macrostructures, they will produce microvariants of that macrostructure, e.g., profit differentiations. Moreover, if they produce microlevel inequalities repeatedly over time, those inequalities become regularized, expected patterns, or normative institutional features of those situations (see Berger and Luckmann 1966; Collins 1989). Ridgeway’s (1991, 2000) status construction theory and Thye’s (Thye 1999, 2000; Thye, Willer, and Markovsky 2006) status value theory of power exemplify the framework of structural social psychology. Both theories illustrate how things that are valued in the larger culture (i.e., resources or characteristics) can affect the value of things in local encounters and, thus, how status value is created and spreads from valued to unvalued elements in encounters. Our purpose is to analyze the emergence of a microstructure under the following scope conditions. First, two individuals are involved in dyadic exchange without alternative partners. Second, they hold resources that are equal in monetary value. Together, these two conditions describe an equal-power, high mutual-dependence situation from which one would not predict differential profits from exchange (Emerson 1972a,b; Willer 1999). Third, the overarching structure is constant such that the individuals have repeated opportunities to exchange and receive rewards or profits. Thus, over time, the actors are essentially negotiating how best to divide the profits from exchange (Markovsky, Willer, and Patton 1988). The repetitiveness allows a microstructure to emerge, that is, a regularized and repeated pattern

CH03-P374095.indd 40

5/26/2008 2:30:17 PM

Social Exchange and the Maintenance of Order

41

of interaction (Berger and Luckmann 1966). Fourth, the individuals differ on one or more status characteristics that carry cultural value (i.e., expectations of relative worth and competence) but have no direct relevance to the exchange. In structural social psychology terms, if repeated exchanges create a pattern of profit differentiation that corresponds with the status inequality, this would be an evidence of an emergent microstructure that reflects the larger macrostructure. Given that such inequalities emerge, an important question is how individuals respond emotionally to those inequalities that are mirrored in the macrostructure.

THEORIES OF STATUS AND POWER This paper draws from two major research programs in structural social psychology: social-exchange theory (Emerson 1972a, b; Homans 1974; Markovsky et al. 1993; Molm and Cook 1995; Willer 1999) and expectation states theory (Berger, Cohen, and Zelditch 1972; Berger, Connor, and Fisek 1974). These research programs have spawned impressive bodies of cumulative theory and supporting evidence on the sources and consequences of differentiation in groups (Cook, Molm, and Yamagishi 1993; Wagner and Berger 1993; Willer 1999). The two programs have developed, respectively, as complementary and non-overlapping accounts of power and status. Social-exchange theory focuses on power in mixed motive situations; expectation states theory focuses on status in collectively oriented groups. Power is defined as the structural capacity to extract valued resources or goods from another (Emerson 1972a, b; Lawler 1992; Willer 1999). Status is defined as actors’ relative standing vis-à-vis others in terms of prestige, value, or honor (Berger, Cohen, and Zelditch 1972). Both theories address inequality: status characteristics theory explains prestige gradients based on culturally valued status characteristics (Ridgeway 1991, 2000); network exchange theory explains resource hierarchies on the basis of network structures and exchange conditions (Willer 1999). Recent years have seen a marked increase in work bridging these traditions by examining the relationship between power and status (see Lovaglia 1994, 1995; Lovaglia and Houser 1996; Lovaglia, Willer, and Troyer 2003; Thye 1999, 2000; Thye, Willer, and Markovsky 2006; Thye and Witkowski 2003). In the case of theory connecting power to status, Lovaglia theorized that those with higher power in an exchange relation are perceived as more competent, and consequently, as possessing higher status and influence (Lovaglia 1994, 1995). However, later empirical research showed that the relation between power and status was more complex than assumed. Lovaglia and Houser (1996) found that power exercise generated negative emotions among low power individuals, and because of this, the influence of high power persons was reduced or undermined (Lovaglia and Houser 1996; Willer, Lovaglia, and Markovsky 1997). Thus, power exercise produces status only when the exercise of power generates perceptions of competence without provoking negative emotions.

CH03-P374095.indd 41

5/26/2008 2:30:17 PM

42

Social Structure and Emotion

In the case of relating status to power, Thye (Thye 1999, 2000; Thye, Willer, and Markovsky 2006) finds conditions under which status differences generate power differences through the spread of status value from a characteristic to a resource. Status value is defined as the honor, prestige, or esteem associated with a characteristic or object. The status value theory of power asserts that status value can spread from a high status person to their resources. For example, a cigar humidor once owned by a former president takes on status value due to the presidential association. The spread of status value makes high status actors more attractive exchange partners, because their goods are more valuable, and thereby enhances their capacity (structural power) to extract more favorable terms of exchange. We step off from the prior work of Thye and associates (Thye 1999, 2000; Thye, Willer, and Markovsky 2006) by developing ideas on the emergence of order and cohesion in social exchange, in particular, as informed by the theory of relational cohesion (Lawler and Thye 2006; Lawler, Thye, and Yoon 2000; Lawler and Yoon 1996; Thye, Yoon, and Lawler 2002).

HOW THEORIES OF POWER AND STATUS EXPLAIN MICROORDER We define microsocial order as a repetitive pattern of interaction (exchange, transaction) among two or more individuals with three attendant dimensions (see Collins 1981; Homans 1951). First, individuals perceive they are a social unit, that is, they come to develop an overarching sense of unity or cohesion. Second, individuals develop emotions or feelings directed at the unit based on their interactions with each other. Finally, they orient their behavior toward the unit in the form of making sacrifices for, giving benefits to, or continuing to interact with others in the unit. This notion of social order interweaves cognitive, emotional, and behavioral dimensions. Frequent or repetitive interaction, continuously reproduced among the same actors, is at the heart of this conception (see also Collins 1981, 2004). Prevailing theories of power and status offer implicit explanations for microsocial order, focusing on how and when exogenous structural (e.g., interdependencies) or cultural factors (e.g., widely held status beliefs) produce stable, repetitive patterns of interaction. Our focus is on whether and how such interactions foster emotions and feelings about the task, others in the situation, and the larger relation or group (Collins 1981; Homans 1951; Lawler and Thye 2006). This emphasis can be traced to Homans’ (1951) tripartite distinction between interaction, activities, and sentiments. The activities are the joint tasks undertaken by the actors, such as social exchange; interactions refer to the frequency with which actors undertake tasks with the same others; and sentiments refer to enduring feelings or affect developed in the course of these recurrent interactions. With this conception in mind, we next explicate how contemporary exchange and status theories address the problem of microsocial order.

CH03-P374095.indd 42

5/26/2008 2:30:18 PM

Social Exchange and the Maintenance of Order

43

ORDER IN SOCIAL-EXCHANGE THEORY

Social-exchange theory explains microorder and stability in terms of the structurally based interdependencies and the associated incentives actors have to exchange valued “goods” with others (Emerson 1972b; Molm and Cook 1995; Willer 1999). Individuals form and maintain relations with others insofar as these relations produce greater rewards than other alternative relations. Interdependencies shape who is likely to exchange with whom and who are advantaged or disadvantaged (Lovaglia et al. 1995; Thye, Lovaglia, and Markovsky 1997). Power differentiation occurs whenever some actors are more dependent on their exchange partners than their partners are on them (Emerson 1972b, 1981); or some actors can exclude others from valued goods without personal cost (Willer 1999). From a social-exchange perspective relations are rather tenuous and precarious. This is because such relations are contingent on stable flows of benefit to each individual assuming that individual preferences do not change. The implication is that people are committed to extant relations only so long as they cannot achieve better individual outcomes elsewhere. Unequal power relations should be especially precarious because they motivate the low power actor to improve his or her power, while the high power actor has an incentive to use and maintain (or even enhance) their power (Emerson 1962, 1964, 1972a, b, 1981; Kipnis 1976). In the context of potential instability, exchange theorists have developed two classes of complementary explanations for microorder, one based on trust and the other based on emotion. Trust, in this literature refers to expectations of cooperation and/or benign behavior based on inferences about another person’s traits or characteristics (Yamagishi and Yamagishi 1994). Emotion is defined as a relatively short-term positive or negative evaluative state with neurological, neuromuscular, and sometimes cognitive manifestations (Damasio 1999; Izard 1991; Kemper 1978). Emotion is distinct from affect, which refers more broadly to any positive or negative evaluation toward an object, including both transitory and enduring feelings (Smith-Lovin 1995). Here, the terms emotion and feeling are used interchangeably. Both trust and emotional explanations seek to explain when actors in exchange are likely to develop a commitment to each other or their relationship (Cook and Emerson 1978; Kollock 1994, 1999; Lawler and Yoon 1993). Commitment is defined as a proclivity to exchange with those one has exchanged with in the past, despite the availability of alternative partners. Commitment, as such, reflects the behavioral dimension of microsocial order as defined above. Let us compare the trust and affect explanations. The trust approach deals with how actors handle risks in social exchange (Molm 2003) and portrays commitment as an uncertainty-reduction strategy. Kollock (1994) compared the propensity of actors in a market to exchange repeatedly with the same others (i.e., commitment behavior) when they know the quality of the products they are buying (low uncertainty) versus when they do not know the quality of products (high

CH03-P374095.indd 43

5/26/2008 2:30:18 PM

44

Social Structure and Emotion

uncertainty). He found that commitments or repeated exchanges with the same actors were more likely under conditions of high compared to low uncertainty. For Kollock (1994), uncertainty was an exogenous market condition promoting exchanges by the same actors. In another study (Lawler, Thye, and Yoon 2000), uncertainty reduction was conceptualized as an endogenous result of repeated exchange, and results show that it facilitated an expansion of the exchange relation. With frequent exchange, actors found each other more predictable, and this in turn made them more willing to risk new ventures with those same others (i.e., expand their relationship). Yamagishi, Cook, and associates (Cook et al. 2005; Yamagishi and Yamagishi 1994; Yamagishi, Cook, and Watabe 1998) contrast commitment and trust solutions, arguing that commitment inhibits the taking of risks to establish new relationships, whereas trust makes this possible. They define trust as a persons’ generalized trust in others (Yamagishi, Cook, and Watabe 1998) and see this as a culturally based orientation to others in general. Commitment leads one to stay with current relationships, thereby missing opportunities for joint activities and mutually profitable exchanges; whereas, generalized trust toward others leads actors to take more risks to seek mutually profitable exchanges and relations. The distinction between commitment and trust is used to understand fundamental differences between American and Japanese business cultures and practices. That is, Americans have higher levels of general trust in others than Japanese and, therefore, rely less on family ties in economic activity (Yamagishi and Yamagishi 1994; Yamagishi, Cook, and Watabe 1998). In short, trust makes order possible under conditions where individuals are self-interested and lack extensive knowledge of each other. This occurs by promoting openness toward new relations and by enhancing efforts to resolve social-dilemma problems. A second approach to microorder focuses on the emotional/affective consequences of exchange and is grounded in relational cohesion theory and research (Lawler 2002; Lawler and Thye 1999, 2006; Lawler Thye, and Yoon 2000, 2006; Lawler and Yoon 1996, 1998; Thye, Yoon, and Lawler 2002; Yoon and Lawler 2006). This theory explicitly speaks to the emotional, cognitive, and behavioral dimensions of microsocial order. The theory is organized around the idea that social exchange has emotional effects on actors, and these emotions have consequences for the ensuing cognitions and strength of exchange relations (i.e., relational cohesion and commitment). The predictions are that (1) more frequent exchange engenders more positive emotions and feelings and (2) these, in turn, generate perceptions that the exchange relation is a unifying force in the exchange situation or encounter. The result (3) is that actors become more committed to their relation, which is manifest in a propensity to continue to exchange with each other (stay behavior), to give token, symbolic gifts to one another (gift giving), and to engage in new joint ventures involving a social dilemma and attendant risks. Experiments on relational cohesion theory consistently have supported this exchange-to-emotion-cohesion causal chain (Thye, Yoon, and Lawler 2002) and, furthermore, have shown that equal power relations have stronger

CH03-P374095.indd 44

5/26/2008 2:30:18 PM

Social Exchange and the Maintenance of Order

45

effects on positive emotions than unequal power relations (Lawler and Yoon 1993, 1996, 1998). To extrapolate, relational cohesion theory and research suggest that stable microorder is likely to emerge if actors are equal in power and if exchanges generate positive emotions and feelings. Within unequal power relations, lower power actors report less pleasure/satisfaction than higher power actors, which is consistent with results of other research (Lovaglia 1995; Willer, Lovaglia, and Markovsky 1997). However, except under extremely low power conditions, relational cohesion research also indicates that lower power actors experience more positive affect if exchanges are consummated than if they are not (Lawler and Yoon 1998). Thus, while unequal power attenuates cohesion, there is still some evidence that cohesion develops due to the positive emotional effects of more frequent exchange. The fact that unequal power relations nevertheless generate some relational cohesion can help us understand how social exchange might foster microorder under conditions of status inequality. ORDER IN EXPECTATION STATES THEORIES

Implicitly, theories within the expectation states tradition explain social order and stability by way of (i) emotional expression (Lovaglia and Houser, 1996; Ridgeway and Johnson 1990) or (ii) legitimation (Ridgeway and Berger 1986; Walker and Zelditch 1993). Here we focus on the latter. An inequality is legitimate when the members of that structure believe that inequality is valid and such differences should exist. Ridgeway and Berger (1986:606) argue that “power and prestige positions based on performance expectations have a normative, moral quality which makes behavior incongruent with those positions an event that is not merely unexpected but also one that should not happen.” Here, we elaborate a multilevel conception of the various ways that theories in the expectation states tradition show how inequalities become legitimate. Ordered inequality was, in some regards, the spark that ignited research and theorizing in the expectation states program. The earliest work in this tradition was aimed at understanding the emergence of influence and prestige hierarchies in small groups (see Wagner and Berger (2002) for an overview). Bales and associates (Bales 1950; Bales et al. 1951) repeatedly found that stable and persistent interactional differences quickly emerged among initially homogenous actors working together on a group task. These early studies led to the development of expectation states theory which centered on the concept of a performance expectation. Performance expectations are non-conscious beliefs about an actor’s future task performance. They are simply “unaware hunches about whose suggestions are likely to be better” (Ridgeway and Walker 1995:288). Now consider how inequalities based on performance expectations might become legitimate. Expectation states theory asserts that, through social interaction, individuals develop performance expectations for themselves and for others regarding task competency. Because performance expectations develop on the

CH03-P374095.indd 45

5/26/2008 2:30:18 PM

46

Social Structure and Emotion

basis of observable acts, these beliefs tend to be consensual within the group. Those perceived to be most competent are higher in the expectation hierarchy, interact more frequently, and have more influence over those perceived to be less competent. Notice that to the extent that group members feel it is necessary and legitimate to take the ideas of others into consideration (i.e., they are collectively oriented) and truly want to do the best job possible (i.e., they are task-oriented), group members should allow those suspected to be most competent to have the most input and influence. Said differently, inequalities in the context of performance expectations become legitimate to the extent that group members are collectively oriented toward furthering the goals of the group. Status characteristics theory is a branch of the larger expectation states program that connects culturally specified beliefs to performance expectations. The theory defines two kinds of status characteristics. A diffuse status characteristic such as gender exists in a particular culture if and only if (i) one state (e.g., male) is more highly valued than the other state (e.g., female), (ii) those with the high state are expected to be more competent at specific tasks such as sports or math, and (iii) those with the high state are generally expected to be more capable in general, i.e., at a wide range of tasks. In contrast, a specific status characteristic satisfies conditions (i) and (ii), but not (iii). For example, math skill is a specific status characteristic if being good at math is preferable to being bad at math, and if one expects a math expert to be competent at other numerical tasks. Next we consider how inequalities based on status characteristics might become legitimate. If a characteristic such as being female is activated and made salient, cultural beliefs about the competence of women are imported into the situation. Once imported, these will shape patterns of deference and interpersonal influence in the interaction (Berger, Connor, and Fisek 1974; Berger et al. 1977; Ridgeway 1991). Individuals with the more highly valued states of the characteristic (e.g., males) will have more action opportunities, receive more deference, and have more influence over others. In short, this inequality is legitimized by a higher level phenomenon – the larger set of cultural beliefs that connect states of the characteristics (male versus female) to states of performance expectations (competent, incompetent). Thus, legitimacy in this setting is traced to the macrolevel. That is: inequalities in the context of status characteristics are seen as legitimate to the extent that such inequalities are anchored in widely shared status beliefs that affirm and support corresponding performance expectations. Finally, we examine how reward inequalities based on referential structures might become legitimate. Referential structures are sets of socially validated beliefs that actors hold in common describing the relationship between various traits and levels of rewards (Berger, Cohen, and Zelditch 1972). Reward expectations theory uses this notion to understand how status characteristics, abilities, and task accomplishments correspond to reward expectations (Berger et al. 1985). The theory assumes that whenever rewards are distributed in a situation, an ability referential structure is activated and actors expect that more able group members will be more highly rewarded for their participation. Categorical

CH03-P374095.indd 46

5/26/2008 2:30:18 PM

Social Exchange and the Maintenance of Order

47

referential structures link reward levels to status characteristics. Such referential structures are socially defined, widely shared, and generally taken for granted. Again, the seeds of legitimation can be traced to the macrolevel. We assert that inequalities in the context of exchange processes are seen as legitimate to the extent that such inequalities activate widely shared referential structures linking ability or status characteristics to levels of reward. Taken together, these theories have much to say about how a status-based inequality is generated and maintained in relations and groups. In general, microorder is not as problematic in the expectation states program as it is in social-exchange theory because of the focus on taken for granted, institutionalized, cultural beliefs that undergird inequality. Consistent with structural social psychology, however, the expectation state tradition makes important connections between a macrolevel social framework and a local task situation or microlevel encounter, as expressed succinctly in the following excerpt from a theoretical piece by Berger, Ridgeway, and Zelditch (2002): … a social framework of already-given status beliefs plays a central role in the emergence of newly constructed status characteristics. But we need to further distinguish the pre-given cultural framework of the larger society … from the subculture the group itself creates in the course of repeated interaction. We assume that over the course of a sequence of situations, a group develops its own local framework – a set of cultural elements that are specific to the group. (p. 160)

Important to note here is the reference to the role of repeated interaction in the construction of microorder. This idea dovetails with the theory of relational cohesion and is important to our analysis of how social exchange promotes order under conditions of status inequality.

SUMMARY

To summarize, theory and research on status and exchange lead us to a number of well-founded propositions that orient and guide our theoretical and empirical analyses. ●

●

●

●

CH03-P374095.indd 47

First, in task situations, differentially valued states of a status characteristic generate differential rewards or profits in social exchange (Thye 1999, 2000; Thye, Willer, and Markovsky 2006). Second, in social exchange, actors who receive more profits have more positive emotions and feelings about their exchanges, while actors who receive fewer profits tend to experience negative emotions (Willer, Lovaglia, and Markovsky 1997). Third, if low power actors experience negative emotions, because power is used against them, it undermines the influence of the high status actor (Lovaglia and Houser 1996). Fourth, in social exchange, exchanges that are repeated more frequently generate greater positive emotions and feelings, especially under equal

5/26/2008 2:30:18 PM

48

●

Social Structure and Emotion

power but also under some unequal-power conditions (Lawler and Yoon 1996, 1998). Fifth, when repeated exchange generates positive emotions and feelings, it also produces relational cohesion and micro social order at the relational or group level (Lawler, Thye, and Yoon 2000; Lawler and Yoon 1996).

Given the above propositions and the scope conditions (see p. 40),we would expect to find empirically: (1) If status characteristics of gender and race are salient in an exchange relation, these status differences should produce differences in rewards or profits from social exchange that advantage the higher status actors. (2) Those actors who earn less in the exchange should experience negative emotions and feelings about the social exchange, and those who earn more from exchange should experience positive emotions and feelings (Kemper 1978, 1984; Willer, Lovaglia, and Markovsky 1997). Next, we turn to an experiment that examined these implications and produced the interesting result, which is the stimulus for our theorizing.

THE EXPERIMENT OVERVIEW

A series of experiments were designed to test how status and resource assignments alter power exercise and emotional reactions in dyadic exchange. Here we briefly describe the experimental conditions (more detail is available in Thye (2000) and Thye and Witkowski (2003). Each experimental session involved two subjects who are higher or lower status than their partner on race and gender dimensions. The two subjects negotiated exchanges of resources (poker chips) with one another over personal computers from different rooms. MANIPULATIONS

The experiment consists of a three-condition completely randomized design. In all conditions white males negotiated with African-American females over computer terminals such that the two are differentiated on race, gender, and meaning insight ability (see below). At no time during the experiment did the two ever meet face-to-face. In condition one, white males negotiated with African-American females and both were told (i) the true race and gender of their partner (along with other personal characteristics that were held constant) and (ii) that the male scored higher on a meaning insight test. Both subjects had blue poker chips to exchange. Condition two was identical to condition one in that each person is again told the true race and gender of their partner; but here subjects had different colored chips. The third and final condition is a control. In this condition, white males negotiated with African-American females, however, both subjects were told their partners were of the exact same race and gender,

CH03-P374095.indd 48

5/26/2008 2:30:18 PM

Social Exchange and the Maintenance of Order

49

and all subjects had the same color chips. Thus, in condition three the subjects were equated on all resource and status characteristics. This allows for a baseline measure of power exercise and emotional reactions unrelated to status or resource differentiation. PROCEDURES

All experiments were conducted at a large southern university. The subjects were college students who were recruited for payment from undergraduate classes. Each experimental session involved a single dyad comprised of one white male and one African-American female. Upon their arrival, subjects were randomly assigned to one of two separate rooms where they remained for the entire session. At no time did subjects learn the true identity of their partner. Except for a few words from the research assistant, all instructions were standardized and delivered to the subjects via the computer (see Thye 2000 for identical procedures). Each experiment had several phases. First, subjects provided the experimenter with information about their gender, age, race, current year in school, academic major, and high school GPA on a background information form. Each individual was told that she or he would be working with one other person via the computer. Each subject was led to believe they were connected to a partner of different race and gender. Depending on the experimental condition, the experimenter either reported the race and gender of the partner (along with consistent photos to add realism) or equated information on race and gender (condition three only). The experimenter also reported that white males had scored higher on a meaning insight test administered just prior to the negotiations (see Berger, Fisek, and Freese 1976 for details). These procedures are comparable to those used in past research to create consistent status differences (Lovaglia and Houser 1996; Markovsky, Smith, and Berger 1984; Thye 2000). Each subject was then assigned a resource (poker chips) with which to exchange. All subjects were told that she or he would be given “blue poker chips in light of your meaning-insight score and the other information we have about you.” A symbolic blue poker chip was given to each subject as a reminder. Then, depending on the condition, subjects were told their partners also possessed blue poker chips or chips of a different color. Thye (2000) suggests that holding different resources is a condition for the spread of status value from a status characteristic to an exchangeable resource, and the difference in chip color was designed to test for such an effect (see also Ridgeway 1991). However, the difference in chip color across conditions one and two does not change the results of concern to us and, therefore, is not relevant to this chapter. Bargaining Protocol Following the poker chip allocations, subjects began the negotiation and exchange phase after a brief tutorial. Subjects completed 60 rounds of negotiation making at most one exchange per round. All subjects were encouraged to gain

CH03-P374095.indd 49

5/26/2008 2:30:18 PM

50

Social Structure and Emotion

as many chips as they could for themselves, and were informed that the number of chips earned in the study would determine their pay. Each subject knew (i) the number of poker chips they requested from their partner and (ii) the number of poker chips earned by self on each round. In reality, subjects negotiated the division of 30 poker chips located in a central pool. Prior to the first round, the central computer initialized all requests so that each subject would request 15 poker chips (half the profit pool) from their partner (details on the specifics are found in Thye, Lovaglia, and Markovsky 1997). On each round, a subject could increase, decrease, or leave unchanged the number of chips acquired through exchange on the previous round. The total number of chips earned is our measure of power use. Importantly, for power to emerge in this setting, one party must increasingly request more chips and the other party must capitulate. Thus, the bargaining protocol requires that both parties jointly participate in the construction of power differences. Following each round, subjects learned the number of poker chips earned, and the supply of 30 poker chips was replenished for the next round. Subjects negotiated for a total of 60 rounds. Dependent Measures We focus on two primary outcomes. First, power exercise is measured as the average profit earned during the final 10 rounds of exchange for each pair (Willer 1999). Second, we administered a computerized measure of pleasure/satisfaction and interest/excitement subsequent to the negotiations. Subjects reported their feelings about the negotiations along a series of bipolar adjectives that could range from ⫺30 to ⫹30. The pleasure/satisfaction index was based on the sum of scores for five items: displeased/pleased, unhappy/happy, not satisfied/ satisfied, not joyful/joyful, and discontented/contented. The interest/excitement index summed the following five items: not interesting/interesting, boring/excited, tired/energetic, unenthusiastic/enthusiastic, and unmotivating/motivating. These items are virtually identical to those used in previous research (Lawler, Thye, and Yoon 2000; Lawler and Yoon 1993, 1996, 1998) and consistently have displayed good measurement properties. RESULTS

First, we observed very large differences in power exercise between high (white male) and low (African-American female) status individuals. Across the two status-differentiated conditions, about 75% of the profit pool went to the higher status person. This level of power exercise is higher than the most comparable condition reported by Thye (2000) using age and education as the status manipulation. In that research, higher status subjects earned around 63% of the total profit, compared to the 75% reported here. The overall implication is that race and gender have especially powerful effects in producing inequality in exchange. This large impact of race and gender combined was unexpected, though a difference of some sort was anticipated. Most striking is the overall size of the status effect. Further,

CH03-P374095.indd 50

5/26/2008 2:30:18 PM

Social Exchange and the Maintenance of Order

51

even when the actors believe they are interacting with a partner who is equated to them on race and gender, a significant difference in profits still favors the while males. Overall, this suggests that African-American women and white men may negotiate differently, and these differences generate a baseline of empirical differences in power exercise (see Thye and Witkowski 2003 for details). Second, despite the large power differences, we observed no significant differences between high status and low status actors in self-reported feelings of pleasure/satisfaction and interest/excitement. For each index, a difference score was created by subtracting the low status actors’ score from the higher status actor’s score. The differences between high and low status scores, by condition (cj), were as follows: (for pleasure/satisfaction, c1 ⫽ 1.02, t ⫽ 0.22, two-tailed p ⫽ 0.83; c2 ⫽ 7.08, t ⫽ 0.97, two-tailed p ⫽ 0.35; c3 ⫽ ⫺3.64, t ⫽ ⫺0.53, two-tailed p ⫽ 0.60; for interest/excitement, c1 ⫽ 2.35, t ⫽ 0.39, two-tailed p ⫽ 0.70; c2 ⫽ 1.58, t ⫽ 0.34, two-tailed p ⫽ 0.74; c3 ⫽ ⫺0.76, t ⫽ ⫺0.09, two-tailed p ⫽ 0.92).2 Contrary to prior theories of emotion and research on exchange, those with less power did not report more negative feelings than those with more power (Kemper 1978, 1984; Lawler and Yoon 1993, 1996; Lovaglia and Houser 1996; Willer, Lovaglia, and Markovsky 1997). Moreover, the means for both measures of emotion, for both higher and lower status actors, are on the positive side of the negative-to-positive dimension. Third, low status individuals in conditions one and two (status activated) did not report significantly different levels of positive emotion than those low status individuals in condition three (status equated). For pleasure/satisfaction, the difference between conditions one and three is 1.04 (t ⫽ 0.14, two-tailed p ⫽ 0.89) and the difference between conditions two and three is 3.88 (t ⫽ 0.66, two-tailed p ⫽ 0.51). The same pattern exists for interest/excitement. Here, the difference between conditions one and three is 3.04 (t ⫽ 0.42, two-tailed p ⫽ 0.68) and the difference between conditions two and three is 0.24 (t ⫽ 0.04, two-tailed p ⫽ 0.96). The exact same trend holds for both emotions reported by high status actors across these conditions. Thus, there were no significant differences in emotion between status differentiated and status equated for either individual. The implication is that processes involving the exchange itself are creating mild positive emotional reactions, despite significant status and power differences. To further explore this, we conducted one final analysis. Consistent with the theory of relational cohesion, OLS regression analyses indicate that more frequent exchange increased positive feelings for both actors, along pleasure/ satisfaction and interest/excitement dimensions. Controlling for realized profit differences and experimental condition, we find that more frequent exchange produced more pleasure/satisfaction (high status  ⫽ 0.41, two-tailed p ⫽ 0.02; low status  ⫽ 0.70, two-tailed p ⬍ 0.001) and also more interest/excitement (high status  ⫽ 0.36, two-tailed p ⫽ 0.06; low status  ⫽ 0.51, two-tailed p ⬍ 0.01). 2 Each t-test is testing for a significant difference in emotional reaction between white males and African-American females.

CH03-P374095.indd 51

5/26/2008 2:30:18 PM

52

Social Structure and Emotion

Thus, repeated exchange in the context of status inequalities and strong power exercise generated a degree of stability and order by enhancing the satisfaction and interest in the exchange for both high and low status actors. THE PUZZLE

In sum, the anomaly in these empirical results boils down to the fact that race and gender produce highly differentiated rewards or outcomes for the actors, but yet they both expressed positive emotions and feelings about the results of exchange. The initial solution, also based on the results, is that repeated exchange with the same partner fosters a positive relationship between the two in spite of power exercise. This simple solution is consistent with the classic work of Homans (1951) on the effects of interaction frequency and interpersonal sentiment, the work of Collins (1981) on repeated conversations and emotional energy, and the theory of relational cohesion (e.g., Lawler and Yoon 1996) which links exchange frequency to emotion and cohesion. However, while this is a plausible interpretation it raises another important question: Why does repeated exchange or interaction have cohesive effects for actors highly stratified in terms of both status and power? What mechanisms are at work? The next section compares alternative answers or explanations to these questions.

THEORETICAL INTERPRETATIONS We assess and integrate a range of explanations based on social-exchange and expectation-states traditions. First, we briefly explore why neither tradition alone offers a satisfactory interpretation for the confluence of large profit differences and similar (consensual) positive emotions. Then we construct an explanation by examining the role of status value (Thye 2000), referential structures (Berger, Ridgeway, and Zelditch 2002), legitimation (Walker and Zelditch 1993), and joint tasks and shared responsibility (Lawler 2001). By interweaving these theoretical notions, we generate a more plausible interpretation than any one or a subset of them can provide. SHORTCOMINGS OF EACH THEORY

From fundamental notions of social-exchange theory, the high interdependence between low and high status actors in our experiment could account for the fact that they were likely to exchange frequently (roughly 70% of the time). Relational cohesion theory (Lawler and Yoon 1996) then could be used to account for the generation of positive emotions and feelings on the part of both actors. Yet, as we noted earlier, several studies in this tradition find low power actors to be less satisfied or to experience more negative emotions than high

CH03-P374095.indd 52

5/26/2008 2:30:18 PM

Social Exchange and the Maintenance of Order

53

power actors (Lawler and Yoon 1993; Willer, Lovaglia, and Markovsky 1997). Moreover, in our experiment, there were significant levels of non-exchange. Across the three conditions, approximately 32% of the possible exchanges did not occur, meaning that actors accepted zero profits. This is a relatively high rate of non-exchange given the high level of structural interdependence, and it may be due to the problems posed by status inequalities. Overall, central ideas about interdependence and emotion from exchange theorizing fall short of providing an adequate explanation for our findings. From status characteristics theory, the effects of status inequality on actualized power (profit) inequality could be explained by the fact that the experimental conditions make salient both gender and race, and the “burden of proof ” process renders them relevant to the task of exchange. The exchange task does not contain properties that objectively connect gender or race to task performance, and there are no other differences (e.g., in structural power). However, it is well-known that people accord meaning to whatever characteristics (or cues) distinguish them in the situation, and they form performance expectations based on these (Berger, Cohen, and Zelditch 1972; Ridgeway and Berger 1986). Given both stereotypical cultural beliefs associated with gender and race, and also the fact that women and African Americans have low power in the larger culture, it is reasonable to suppose that actors translate or connect low status and low power in the larger culture to status and power in the local situation (the experiment). This implies that referential structures are activated and acted upon in the exchange. Thus, while status theories help explain initial inequalities, they still are challenged to explain why positive affect is produced for the actors despite large differences in power use. We propose that elements of both social exchange and expectation states theories are necessary to account for the puzzle. Figure 3.1 portrays an integrative explanation in model form. There are three central points in the theoretical argument. First, the status differences activate dual processes that serve to legitimate an initial inequality in the system. One process is the spread of status value from a person to his or her resources (Thye 2000; Thye, Willer, and Markovsky 2006). The other process is the activation of a referential structure affirming that higher status individuals should earn more from exchange (see Figure 3.1). These dual forces validate the correspondence between status and power in the local encounter. Second, in the context of the spread of status value and activation of referential structures, small initial inequalities are repeated. This repetition alone can build a valued relationship between the actors by generating positive emotions and feelings. These emotions stem from repeated success at reaching negotiated exchanges and occur while initial inequalities are also repeatedly affirmed or validated. Third, in the context of repeated exchange, the “validity” of the inequality generates “propriety” and self-expectations consistent with the hierarchy. Overall, status processes generate initial profit differentiation and exchange processes produce positive relations despite that differentiation. Below we discuss further each of the theoretical links in the model (Figure 3.1).

CH03-P374095.indd 53

5/26/2008 2:30:18 PM

54

CH03-P374095.indd 54

SPREAD OF STATUS VALUE

INITIAL STATUS DIFFERENCES

EXCHANGE FREQUENCY-TOEMOTION-TO-COHESION

INITIAL INEQUALITIES THAT ARE REPEATED OVER TIME

FIGURE 3.1

VALIDITY-TO-SELF-EVALUATIONTO-PROPRIETY The effects of status in systems of exchange.

5/26/2008 2:30:18 PM

Social Structure and Emotion

REFERENTIAL STRUCTURES

Social Exchange and the Maintenance of Order

55

From Status to Initial Inequality First, Thye’s (2000) status value theory of power asserts that when resources are relevant to the positive states of status characteristics, those resources acquire consistent status value. For example, items owned by former presidents take on very high status value due to that association (Thye 2000). In the experimental situation described above, each subject was allocated poker chips on the basis of their characteristics and performance on a status-valued task. Therefore, the conditions are ripe for those chips to acquire status value consistent with the states of the status characteristics. That is to say, chips held by the white males may have been perceived as more honorary or valuable than those held by the AfricanAmerican females. If so, then the exchanges where greater numbers of chips from the “lower status” actor are traded for fewer “higher status” chips would appear as equal; and therefore negative emotions should be mitigated. As Thye and colleagues explain: It remains to be seen whether status-driven power produces similar negative emotions, but we suspect not. Status value leads to power exercise but, from the perspective of a low status actor buying a status valued good, the exchange does not appear unequal. To the contrary, because the exchange ratio appears as equipower there is no basis for negative reaction. The buyers of Michael Jordan shoes generally do not resent the retailer from which they were purchased. Similarly, collectors who pay high prices for status valued goods do so willingly because the goods are highly valued to them. (Thye, Willer, and Markovsky 2006:1488)

Thus, the inequalities observed in the experiment may be perceptually attenuated for the low status person, because each chip from the higher status individual has more “status value” than each chip from the lower status individual. At the same time, status characteristics can activate consistent referential structures (Berger et al. 1985; Thye, Willer, and Markovsky 2006; Thye and Witkowski 2003). A referential structure is defined as a set of cultural beliefs about the association between a status characteristic such as gender and goal objects like rewards or profits (Wagner and Berger 1993). In fact, reward expectations theory asserts that whenever resources are allocated (as in the exchange experiments described above) referential structures become salient and provide a point of reference for individuals. In this way, a referential structure can legitimate an association between status and power exercise. To the extent gender and race are salient, the culturally based referential structure is likely to shape how individuals perceive reward allocations. Cultural definitions of “what is” become “what is expected to be” in the local situation (Wagner and Berger 1993:38), and this may lead individuals to assume and act on consensual definitions of gender and race, regardless of the particular focus of the situation. It is still not entirely clear, however, why low status actors would feel as positive as high status actors and be accepting the inequality. Legitimacy theory provides a partial answer (Zelditch and Walker 1984).

CH03-P374095.indd 55

5/26/2008 2:30:18 PM

56

Social Structure and Emotion

From Inequality to Legitimacy Status value and referential structure effects can be construed as a legitimizing force, producing and reproducing local status hierarchies that mirror those of the larger culture. Zelditch and Walker’s (1984) theory of legitimacy includes two relevant dimensions of legitimacy: validity, and propriety. A social hierarchy is “valid” if actors treat norms, beliefs and procedures as taken for granted and matters of objective fact; that is, when actors accept that this is “the way things are” done. Expectations associated with different states of gender or race often exude the property of validity, that is, a collective kind of legitimacy that occurs when others accept something as “what is” (Wagner and Berger 1993; Weber [1918] 1968). Validity generates acceptance of the world which helps to explain the strong, persistent impact of referential structures as well as the resilience of gender and race effects in a variety of contexts. Propriety, on the other hand, refers to actors’ evaluation (positive–negative) of the structural hierarchy. The concept of propriety captures an evaluative dimension, defined as the degree that actors believe the structure and their place in it is correct, proper, or just (Dornbusch and Scott 1975). In other words, propriety describes an individual’s acceptance and support of the social structure, its norms and procedures as desirable and correct. Impropriety is a potential source of resistance and challenge to the existing hierarchy and, thus, a source of social change. One can think of propriety as legitimacy at the individual level, whereas validity is legitimacy at a collective level. Legitimacy theory asserts that validity can have effects on propriety, and related research has documented such validity-to-propriety effects (Ridgeway and Correll 2006; Walker and Zelditch 1993). Repeated exchange is a mechanism that could generate validity-to-propriety effects under conditions of status inequality, especially if it fosters positive emotions or feelings as we observe. We propose that, all things being equal, validity generates propriety if it is repeatedly experienced across time without challenge from the participants or bystanders in the situation (Ridgeway and Correll 2006; Ridgeway and Diekma 1989). In our experiment, which involved a dyad, the transformation from validity to propriety may stem from the simple repetition of inequality initiated by the spread of status value and supporting referential beliefs. Status value effects and referential beliefs may be sufficient to generate unequal profits in the first few exchange episodes, and only a few instances of unequal exchanges may be necessary to generate expectations of advantage/disadvantage in future exchanges. Berger and Luckmann (1966) ask how institutional patterns develop at the micro level, and they suggest that repetition of a pattern is the immediate or first step. Further, they argue that it does not take many repetitions – sometimes only one – to make that pattern a “reality” to actors and to produce incipient legitimation of it. In this context, the observed effect of exchange frequency on positive emotions for both actors may reflect repeated affirmations of a legitimate inequality. Affirming a legitimate inequality may have positive emotional consequences, or at least enable the exchange-to-emotion effects posited by relational cohesion theory to operate freely. Thus, inequality that is initially set in motion by the

CH03-P374095.indd 56

5/26/2008 2:30:18 PM

Social Exchange and the Maintenance of Order

57

spread of status value and supporting referential structures may promote micro order in part because acting in accord with this inequality generates consensual emotions and feelings. Social exchange is the interactional mechanism that drives this process. To more fully understand and explain the effects of social exchange, we turn to the affect theory of social exchange (see Lawler 2001, 2006; Lawler and Thye 2006). Exchange as the Mediating Link The affect theory of social exchange (Lawler 2001) builds on relational cohesion theory and explicitly theorizes when and how the emotions produced by exchange generate stronger and more cohesive relationships among actors. A key assumption is that social exchange inherently involves a joint task. In our experiment, consummating an exchange is a joint task that produces joint benefit. Jointness is important because it makes salient to actors that they cannot generate profits without cooperating or taking into account the other person in the situation (Lawler 2001). The affect theory of social exchange argues that in the context of joint tasks, individuals are likely to attribute their individual emotions to relational or group affiliations. This occurs to the degree that joint tasks generate a sense of shared responsibility for the results of social exchange (see Lawler 2001). Thus, if exchanges in our experiment generated a sense of shared responsibility, it is reasonable to infer that repeated exchange produces relational attributions of positive emotion and thereby strengthens the cohesion of these relations. The relation, as such, becomes important to actors, that is, an object of value in its own right. As a result, actors are more prepared to nurture and maintain the relationship and to make individual sacrifices on behalf of it (Lawler 2001). We can now integrate several strands of the above theoretical arguments. First, relational cohesion theory predicts the empirical results for emotion that we observe – namely, that more frequent exchange produces more positive emotions on the part of both low and high status actors. The affect theory of social exchange goes a step further and explains why positive feelings strengthen the exchange relation – namely, because repeated exchanges invoke a sense of shared responsibility, especially given the high degree of mutual dependence among actors in our experiment. While repeated exchange creates a relationship that has value, the negative emotions that normally accompany such inequality are attenuated because of supporting status value and referential structure effects. The combined impact of these parallel exchange and status processes are to mitigate the social order problem posed by power and status inequalities.

CONCLUSION We interpret a provocative experimental finding – that status-driven power does not elicit negative emotions – with two parallel, but interwoven processes. One involves an exchange-to-emotion-to-relation process through which social exchange builds cohesive and valued relations among actors. The sense of shared

CH03-P374095.indd 57

5/26/2008 2:30:19 PM

58

Social Structure and Emotion

responsibility for exchange is central here. The second involves a validity-toself-evaluation-to-propriety process in which legitimate inequalities come to be accepted and internalized by individuals. This leads actors to make selfevaluations that are consistent with profit differentiations and to act as if the profit differentiations are proper. These exchange and status processes converge as a relationship is constructed, as action is oriented to that relation, and as self-evaluations are adjusted to be consistent with both relative profits and their shared capacity to consummate exchanges. There are multiple components of our theoretical argument, summarized in Figure 3.1. First, status differences activate the spread of status value and a consistent referential structure. Second, because of status value and referentialstructure effects, initial inequalities emerge in social exchange, and these are experienced repeatedly by the actors. On the one hand, these repeated experiences validate (legitimize) the congruence of status/power differences in the local encounter and larger culture. On the other hand, repeated experiences of exchange give actors a sense of shared responsibility and promote positive emotions and feelings. Third, repeated exchange and repeated experience of inequalities promote both (a) emotion-to-cohesion effects from which valued relations develop between the actors and (b) validity-to-propriety effects for the hierarchy, from which both high and low status actors come to accept the inequality as proper. In this way status-based inequality may become stable and orderly. While our analysis approaches inequalities of power and status from socialexchange and expectation states theories, these also could be approached from the perspective of identity theory. For instance, identity control theory (Burke and Stets 1999; Stets and Burke 2002) presumes that individuals hold identity standards (or meanings) for identities like male or African American. When the self-in-situation meanings (or transient impressions) verify these identity standard meanings, then positive emotions are predicted to be the result. As such, if it is the case that behaviors such as “competing,” “acquiring,” or “negotiating” match the identity standard associated with white or male then one would expect positive emotions. At the same time, if behaviors such as “agreeing,” “making concessions,” or “exchanging” match the identity standard associated with African American or female, one would again expect positive emotions. This is an intriguing possibility that should be the target of future research along with our integration of social-exchange and status theories. In closing, social-exchange and expectation states theories are among the most successful programs of theory and research in sociological social psychology. Each has clear strengths and each has a somewhat different focus – i.e., expectations, status, and influence versus incentives, networks, and power. Several theoretical and research efforts in recent years have created interesting and important bridges among these traditions while preserving the unique strengths of each (Lovaglia and Houser 1996; Thye 1999, 2000; Thye, Willer, and Markovsky 2006). This paper adds to that bridging literature. We take an

CH03-P374095.indd 58

5/26/2008 2:30:19 PM

Social Exchange and the Maintenance of Order

59

important phenomenon – the capacity of a group, organization, or society to produce and reproduce an acceptance of widely disparate allocations of status, power, and wealth – and try to explain such acceptance based on extant theories. In a world with growing inequalities between those who have status and power and those who do not, this is a timely problem to address. The unexpected results of a recent experiment fortuitously provided us with the incentive to theorize the problem. In our theoretical analysis we found status and exchange theories, singly, to be less satisfactory as explanations than a more complicated formulation combining elements from each tradition. While the resulting explanation is indeed complicated, the core message is quite straightforward: Inequalities are initiated and legitimized through status processes, while valued relations are generated through social-exchange processes. In this context, both advantaged and disadvantaged individuals act in accord with, but also come to believe in and accept, the social inequalities they encounter in their everyday lives.

ACKNOWLEDGMENTS This research was supported by a Research and Productive Scholarship Grant at the University of South Carolina to Shane R. Thye. The order of authorship is random and does not reflect differential contributions.

REFERENCES Bales, Robert F. 1950. Interaction Process Analysis: A Method for the Study of Small Groups. Cambridge, MA: Addison-Wesley Press. Bales, Robert F., Fred L. Strodtbeck, Theodore M. Mills, and Mary E. Roseborough. 1951. “Channels of Communication in Small Groups.” American Sociological Review 16:461–468. Berger, Joseph, Bernard P. Cohen, and Morris Zelditch, Jr.. 1972. “Status Characteristics and Social Interaction.” American Sociological Review 37:241–255. Berger, Joseph, Thomas L. Conner, and M. Hamit Fisek (eds). 1974. Expectation States Theory: A Theoretical Research Program. Cambridge, MA: Winthrop. Berger, Joseph, M. Hamit Fisek, and Lee Freese. 1976. “Paths of Relevance and the Determination of Power and Prestige Orders.” Pacific Sociological Review 19:45–62. Berger, Joseph, M. Hamit Fisek, Robert Z. Norman, and Morris Zelditch, Jr.. 1977. Status Characteristics and Social Interaction. New York: Elsevier. Berger, Joseph, M. Hamit Fisek, Robert Z. Norman, and David G. Wagner. 1985. “Formation of Reward Expectations in Status Situations.” In: Joseph Berger and M. Zelditch, Jr. (eds), Status, Rewards, and Influence, pp. 215–261. San Fransisco: Jossey-Bass. Berger, Joseph, Cecilia Ridgeway, and Morris Zelditch. 2002. “Construction of Status and Referential Structures.” Sociological Theory 20(2):157–179. Berger, Peter L. and Thomas Luckmann. 1966. The Social Construction of Reality. New York: Doubleday. Burke, Peter J. and Jan E. Stets. 1999. “Trust and Commitment Through Self-Verification.” Social Psychology Quarterly 62:347–366.

CH03-P374095.indd 59

5/26/2008 2:30:19 PM

60

Social Structure and Emotion

Collins, Randall. 1981. “On the Microfoundations of Macrosociology.” American Journal of Sociology 86:984–1014. Collins, Randall. 1989. “Toward a Neo-Meadian Sociology of Mind.” Symbolic Interaction 12:1–32. Collins, Randall. 2004. Interaction Ritual Chains. Princeton, NJ: Princeton University Press. Cook, Karen S. and Richard M. Emerson. 1978. “Power, Equity and Commitment in Exchange Networks: Theory and Experimental Results.” American Sociological Review 43:721–739. Cook, Karen S., Linda D. Molm, and Toshio Yamagishi. 1993. “Exchange Relations and Exchange Networks: Recent Developments in Social Exchange Theory.” In: Joseph Berger and Morris Zelditch, Jr. (eds), Theoretical Research Programs: Studies in the Growth of Theory, pp. 296–322. Stanford, CA: Stanford University Press. Cook, Karen S., Toshio Yamagishi, Coye Cheshire, Robin Cooper, Masafumi Matsuda, and Rie Mashima. 2005. “Trust Building via Risk Taking: A Cross-Societal Experiment.” Social Psychology Quarterly 68:121–142. Dahrendorf, Ralf. 1959. Class and Class Conflict in Industrial Society. Stanford, CA: Stanford University Press. Damasio, Antonio. 1999. The Feeling of What Happens. New York: Harcourt Brace. Davis, Kingsley and Wilbert E. Moore. 1945. “Some Principles of Stratification.” American Sociological Review 42:639–651. Della Fave, L. Richard. 1980. “The Meek Shall Not Inherit the Earth: Self-Evaluation and the Legitimacy of Stratification.” American Sociological Review 45:955–971. Dornbusch, Sanford M. and W. Richard Scott. 1975. Evaluation and the Exercise of Authority. San Francisco: Jossey-Bass. Durkheim, Émile. [1933] 1964. The Division of Labor in Society. Glencoe, IL: Free Press. Emerson, Richard M. 1962. “Power-Dependence Relations.” American Sociological Review 27:31–40. Emerson, Richard M. 1964. “Power-Dependence Relations: Two Experiments.” Sociometry 27:282–298. Emerson, Richard M. 1972a. “Exchange Theory, Part I: A Psychological Basis for Social Exchange.” In: Joseph Berger, Morris Zelditch, Jr., and Bo Anderson (eds), Sociological Theories in Progress, Vol. 2, pp. 38–57. Boston: Houghton-Mifflin. Emerson, Richard M. 1972b. “Exchange Theory, Part II: Exchange Relations and Networks.” In: Joseph. Berger, Morris Zelditch, and Bo Anderson (eds), Sociological Theories in Progress, Vol. 2, pp. 58–87. Boston: Houghton-Mifflin. Emerson, Richard M. 1981. “Social Exchange Theory.” In: M. Rosenberg and R. H. Turner (eds), Social Psychology: Sociological Perspectives, pp. 30–65. New York: Basic. Habermas, Jürgen. 1973. Legitimation Crisis. Boston: Beacon. Homans, George L. 1951. Human Group. New Brunswick, NJ: Transaction Publishers. Homans, George L. 1974. Social Behavior: Its Elementary Forms. New York: Harcourt Brace Jovanovich. Izard, Carroll E. 1991. The Psychology of Emotion. New York: Plenum Press. Kemper, Theodore D. 1978. A Social Interactional Theory of Emotions. New York: Wiley. Kemper, Theodore D. 1984. “Power, Status, and Emotions: A Sociological Contribution to a Psychophysiological Domain.” In: Klaus R. Scherer and Paul Ekman (eds), Approaches to Emotion, pp. 369–383. Hillsdale, NJ: Erlbaum. Kemper, Theodore D. and Randall Collins. 1990. “Dimensions of Microinteraction.” American Journal of Sociology 96(1):32–68. Kipnis, David. 1976. The Powerholders. Chicago: University of Chicago Press. Kollock, Peter. 1994. “The Emergence of Exchange Structures: An Experimental Study of Uncertainty, Commitment, and Trust.” American Journal of Sociology 100:315–345. Kollock, Peter. 1999. “The Production of Trust in Online Markets.” In: Shane R. Thye, Edward J. Lawler, Michael W. Macy, and Henry A. Walker (eds), Advances in Group Process, Vol. 16. Stamford, Conn: JAI Press. Lawler, Edward J. 1992. “Power Processes in Bargaining.” The Sociological Quarterly 33:17–34.

CH03-P374095.indd 60

5/26/2008 2:30:19 PM

Social Exchange and the Maintenance of Order

61

Lawler, Edward J. 2001. “An Affect Theory of Social Exchange.” American Journal of Sociology 107:321–352. Lawler, Edward J. 2002. “Micro Social Orders.” Social Psychology Quarterly 65:4–17. Lawler, Edward J. 2006. “Exchange, Affect, and Group Relations.” In: Javier Trevino (ed.), George C. Homans: History, Theory, and Method. Boulder, CO: Paradigm Publishers. Lawler, Edward J., Cecilia Ridgeway, and Barry Markovsky. 1993. “Structural Social Psychology and the Micro-macro Problem.” Sociological Theory 11:268–290. Lawler, Edward J. and Shane R. Thye. 1999. “Bringing Emotions into Social Exchange Theory.” Annual Review of Sociology 25:217–244. Lawler, Edward J. and Shane R. Thye. 2006. “Social Exchange Theory of Emotions.” In: Jan E. Stets and Jon H. Turner (eds), Handbook of the Sociology of Emotions, pp. 295–320. New York: Springer. Lawler, Edward J., Shane R. Thye, and Jeongkoo Yoon. 2000. “Emotion and Group Cohesion in Productive Exchange.” American Journal of Sociology 106(3):616–657. Lawler, Edward J., Shane R. Thye, and Jeongkoo Yoon. 2006. “Commitment in Structurally Enabled and Induced Exchange Relations.” Social Psychology Quarterly 69(2):183–200. Lawler, Edward J. and Jeongkoo Yoon. 1993. “Power and the Emergence of Commitment Behavior in Negotiated Exchange.” American Sociological Review 58:465–481. Lawler, Edward J. and Jeongkoo Yoon. 1996. “Commitment in Exchange Relations: Test of a Theory of Relational Cohesion.” American Sociological Review 61:89–108. Lawler, Edward J. and Jeongkoo Yoon. 1998. “Network Structure and Emotion in Exchange Relations.” American Sociological Review 63:871–894. Lenski, Gerhard E. 1966. Power and Privilege. New York: McGraw-Hill. Lovaglia, Michael J. 1994. “Relating Power to Status.” Advances in Group Processes 11:87–111. Lovaglia, Michael J. 1995. “Power and Status: Exchange, Attribution, and Expectation States.” Small Group Research 26(3):400–426. Lovaglia, Michael J. and Jeffrey A. Houser. 1996. “Emotional Reactions and Status in Groups.” American Sociological Review 61:867–883. Lovaglia, Michael J., John Skvoretz, David Willer, and Barry Markovsky. 1995. “Negotiated Outcomes in Social Exchange Networks.” Social Forces 74(1):123–155. Lovaglia, Michael J., Robb Willer, and Lisa Troyer. 2003. “Power, Status, and Collective Action: Developing Fundamental Theories to Address a Substantive Problem.” In: Shane R. Thye and John Skvoretz (eds), Advances in Group Processes, Vol. 20, pp. 105–131. Elsevier. Markovsky, Barry, John Skvoretz, David Willer, Michael J. Lovaglia, and Jeffery Erger. 1993. “The Seeds of Weak Power: An Extension of Network Exchange Theory.” American Sociological Review 58:197–209. Markovsky, Barry, LeRoy F. Smith, and Joseph Berger. 1984. “Do Status Interventions Persist?” American Sociological Review 49:373–382. Markovsky, Barry, David Willer, and Travis Patton. 1988. “Power Relations in Exchange Networks.” American Sociological Review 53:220–236. Marx, Karl and Frederick Engels. 1959. “Manifesto of the Communist Party.” In: L. S. Feuer (ed.), Marx and Engels: Basic Writings on Politics and Philosophy. Garden City: Doubleday. Molm, Linda. 2003. “Power, Trust, and Fairness: Comparisons of Negotiated and Reciprocal Exchange.” In: Shane Thye and John Skvoretz (eds), Advances in Group Processes, Vol. 20, pp. 31–65. London: Elsevier. Molm, Linda and Karen Cook. 1995. “Social Exchange and Exchange Networks.” In: K. Cook, G. Fine, and J. House (eds), Sociological Perspectives on Social Psychology, pp. 209–235. Boston: Allyn and Bacon. Parsons, Talcott. 1949. “An Analytic Approach to the Theory of Social Stratification.” In: Talcott Parsons (ed.), Essays in Sociological Theory. New York: Free Press. Parsons, Talcott. 1951. The Social System. New York: Free Press. Parsons, Talcott. 1953. “A Revised Analytic Approach to the Theory of Social Stratification.” In: Reinhard Bendix and Seymour Martin Lipset (eds), Class, Status, and Power. New York: Free Press.

CH03-P374095.indd 61

5/26/2008 2:30:19 PM

62

Social Structure and Emotion

Pareto, Vilfredo. 1935. The Mind and Society: A Treatise on General Sociology. New York: Harcourt, Brace. Ridgeway, Cecilia L. 1991. “The Social Construction of Status Value: Gender and Other Nominal Characteristics.” Social Forces 70(2):376–386. Ridgeway, Cecilia L. 2000. “The Formation of Status Beliefs: Improving Status Construction Theory.” In: Shane R. Thye, Edward J. Lawler, Michael W. Macy, and Henry A. Walker (eds), Advances in Group Processes, Vol. 17, pp. 77–102. Ridgeway, Cecilia L. and Joseph M. Berger. 1986. “Expectations, Legitimation, and Dominance Behavior in Task Groups.” American Sociological Review 51:603–617. Ridgeway, Cecilia L. and Shelley J. Correll. 2006. “Consensus and the Creation of Status Beliefs.” Social Forces 85(1):431–453. Ridgeway, Cecilia L. and David Diekma. 1989. “Dominance and Collective Hierarchy Formation.” American Sociological Review 54:79–93. Ridgeway, Cecilia L. and Kristan Glasgow Erikson. 2000. “Creating and Spreading Status Beliefs.” American Journal of Sociology 106(3):579–615. Ridgeway, Cecilia L. and Cathryn Johnson. 1990. “What is the Relationship Between Socioemotional Behavior and Status in TASK Groups.” American Journal of Sociology 95:1189–1212. Ridgeway, Cecilia L. and Henry A. Walker. 1995. “Status Structures.” In: Karen S. Cook, Gary Alan Fine, and James S. House (eds), Sociological Perspectives on Social Psychology, pp. 281–310. Boston: Allyn and Bacon. Smith-Lovin, Lynn. 1995. “The Sociology of Affect and Emotion.” In: Karen S. Cook, Gary Alan Fine, and James S. House (eds), Sociological Perspectives on Social Psychology, pp. 118–148. Boston: Allyn and Bacon. Smith-Lovin, Lynn and David R. Heise. (eds.) 1988. Analyzing Social Interaction: Advances in Affect Control Theory. New York: Gordon and Breach. Stets, Jan E. and Peter J. Burke. 2002. “A Sociological Approach to Self and Identity.” In: M. R. Leary and J. P. Tangney (eds), Handbook of Self and Identity, pp. 128–152. New York: Guilford Press. Thye, Shane R. 1999. “Status Influence and Status Value.” In: David Willer (ed.), Network Exchange Theory, pp. 248–255. Westport, CT: Praeger. Thye, Shane R. 2000. “A Status Value Theory of Power in Exchange Relations.” American Sociological Review 65(3):407–432. Thye, Shane R., Michael J. Lovaglia, and Barry Markovsky. 1997. “Responses to Social Exchange and Social Exclusion in Networks.” Social Forces 75(3):1031–1047. Thye, Shane, David Willer, and Barry Markovsky. 2006. “From Status to Power: New Models at the Intersection of Two Theories.” Social Forces 84(3):1471–1495. Thye, Shane and Christine Witkowski. 2003. “The Status Value Theory of Power: The Effect of Status and Resource Differentiation on Power in Exchange.” Presented at the Annual Meetings of the American Sociological Association, Atlanta, GA. Thye, Shane R., Jeongkoo Yoon, and Edward J. Lawler. 2002. “The Theory of Relational Cohesion: Review of a Research Program.” In: Shane R. Thye and Edward J. Lawler (eds), Advances in Group Process, Vol. 19. Oxford: Elsevier. Wagner, David G. and Joseph Berger. 1993. “Status Characteristics Theory: The Growth of a Program.” In: Joseph Berger and Morris Zelditch, Jr. (eds), Theoretical Research Programs: Studies in the Growth of Theory, pp. 23–63. Stanford, CA: Stanford University Press. Wagner, David G. and Joseph Berger. 2002. “Expectation States Theory: An Evolving Research Program.” In: Joseph Berger and Morris Zelditch, Jr. (eds), New Directions in Contemporary Sociological Theory, pp. 41–76. New York: Rowman and Littlefield. Walker, Henry A. and Morris Zelditch. 1993. “Power, Legitimacy, and the Stability of Authority: A Theoretical Research Program.” In: Joseph Berger and Morris Zelditch, Jr. (eds), Theoretical Research Programs: Studies in the Growth of Theory, pp. 364–381. Stanford, CA: Stanford University Press.

CH03-P374095.indd 62

5/26/2008 2:30:19 PM

Social Exchange and the Maintenance of Order

63

Weber, Max. [1918] 1968. Economy and Society. Berkeley: University of California Press. Weber, Max. [1916] 1946. From Max Weber: Essays in Sociology. Translated and edited by Hans H. Gerth and C. Wright Mills. New York: Oxford University Press. Willer, David. 1999. Network Exchange Theory. Westport CT: Praeger. Willer, David, Michael J. Lovaglia, and Barry Markovsky. 1997. “Power and Influence: A Theoretical Bridge.” Social Forces 76(2):571–603. Yamagishi, Toshio, Karen S. Cook, and Motoki Watabe. 1998. “Uncertainty, Trust and Commitment Formation in the United States and Japan.” American Journal of Sociology 104(1):165–194. Yamagishi, Toshio and Midori Yamagishi. 1994. “Trust and Commitment in the United States and Japan.” Motivation and Emotion 18:129–166. Yoon, Jeongkoo and Edward J. Lawler. 2006. “Relational Cohesion Model of Organizational Commitment.” In: Olympia Kyriakidou and Mustafa Ozbilgin (eds), Relational Perspectives in Organizational Studies. Cheltenham, UK: Edward Elgar Publishers. Zelditch, Morris and Henry A. Walker. 1984. “Legitimacy and the Stability of Authority.” In: Edward J. Lawler (ed.), Advances in Group Processes, Vol. 1, pp. 1–25. Greenwich, CT: JA.

CH03-P374095.indd 63

5/26/2008 2:30:19 PM

PA R T

II Emotion and Identity Work

CH04-P374095.indd 65

5/26/2008 2:32:58 PM

4 Commentary Dawn T. Robinson University of Georgia, Athens, GA, USA

The relationship between identity and emotion is one that has received considerable theoretical and empirical attention in the last several decades. An emerging literature has organized around a fairly consensual understanding of the self as a control system comprised of various identities that give it meaning and of emotions arising out of this control process to energize social action. The reciprocal relationship between self, identity, and emotion is a nearly universal element in identity theories. Social Identity Theory (Tajfel and Turner 1986) was developed by European psychologists to describe the relationship between social structure and the self-concept. This theory focuses on the meanings afforded to the self by identification with various groups. According to Social Identity theory, positive emotion arises from group identification – especially with culturally valued categories. Identity Theory, developed by Stryker (1980) as a blend of core ideas from role theory (e.g., Merton 1957) and symbolic interaction (e.g., Mead 1934), also conceptualizes the self as comprising a set of identities that give it meaning. According to Identity Theory, these identities are linked to social positions and are arranged within the self in a salience hierarchy which reflects the importance of each identity to the self-concept. Social relationships create varying degrees of dependence on performance of different identities. Identities whose performances impact more, or more valued, relationships are driven upward in the salience hierarchy. Identities with higher salience have greater potential for enactment across situations. Focusing on valued social identities, Stryker (1987, 2004) argued that positive emotions arise from successful performance of roles, while negative emotion arises from failure to meet role expectations. This argument closely relates to Goffman’s ideas about emotions arising from the dramaturgical process of impression management. According to Goffman (1959), Social Structure and Emotion

CH04-P374095.indd 67

67

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 2:32:58 PM

68

Social Structure and Emotion

negative emotions – most notably, embarrassment – arise from failures in selfpresentation. While Goffman did not offer a highly developed theory about the relationship between emotion and identity, he used his dramaturgical metaphor to describe the emotion dynamics associated with playing tightly scripted cultural roles. His work on the importance of backstage activities to support frontstage performances, became fundamental to Hochschild’s theory of emotion management (see Moore, this volume), who expanded the acting metaphor to include both “superficial” and “deep” performances to comply with expression norms and feeling norms, respectively. The idea that emotions arise as the result of successful – and unsuccessful – performances of identities is key to the two most developed, formal theories of identity and emotion in contemporary sociology, as well. The two major theories of emotion in the more quantitative, structural branch of symbolic interaction are Affect Control Theory (Heise 1979, 2007; MacKinnon 1994; Smith-Lovin and Heise 1988) and Identity Control Theory (Burke 1991, 2004; Burke and Reitzes 1991). Both theories emphasize the central role of identity meanings and the ability of social interactions to sustain them in producing emotional responses. Both theories share a common basis in Powers’ (1973) cybernetic control model of perception and so have as their central mechanism the comparison between a reference signal (a culturally established desired set of meanings) and input from the situational context in which the actor is embedded (meanings generated from local interactions). In Affect Control Theory, the reference signal meanings are fundamental sentiments acquired through past interactions, exposure to cultural materials, and other sources. These sentiments have cognitive and affective components, which are inseparably evoked when self, other interactants and actions are labeled in the definition of a situation (MacKinnon 1994). Affect Control Theory uses as its metric the three dimensions of affective meanings based on the work of Charles Osgood and his colleagues (Osgood et al. 1957, 1975) – evaluation (goodness versus badness), potency (powerfulness versus powerlessness), and activation (lively versus quiet). Identity Control Theory presumes a somewhat less cultural and more personalized set of self-identity meanings that are arranged in a salience hierarchy within a stable self-structure (Burke 1991; Stryker 1980). The relevant number and type dimensions can vary, making assessment of the reference signal is more domain specific in Identity Control Theory. Both control theories presume that actors operate to maintain reference meanings within a situation. In Identity Control Theory, the self is a distinct control system with individuals behaving to control their own self-meanings. The perceiver is responsible for controlling his or her own identity standards. Individuals may simultaneously be controlling sets of hierarchically nested identity standards, but the highest level in the hierarchy is still in the self (Tsushima and Burke 1999). In Affect Control Theory, the interactional event is the control system, with individual actors striving to maintain the set of meanings invoked by situation as a whole (Smith-Lovin and Robinson 2006).

CH04-P374095.indd 68

5/26/2008 2:32:58 PM

Commentary

69

In Affect Control Theory, the actor processes an event of the general form “the actor does behavior to the object person within this setting” to produce a set of situated impressions for the actor, behavior, object–person and setting, which are then compared to the fundamental sentiments associated with those event elements. This comparison process results in an assessment of deflection, which is the amount of discrepancy between the impressions and the sentiments (Heise 1979). The actor is then motivated to create new behaviors that realign the situated impressions with the more stable sentiments. In Identity Control Theory, the reference level is an identity standard (a stored set of self-relevant meanings) that serves as a guide for the actors behavior. When reflected appraisals fail to confirm identity standards, this disconfirmation creates a discrepancy leading directly to both behavior and emotion in the theory. Specifically, when reflected appraisals do not support identity meanings, actors experience negative emotion and generate new actions to restore meanings (Burke 1991). In Affect Control Theory, however, emotions are more closely related to impressions than to deflection. Thus, deflecting events are only predicted to produce negative emotion when the events themselves are negative, such as an unsuccessful performance of a culturally valued identity. On the other hand, Affect Control Theory says that when events are too positive, actors may feel great even as they engage in efforts that realign the meanings. Confirming events, then, can also produce either positive or negative emotion in Affect Control Theory. An actor will feel positive emotions – pride, satisfaction, warmth – when confirming socially valued identities, and negative emotions – embarrassment, guilt, depression – when confirming stigmatized identities. The two theories are highly parallel, with the primary differences being (1) reliance on a stable set of dimensions (Affect Control Theory) or a more domain-specific measurement scheme (Identity Control Theory), (2) the location of the control process in the self (Identity Control Theory) or the event (Affect Control Theory), and (3) the mechanism underlying emotion production. Consequently, the predictions of the two theories largely overlap as well. Numerous studies support the basic control process predictions employed by these theories (e.g., Burke 1991, 2004; Burke & Cast 1997; Burke and Stets 1999; Riley and Burke 1995; Robinson and Smith-Lovin 1992; Stets and Carter 2006; Swann et al. 1987). Empirical research in the Identity Control Theory tradition also provides support for the notion of hierarchy in identity control processes (e.g., Tsushima and Burke 1999). Since predictions about emotional response are one of the few places where the control theories differ (Smith-Lovin and Robinson 2006), there has been some recent empirical attention paid to this question – with mixed results so far. Robinson and Smith-Lovin (1992) showed that deflections in a positive direction produced positive emotion, even when experimental participants acted to reduce that positive deflection by choosing a interaction partner that had evaluated them negatively (consistent with their negative self-image). Experimental research by Stets (2003, 2005) also supports Affect Control Theory predictions, by showing

CH04-P374095.indd 69

5/26/2008 2:32:58 PM

70

Social Structure and Emotion

that deflections in a positive direction (over-reward) create positive rather than negative emotion. On the other hand, longitudinal research of newly married couples by Burke and Harrod (2005) found that individuals whose views of their own identities were highly discrepant from their perceptions of their partners views of their identities experience more negative emotion. A series of studies in the Affect Control Theory tradition demonstrates how observers use emotional expressions to infer identity information about the emoter – and others – in the situation. The theory presumes that emotion displays reflect information about the confirmation/disconfirmation of the emoter’s definition of the situation (Smith-Lovin 1990). Under the theory, then, emotions can be used by observers to glean information about the identities occupied by actors in a situation – and about the meanings of observed behaviors. Robinson and Smith-Lovin (1999) used vignette studies to investigate how emotion displays imbedded in transcripts of self-disclosures affected readers’ reports about the identities of the person telling the story. Consistent with Affect Control Theory predictions (Heise 1989), they found that when describing charitable behaviors actors were evaluated more positively when they displayed positive emotion, while when describing harmful behaviors actors were evaluated more positively when they displayed negative emotion. Robinson, Smith-Lovin and Tsoudis (1994) applied this to evaluations of criminal behavior, showing that readers of mock criminal confessions recommended longer sentences for confessors who appeared calm than for those who showed remorse while describing their crime. Tsoudis and Smith-Lovin (1998, 2001) and Tsoudis (2000) extended this logic to show how, consistent with Affect Control Theory predictions, the emotions displays of a victim also impact our impressions of a criminal. Emotion recognition accuracy is generally higher within racial and cultural groups than between racial and cultural groups (for meta-analysis, see Elfenbein and Embady 2002). Meanwhile judges and probation officers are predominantly white. These effects are consistent with research revealing greater resistance of probation officers and federal judges to recommending or awarding remorse-based sentence reductions to members of ethnic and racial minority groups (Everett and Neinstedt 1999). Affect control-based research on the relationship between identity and emotion has other clear structural implications. Affect Control Theory’s emotion equations predict that people occupying high positive, powerful identities will tend to experience more positive emotion, and less negative emotion, on average (Heise 1989; Smith-Lovin 1990). Recent research based on nationally representative data from the US offers some empirical support for this generalization (Lively and Heise 2002; Simon and Nath 2004), as does a representative sample of residents of Toronto, Canada (Turner, Wheaton and Lloyd 1995). Identity Control Theory, with its roots in Identity Theory, also has a clear structural component with its emphasis on positions and relationships as the origins of identities and as determinants of their relative salience. One Identity Control Theory based implication of this is that identities activated in domains

CH04-P374095.indd 70

5/26/2008 2:32:58 PM

Commentary

71

with closer, more intense, personal relationships (like families) should lead to identities being controlled more tightly and produce more intense negative emotions when those identities are not verified (Burke 1991), compared to identities activated within domains or institutions with less intense relationships (e.g., workplaces, organizations). Stets and Tsushima (2001) find evidence of this using national representative data. In particular, they found that people reported experiencing more intense feelings of anger when their family identities were disconfirmed than when occupying worker identities were disconfirmed. The first chapter in this section, by Peter Burke, emphasizes another structural force on identity process – access to resources. This exciting advance Identity Control Theory (Burke 2004; Freese and Burke 2004) leads our thinking about the relationship between social structure, identity, and emotion in important new directions. According to Freese and Burke (1994) system structures can be thought of as the configurations of resource flows, with the fundamental goal of interaction being to connect resource flows, creating that configuration. Burke (this volume) argues that access to resources (both actual and potential) makes the self-verification process more successful and also alleviates some of the negative consequences of identity disconfirmation. He tests these ideas using a longitudinal study of couples in their first years of marriage, focusing on the correspondence between respondents’ self-impressions and their spouses’ expectations for them as a measures of self-verification. Following the assumption that high status positions generate access to resources and that resources, in turn, often signal status, Burke treats social status as a proxy for resources in this investigation. His findings are consistent with past research about the relationship between status and emotion described earlier in this comment and also largely support these new ideas about resources (here, status) as protection against the negative emotional consequences of verification failure. The chapter in this section by Francis and Heise harkens back to Goffman’s dramaturgical theory about the relationship between impression management and emotion. Although these authors did not make use of Affect Control Theory or its equations in this chapter, they did exploit the dimensions of meaning (evaluation, potency, activity) traditionally used in Affect Control Theory research. They focus on two types of “face” concerns: moral character and social status and investigate their implications for emotional experience – which they measure in evaluation, potency and activity. In a fairly complex vignette experiment, they find general support for Goffman’s arguments about the relationship between threats to face and negative emotion. More interestingly, they find that these two different types of face threat (moral character versus social status) produce different kinds of negative emotions. In particular, and consistent with Affect Control Theory would predict, threats to moral face (more evaluation related) yielded somewhat more unpleasant emotions than threats to status face (more potency related). The final chapter in this section is Lisa Troyer’s fascinating proposal for how to exploit the rich theoretical knowledge we already have about the relationship between identity and emotion for creating a model of social intelligence for

CH04-P374095.indd 71

5/26/2008 2:32:58 PM

72

Social Structure and Emotion

artificial (or virtual) actors. She points out that considerable attention and resources have been directed toward developing artificial actors with vast analytic intelligence and physical sophistication. Yet, little attention as been paid to developing artificial actors whose social behaviors are meaningful and normative. She makes the case for using the apparatus of Affect Control Theory as the architecture for such an enterprise. She further argues for the utility of bringing in key aspects of Identity Theory (especially identity commitment) to provide a model of identity activation, and bringing in ideas about social risk to highlight the importance of social realism in shaping emotional and behavioral responses in virtual environments. These three papers could hardly take more different approaches to expanding our knowledge about the relationship between social structure, identity, and emotion. Collectively, they offer a rich set of new theoretical ideas, new empirical findings, and radical new strategies for making use of old ideas. Burke’s findings about the relationship between status, identity maintenance, and emotion, along with his findings about the differential effects of different statuses (e.g., education versus race) on identity maintenance and emotion, together with Francis’ and Heise’s findings about the different implications of different kinds of identity threat (moral face versus status/competence), suggest that we need some more nuanced theorizing and investigation of status processes as they relate to identity and emotion. Finally, Troyer’s woxrk opens the door to a bold new world of possibilities – both for investigating these nuances and for putting to good use the store of theoretical knowledge we have already developed.

REFERENCES Burke, Peter J. 1991. “Identity Processes and Social Stress.” American Sociological Review 56:836–849. Burke, Peter J. 2004. “Identities and Social Structure: The 2003 Cooley-mead Award Address.” Social Psychology Quarterly 67:5–15. Burke, Peter J. and Alicia D. Cast. 1997. “Stability and Change in the Gender Identities of Newly Married Couples.” Social Psychology Quarterly 60(4):277–290. Burke, Peter J. and Michael M. Harrod. 2005. “Too Much of a Good Thing.” Social Psychology Quarterly 68:359–374. Burke, Peter J. and Donald C. Reitzes. 1991. “An Identity Approach to Commitment.” Social Psychology Quarterly 54:239–251. Burke, Peter J. and Jan E. Stets. 1999. “Trust and Commitment Through Self-verification.” Social Psychology Quarterly 62:347–366. Cast, Alicia D. and Peter J. Burke. 2002. “A Theory of Self-esteem.” Social Forces 80:1041–1068. Cast, Alicia D., Jan E. Stets, and Peter J. Burke. 1999. “Does the Self Conform to the Views of Others?.” Social Psychology Quarterly 62:68–82. Elfenbein, Hillary Anger and Nalini Embady. 2002. “On the Universality and Cultural Specificity of Emotion Recognition: A Meta-Analysis.” Psychological Bulletin 128(2):203–235. Everett, Ronald and Barbara Nienstedt. 1999. “Race, Remorse, and Sentence Reduction: Is Saying You’re Sorry Enough?” Justice Quarterly 16:99–122.

CH04-P374095.indd 72

5/26/2008 2:32:58 PM

Commentary

73

Freese, Lee and Peter J. Burke. 1994. “Persons, Identities, and Social Interaction.” Advances in Group Processes 11:1–24. Goffman, Erving. 1959. The Presentation of Self in Everyday Life. New York: Doubleday. Goffman, Erving. 1967. Interaction Ritual: Essays on Face-to-face Behavior. New York: Anchor Books. Heise, David R. 1979. Understanding Events: Affect and the Construction of Social Action. New York: Cambridge University Press. Heise, David R. 1989. “Effects of Emotion Displays on Social Identification.” Social Psychology Quarterly 52(1):10–21. Heise, David R. 2007. Expressive Order: Confirming Sentiments in Social Actions. New York: Springer. Heise, David R. and Lisa Thomas. 1989. “Predicting Impressions Created by Combinations of Emotion and Social Identity.” Social Psychology Quarterly 52:141–148. Mackinnon, Neil J. 1994. Symbolic Interactionism as Affect Control. Albany: State University of New York Press. Mead, George Herbert. 1934. Mind, Self, and Society. Chicago: University of Chicago Press. Merton, Robert K. 1957. Social Theory and Social Structure. New York: Free Press. Osgood, Charles E., William H. May, and Murray S. Miron. 1975. Cross-cultural Universals of Affective Meaning. Urbana: University of Illinois Press. Osgood, Charles E., George J. Suci, and Percy H. Tannenbaum. 1957. The Measurement of Meaning. Urbana: University of Illinois Press. Powers, William T. 1973. Behavior: the Control of Perception. Chicago: Aldine. Riley, Ann and Peter J. Burke. 1995. “Identities and Verification in the Small Group.” Social Psychology Quarterly 58(2):61–73. Robinson, Dawn T. and Lynn Smith-lovin. 1992. “Selective Interaction as a Strategy for Identity Maintenance: An Affect Control Model.” Social Psychology Quarterly 55:12–28. Robinson, Dawn T. and Lynn Smith-Levin. 1999. “Emotion Displays as a Strategy for Identity Negotiation.” Motivation and Emotions 23:73–104. Robinson, Dawn T., Lynn Smith-lovin, and Olga Tsoudis. 1994. “Heinous Crime or Unfortunate Accident? The Effects of Remorse on Responses to Mock Criminal Confessions.” Social Forces 73:175–190. Simon, Robin W. and Leda E. Nath. 2004. “Gender and Emotion in the United States: Do Men and Women Differ in Self-reports of Feelings and Expressive Behavior?.” American Journal of Sociology 109(5):1137–1176. Smith-lovin, Lynn. 1990. “Emotion as the Confirmation and Disconfirmation of Identity: An Affect Control Model.” In: T. D. Kemper (ed.), Research Agendas in the Sociology of Emotions, Chapter 9. Albany: State University of New York Press. Smith-lovin, Lynn and David R. Heise. 1988. Analyzing Social Interaction: Advances in Affect Control Theory. New York: Gordon and Breach Science Publishers. (Reprint of a Special Issue of the Journal of Mathematical Sociology, Vol. 13.) Smith-lovin, Lynn and Dawn T. Robinson. 2006. “Control Theories of Identity, Action, and Emotion: in Search of Testable Differences Between Affect Control Theory and Identity Control Theory.” In: Kent Mcclelland and Thomas Fararo (eds), Purpose, Meaning, and Action: Control System Theories in Sociology. New York: Palgrave Macmillan Press. Stets, Jan E. 2003. Justice, Emotion, and Identity Theory. Advances in Identity Theory and Research. New York: Kluwer Academic/plenum Publishers. Stets, Jan E. 2005. “Examining Emotion in Identity Theory.” Social Psychology Quarterly 68:39–74. Stets, Jan E. and Peter J. Burke. 2005. “New Directions in Identity Control Theory.” Advances in Group Processes 22:43–64. Stets, Jan E. and Michael J. Carter. 2006. “The Moral Identity: A Principle Level Identity.” In: K. Mcclelland and T. J. Fararo (eds), Purpose, Meaning, and Action: Control Systems Theories in Sociology, pp. 293–316. New York: Palgrave.

CH04-P374095.indd 73

5/26/2008 2:32:58 PM

74

Social Structure and Emotion

Stets, Jan E. and Michael M. Harrod. 2004. “Verification Across Multiple Identities: The Role of Status.” Social Psychology Quarterly 67:155–171. Stets, Jane E. and Teresa Tsushima. 2001. “Negative Emotion and Coping Responses within Identity Control Theory.” Social Psychology Quarterly 64(3):283–295. Stryker, Sheldon. 1980. Symbolic Interactionism: A Social Structural Version. Menlo Park, CA: the Benjamin/Cummings Publishing Company. Stryker, Sheldon. 1987. “The Interplay of Affect and Identity: Exploring the Relationships of Social Structure, Social Interaction, Self, and Emotion.” Presented at the American Sociological Association Meetings in Chicago. Stryker, Sheldon. 2004. “Integrating Emotion into Identity Theory.” Advances in Group Processes 21:1–32. Swann, William B., John J. Griffin, Steven C. Predmore, and Bebe Gaines. 1987. “The Cognitive– affective Crossfire: When Self-consistency Confronts Self-enhancement.” Journal of Personality and Social Psychology 52(5):881–889. Tajfel, Henri and John C. Turner. 1986. “The Social Identity Theory of Intergroup Behavior.” In: S. Worchel and L. W. Austin (eds), Psychology of Intergroup Relations. Chicago: Nelson-hall. Tsoudis, Olga. 2000. “The Likelihood of Victim Restitution in Mock Cases: Are the ‘Rules of the Game’ Different from Prison and Probation?.” Social Behavior and Personality 28:483–500. Tsoudis, Olga and Lynn Smith-lovin. 1998. “How Bad Was It? The Effects of Victim and Perpetrator Emotion on Responses to Criminal Court Vignettes.” Social Forces 77:695–722. Tsoudis, Olga and Lynn Smith-lovin. 2001. “Criminal Identity: the Key to Situational Construals in Mock Criminal Court Cases.” Sociological Spectrum 21:3–31. Tsushima, Teresa and Peter J. Burke. 1999. “Levels, Agency, and Control in the Parent Identity.” Social Psychology Quarterly 62:173–189. Turner, R. Jay, Blair Wheaton, and Donald A. Lloyd. 1995. “The Epidemiology of Social Stress.” American Sociological Review 60(1):104–125.

CH04-P374095.indd 74

5/26/2008 2:32:59 PM

5 Identity, Social Status, and Emotion Peter j. Burke University of California, Riverside, CA, USA

ABSTRACT Identity control theory was formed in the context of structural symbolic interaction, which attends to the different positions in the social structure in which persons with given identities are embedded and to the impact of structural position on identity processes. One of the differences among social positions is the amount of resources the position controls and hence the amount of status accorded to the occupant. I examine the effects of social status on the emotional impacts of the lack of identity verification. Status, conceptualized as a symbolic marker indicating who has control of resources, is hypothesized to have two effects. Status and resources help persons verify their identities, and at the same time, help to buffer the consequences of a lack of verification. Data from the spousal identity of 286 newly married couples are examined and generally support these hypotheses. The implication of these findings for existing theory and future research are discussed. Key Words: Identity Theory, Social Status, Emotion, Resources, Identity Verification

INTRODUCTION Identity theory has its roots in the structural symbolic interaction framework (Stryker [1980] 2002), which both acknowledges and attends to the fact that identities operate in culturally defined positions (roles and groups) within the social structure. Role identities are defined by the cultural meanings and expectations associated with role positions, and social identities are defined by the cultural meanings and expectations associated with groups and socially defined Social Structure and Emotion

CH05-P374095.indd 75

75

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 2:34:18 PM

76

Social Structure and Emotion

categories. Groups, social categories, and roles operate to manipulate resources. At a very abstract level, connected resource flows form the fundamental interaction process of any social system. Indeed, system structures may be viewed as connected resource flows (Freese and Burke 1994). By verifying identities, people act to counteract situational disturbances to bring self-relevant situational meanings into alignment with their identity standards. Changing situational meanings involves the manipulation of active (currently in use) and potential (not currently in use) resources through the control of signs and symbols.1 Thus, identity verification is the way in which the resources that sustain individuals, interactions, and the social structure are appropriately marshaled, transformed, and transferred to maintain personal and structural integrity. In the process, identity verification also uses active resources. Without resources the process of identity verification cannot occur. Since persons are the agents through which identities operate, and each person is the host to many identities, in terms of the multiple roles, groups, and relationships in which the person is involved (James 1890), some of the resources that are manipulated by these multiple identities (in the verification processes) must be directed toward the maintenance of the person as well.2 As a biological entity, a person needs sustenance, but also as a social and psychological being, a person needs resources to maintain him or herself as a functioning unit; a unit through which the identities accomplish their goals. The emergence of a person identity with its own identity standard accomplishes this task of maintaining the person through the manipulation of the signs and symbols that control the necessary active and potential resources. We thus have a picture of role and social identities engaging in identity verification that counteracts disturbances and maintains the social structure through the transfer and transformation of the resources to which the identities are given access, and of person identities engaging in identity verification that maintains the person as an active agent through the control of some subset of these resources. Both of these maintenance functions manipulate active and potential resources, but they also require resources for their accomplishment. Identities control resources through verification and identities use resources to carry out the verification. Adequate resources are thus necessary for verification to occur. Identity theory suggests, and prior research has found that negative emotional reactions occur to the extent that identities are not verified (Burke 1991; Burke and Harrod 2005; Burke and Stets 1999; Cast and Burke 2002; Stets and Burke 2005). The present research examines the consequences of access to resources through one’s position in the social structure for the verification process with a special focus on the emotional consequences of identity verification/non-verification. 1 Signs are a more general class than symbols, the latter being restricted to those signs that have shared conventional meanings. Non-symbolic signs provide a direct experience of the situation that is not necessarily shared (Lindesmith and Strauss 1956). 2 This siphoning of resources from the flows for verification of person identities would include both legitimate siphoning that occurs through salary and benefits associated with the role, as well as non-legitimate siphoning such as embezzlement, backshish, bribery, and the like.

CH05-P374095.indd 76

5/26/2008 2:34:18 PM

77

Identity, Social Status, and Emotion

THEORY IDENTITIES AND MEANINGS

An identity contains the set of meanings defining who one is in terms of his or her roles (e.g., truck driver, wife, or professor), group or social category memberships (e.g., American, fraternity member, or female), or personal characteristics (e.g., dominant, sweet, or supportive). These self-meanings compose what are called identity standards (one standard for each identity). In addition to the identity standard (as shown in Figure 5.1), an identity also consists of the set of perceptions of self-relevant meanings in the interactive situation, and a comparator that functions to compare the perceived self-relevant meanings with the corresponding meanings in the identity standard. Any difference or discrepancy signals an error, which is output ultimately in the form of social behaviors that are aimed at altering the meanings in the situation so as to bring perceptions of those altered meanings into alignment with the identity standard and reduce the error or discrepancy to zero. If there is no discrepancy, people’s behavior reflects in the situation the selfmeanings held in their identity standard. When those meanings are disturbed, usually by others interacting in the situation who are also trying to keep their self-relevant meanings in alignment with their identities, people behave in such a way as to counteract the disturbed self-relevant meanings and bring them back into alignment with their identity standards (Burke 1991). This is the process Identity standard

Comparator Emotion

Perceptions

Error Person

Environment

Output

Reflected appraisals

Symbol and resource flows in the environment

Social behavior

Disturbances

FIGURE 5.1

CH05-P374095.indd 77

Basic identity model.

5/26/2008 2:34:18 PM

78

Social Structure and Emotion

of self-verification: acting to counteract disturbances and adjust the self-relevant meanings in the situation so that the perceptions of them match those in the identity standard (Burke 2004a,b). Thus, an identity standard serves as a reference against which perceived selfrelevant meanings in the interactive situation are compared. When the perceived meanings match the meanings in the identity standard, there is identity verification and people both feel good about that and will continue to act has they have been acting. If the perceived meanings do not match the identity standard because of some disturbance in the situation, a discrepancy exists which is subjectively felt as a negative emotion, and which leads to behavior designed to counteract the discrepancy and restore the situationally self-relevant meanings to match those in the identity standard (Burke 1991; 2004a). Meaning is defined, following Osgood and his colleagues (Osgood, May, and Miron 1975; Osgood, Suci, and Tannenbaum 1957) as a mediational response to signs (arising through immediate experience) and symbols (socially shared and consensual). These responses vary across bipolar dimensions of meaning such as those universal dimensions that Osgood and his colleagues discovered: evaluation (good–bad), potency (strong–weak), and activity (active–passive). These universal dimensions are found in most cultures, but together they only account for about 50% of the variance in people’s responses to stimuli – i.e., these three dimensions capture about half the meaning. The other half of meaning lies in the multitude of socially relevant dimensions we find important for understanding the world: dimensions like moral–immoral, masculine–feminine, empty–full, ordered–disordered, or eager–reticent. For each relevant dimension of meaning, the identity standard sets a level that should be maintained for the person to be who they are. If degree of riskiness is an important dimension of one’s gender identity, each person tries to maintain an impression of riskiness that corresponds to how they define their self. If they are being too risky, they act to become more conservative. If they are too conservative, they increase their risk behavior to the level that satisfies the level set in their gender identity standard. IDENTITIES AND RESOURCES

Tied to the meanings arising from signs and symbols are resources. By controlling meanings, people control resources. Freese and Burke (1994) distinguished between active resources (those that are functioning in the current situation to sustain persons, groups, and interaction) and potential resources (those that are not currently being used in the situation, but are being readied, transformed, transported, etc. to the time and location where they will be used as active resources). Examples of active resources are the heat and light that keep us comfortable and allow us to see to interact; the table on which the food we are eating rests; the utensils we are using to eat the food; the love and friendship expressed by those around us; and so on. Examples of potential resources are the car that is in

CH05-P374095.indd 78

5/26/2008 2:34:18 PM

Identity, Social Status, and Emotion

79

the garage, food that is in the refrigerator, the friend with whom we are not currently interacting, and oil in the pipeline. Active resource transfers are the objects of immediate experience, indicated by signs. Potential resources are the objects of reflective experience, indicated by symbols. As Freese and Burke (1994) point out, at a very abstract level, connecting resource flows is the fundamental goal of interaction, while system structures are the connected resource flows. From this perspective, then, identity verification is accomplished by manipulation of meanings in the situation, but as meanings (arising from both signs and symbols) are tied to resource flows, identity verification is accomplished by the manipulation of active and potential resources. Resources are thus central for the maintenance of identities (Stets and Cast 2005). The manipulation of resources by individual identities in the process of identity-verification brings about the resource flows and transformations that together define the social system (Burke 2004a). As a truck driver, I verify my truck driver role identity by doing the things that define me as a truck driver, and in the process, I have moved (potential) resources between two locations as I haul containers from the Port of Los Angeles to the Inland Empire. I have also used the (actual) resources of my skills and knowledge, as well as the truck, gasoline, oil, tires, roads, etc. to accomplish this. All of my behaviors as a truck driver serve both to portray my identity (e.g., my uniform, my interactions, and conversations with others) and to counteract unexpected and unpredictable disturbances that arise (a flat tire, engine trouble, speed traps, heavy traffic). These disturbances cause me to engage in unanticipated behaviors to accomplish what I need to accomplish in order to maintain meanings consistent with my role as truck driver. Indeed, without this ability to counteract disturbances, the social system would quickly fall apart. Being able to accomplish identity goals (i.e., make perceptions of self-relevant meanings match meanings in the identity standard for the way they are supposed to be) in spite of disturbances is the unique viewpoint of identity theory. The goals have to do with keeping (potential) resource levels and flows at the levels set by the identity standard. The means for reaching the goals use (active) resources in the process. Thus, identity verification requires resources, and those persons with more resources should be better able to verify their identities (Stets and Cast 2005; Stets and Harrod 2004). IDENTITIES, RESOURCES, AND SOCIAL STRUCTURE

Because every position in the social structure is expected (through identity verification) to control relevant resources for the maintenance of the social structure (by controlling sign and symbolic meanings), legitimate access to those resources is built into every position. The nature of the access and the nature of the resources varies by position, as does the amount of resources made available to the position. The role identity of CEO of Microsoft has more resources available for control than does a programmer hired to code a section of a new product.

CH05-P374095.indd 79

5/26/2008 2:34:18 PM

80

Social Structure and Emotion

Control of resources is not the same as status, but status is accorded to persons in part by the amount of resources that are controlled by the person. Indeed, this is the original coding of occupational status in the Duncan SEI scores (1961), using information about the median education and income for persons in any given occupation to predict the level of esteem and respect that people accord persons in that occupation. In this way, occupational category is a status characteristic indicating not only the status rank a person has, but the level of resources someone in that group could be expected to have (on the average). Status construction theory (Ridgeway 2006) makes this point. Arbitrary social categories can become status indicators when the neutral social categories are seen to be associated with superior material resources. If persons in one category are seen to have more resources, membership in that category can in itself come to signal status. In this way, status characteristics become substitutes for knowledge about real control of resources, but they also become a visible guide to people in society as to where the resources are likely to be found – that is, who has control of the resources. Thus, I expect that those with higher status have more resources and they should, as a consequence, be better able to verify their identities (Cast, Stets, and Burke 1999; Stets and Cast 2005). Usually, we distinguish between two types of status characteristics, those that are achieved and those that are ascribed. Higher levels of education or occupational position signal higher achieved status, while racial categories signal relative ascribed status with higher status being given to persons classified as white as opposed to non-white. Being in the higher status group opens doors to the control of more resources and the control of more resources brings about higher status. In the present research, I will use a combined index of achieved status based on both education and occupational status, as well as an index of ascribed status consisting of the categories of white and non-white. Based on the above discussion, I hypothesize: Hypothesis 1:

Hypothesis 2 :

Persons with lower education and occupational status will have higher levels of discrepancy between identity meanings and situationally perceived meanings. Persons in the lower status non-white group will have higher levels of discrepancy between identity meanings and situationally perceived meanings.

IDENTITIES AND EMOTIONS

The identity verification process is guided in part by the emotional responses that people have as a function of the error signal or discrepancy between perceptions and identity standard (Burke 1991). In general, people feel bad when the discrepancy is large or increasing and they feel good if the discrepancy is small or decreasing. More recently Stets and Burke (2005) have made a number of more specific predictions about emotional responses that take into account not

CH05-P374095.indd 80

5/26/2008 2:34:18 PM

Identity, Social Status, and Emotion

81

just the size of the discrepancy, but also a number of other situational features that provide additional meanings. For example, they point out that if the self is the source of the discrepancy, people are more likely to feel embarrassment or shame when the identity standard meanings are set by others (oughts) and to feel disappointment and sadness when the identity standard meanings are set by the self (ideals). Similarly, if another is the source of the discrepancy, people are likely to feel annoyance or hostility when the identity standard consists of oughts, but feel anger or rage if the identity standard consists of ideals. Even these emotions may be modified depending upon the relative status of the other involved in the situation. One may feel anger at a subordinate who creates a discrepancy between perceived meanings and the identity standard, and one may feel fear of a superordinate who creates the discrepancy. The emotions that arise can serve as guides to tell us about the success of our efforts to verify our identity. The good feelings tell us we are doing the right thing. The bad feelings tell us that we need to change our behavior, and the more specific emotions guide us to behavior that is appropriate in the situation for the reduction of the discrepancy. The particular emotions and actions that are generated by a discrepancy are the ones that best serve not only to reduce the discrepancy, but to preserve the social structural relations within which the identity exists. There is another factor that plays into the emotional responses generated by the discrepancy. As Burke (1966) pointed out with respect to the emergence of authority structures in small groups, there are two levels of problems to be solved by a group that he distinguished as first- and second-order problems. In order for a group to decide who is going to be the leader (the second-order problem), they must first decide on how such a decision is going to be made in the first place (the firstorder problem). First-order problems are infrastructure problems – putting things in place that can be used later in solving second-order problems. In that study, Burke found that the emotional reactions to second-order problems were much more severe if first-order problems had not been solved prior to the onset of the second-order problems. The reaction could be described as “we have a problem and we have no way to solve it.” If first-order problems were already solved, then the response could be described as “we have a problem, but we know how to solve it.” In the present situation dealing with identity verification, recall that resources are needed to accomplish identity verification. The first-order problem is solved if there are sufficient resources present to allow an identity to exert the necessary control over meanings in the situation when disturbances (second-order problems) create discrepancies between perceptions and the identity standard. The parent who is short on cash to buy food for her children (a disturbance to verifying her parent identity through supporting the children) is much more upset when there is no way to get cash (for example, with a debit or credit card). With the backup resources available (first-order problem solved), the disturbance (a secondorder problem) is minor and temporary. In this way, the additional resources held by higher status persons can serve to prevent strong negative emotional

CH05-P374095.indd 81

5/26/2008 2:34:18 PM

82

Social Structure and Emotion

reactions to the lack of identity verification that would be felt by lower status persons without those resources. These points lead to two additional hypotheses, again on the assumption that persons with more status have access to more resources: Hypothesis 3:

Hypothesis 4:

Persons with lower education and occupational status will feel the emotional impact of a lack of identity verification to a greater extent than those with higher status. Persons in the lower status non-white group will feel the emotional impact of a lack of identity verification to a greater extent than those in the higher status white group.

PROCEDURES SAMPLE

The data for this study come from a longitudinal study investigating marital dynamics in the first 2 years of marriage (Tallman, Burke, and Gecas 1998). Three data collection points, each one year apart, included a 90-minute face-toface interview, a 15-minute videotaping of a conversation focused on solving an area of disagreement, and four consecutive one-week daily diaries kept by each respondent. The present analyses are based on information gathered during the face-to-face interview at each point in time. The sample was drawn from marriage registration records in 1991 and 1992 in two mid-size communities in Washington State. Of the 1,295 couples registered to marry, 574 met the criteria for involvement (both were over the age of 18, were marrying for the first time, and had no children). These couples were contacted and asked to participate; 286 completed all data collection processes in the first period. There was a 15% attrition rate from the first data collection period to the second period and an additional 4.2% attrition rate from the second to the third period of data collection.3 MEASURES

Problems with identity verification were indicated by a measure of the discrepancy between the respondent’s spousal identity standard and an indicator of their self-relevant perceptions developed following a method used by Swann, De La Ronde, and Hixon (1994) and Swann, Hixon, and De La Ronde (1992). These researchers examined the extent to which an individual’s view of their spouse was congruent with the spouse’s self-views on attributes relevant to their self-concept.

3

A more complete description of the sample can be found in Tallman, Burke, and Gecas (1998).

CH05-P374095.indd 82

5/26/2008 2:34:18 PM

Identity, Social Status, and Emotion

83

In the present study, the focus is on the spousal role identity and I compare the individual’s own meanings and expectations contained in their spousal role identity standard with the expectations held for them by their partner. Selfverification for a respondent exists when the respondent’s self-view is confirmed by the view that the respondent’s spouse has for the respondent. While it would be theoretically preferable to measure each respondent’s actual perceptions of the spouse’s expectations for him or her, this data was not collected; instead, I use the spouse’s actual expectations as a proxy for the respondent’s perceptions (cf. Burke and Stets 1999; Cast and Burke 2002). In the use of the spouse’s expectations as a proxy for the actor’s reflected appraisals several assumptions are made: (1) the meanings in the actor’s identity standard generally correspond to the meaning of the actor’s role performance; (2) the spouse appraises the meanings of the actor’s role performance negatively if they differ from the spouse’s expectations as to how the actor should behave; (3) the spouse acts so as to convey the meanings of his or her appraisals and expectations to the actor; and (4) the actor perceives the spouse’s behavior and infers the spouse’s appraisal (reflected appraisals). The proxy measure of the expectations mentioned in the second assumption is two steps away from the desired measure of reflected appraisals mentioned in assumption 4. Among casual acquaintances or in one-time encounters, these would be large steps, where error could enter easily. Kinch (1963), however, suggests that several conditions can reduce the level of error which might affect these assumptions: (1) self ’s familiarity with the other, (2) the level of familiarity with the situation, and (3) the actor’s past experiences in interpersonal situations. All of these conditions should be met among the newly married couples in the present sample. Therefore, it seems reasonable to assume that a spouse’s expectations for an actor are correlated highly both with his or her appraisals of the actor and with the actor’s perceptions of those appraisals. Thus, although the measure of identity verification is not exactly what is desired, it is a close approximation and a useful proxy. To assess the spousal role identity standard, respondents rated 11 different spousal role activities in terms of the degree to which they felt they themselves should engage in the activity as part of their spousal role. They also rated the activities in terms of the degree to which they felt that their spouse should engage in that activity. Examples of the activities include “being responsible for maintaining contact with parents and in-laws or other members of the family,” “being responsible for taking care of bills and accounts,” and “being responsible for hard work.” It should be noted that it is not the activities themselves that are important, but rather the meanings that are implied by engaging or not engaging in these activities. Response categories ranged across a 5-point scale from doing all of the activity in the relationship (coded 4) to doing none of the activity (coded 0). While most of these items dealt with household activities, and the spousal role in all of its complexity clearly includes more than this, the meanings controlled by these activities are, nevertheless, important aspects of the spousal role

CH05-P374095.indd 83

5/26/2008 2:34:18 PM

84

Social Structure and Emotion

identity. However, to the extent that the identity is not fully measured with this scale, there will be measurement error and tests of hypotheses will be weakened. The degree of disagreement or discrepancy is assessed by calculating the average absolute difference between the respondent’s self-description scores and the perception of the respondent held by the respondent’s spouse. This score could range from 0 indicating perfect agreement to 4 indicating maximum disagreement. The scale was based on the mean of the 11 differences. The resulting identity discrepancy scale scores was standardized and ranged from ⫺1.77 to 6.61. The omega reliability for the scale is 0.88. Two indicators of status position were used to assess relative relationship to available resources. The first was a combined index of years of education and Duncan’s SEI index of occupational prestige, as updated by Stevens and Cho (1985). The combination of education and occupational status was achieved by first standardizing each of the components, adding them together, and then standardizing the resulting Ed/Occ scale. Higher scores represent higher status. The second indicator was race (as white/non-white coded 1 for whites).4 In terms of indicators of emotional outcomes, four were used: anger, depression, distress, and love (felt for the spouse). Anger, as one of the primary emotions, is focused outwards and helps orient a person toward removing a perceived external barrier that hinders identity verification (Turner and Stets 2005). Depression, also considered a primary emotion by Kemper (1987) and often linked with sadness, is focused more inwards and helps a person attempt to reintegrate with others in the group (Turner and Stets 2005). Distress or anxiety is viewed as a secondary emotion that combines both fear and expectancy that often arises when important matters are out of our control (Kemper 1991). Finally, love is a socially derived emotion in which one person gives high amounts of sociability, solidarity, and affection to another (Kemper 1989).5 Because the identity discrepancy directly involved the other spouse, the emotions generated between them are particularly relevant to understanding the emotional consequences of a lack of identity verification. Anger was measured using five items from the hostility subscale of the SCL-90 (Derogatis 1977). Respondents were asked, for example, how many days during the last week they “got angry over things that weren’t really too important,” “had temper outbursts,” or “wanted to hurt or smash something.” Response categories range from 0 (“not at all”) to 7 (“seven days”). Items were aligned in the same direction and summed. The omega reliability for the scale is 0.83 with a high score indicating high levels of anger. The final scale was standardized. Depression was measured using 12 items from the CES-D scale (Radloff 1977). Example items include asking respondents how many days during the last week

4 Using gender as a status indicator is not feasible in the present sample because of the lack of independence between husbands and wives. 5 Kemper calls this the “status” dimension of social relationships (in contrast to the power dimension), though his use of the term “status” is quite different from the use of the term status in this paper.

CH05-P374095.indd 84

5/26/2008 2:34:18 PM

85

Identity, Social Status, and Emotion

they had experiences such as “feel lonely,” “sleep restlessly,” and “feel you could not get going.” Response categories range from 0 (“not at all”) to 7 (“seven days”). The items form a single factor with an omega reliability of 0.95. Items were aligned to the same direction and summed. Possible scores range from 0 to 84. A high score indicates high levels of depression. The final scale was standardized. Distress was measured using nine items from the anxiety subscale of the SCL-90 scale (Derogatis 1977). Respondents were asked, for example, how many days during the week they had “felt keyed up or excited,” “felt hands trembling,” and “felt nervous or have an upset stomach.” Response categories range from 0 (“not at all”) to 7 (“seven days”). Items were aligned and summed. The omega reliability for the anxiety measure is 0.89. The final scale was standardized, with a high score indicating high distress. Finally, love was measured using the Rubin Love Scale (Rubin 1973). Respondents were asked to indicate the extent to which each of 13 statements reflecting how they felt about their spouse, such as “I would do almost anything for (spouse),” “If I could never be with (spouse), I would feel miserable,” and “One of my primary concerns is (spouse’s) welfare.” Responses ranged from “not at all true” to “definitely true” (coded 0–8). The items were summed to form a scale in which a high score reflected high levels of love toward the spouse. The scale was standardized and had an omega reliability of 0.88. Means, standard deviations, and correlations for all of these variables are given in Tables 5.1 and 5.2. ANALYSIS

Because there are three time points in the data, not all the observations are independent. As a consequence, cross-sectional time-series analysis procedures were used to analyze the data (Baltagi 2001; Greene 1990). This procedure allows for correlated errors due to the non-independence of these observations and it combines information about the cross-sectional parts of the data with information on the time-series parts to provide estimates of effects.

TABLE 5.1

Means, standard deviations, minima, and maxima for the variables

Variable

Mean

Standard deviation

Minimum

Maximum

Ed/Occ

0.00

1.00

⫺2.13

2.98

White

0.92

0.27

0.00

1.00

Discrepancy

0.00

1.00

⫺1.78

6.61

Anger

0.00

1.00

⫺0.89

6.50

Depression

0.00

1.00

⫺1.11

6.69

Distress

0.00

1.00

⫺1.02

6.84

Love

0.00

1.00

⫺4.01

1.71

CH05-P374095.indd 85

5/26/2008 2:34:18 PM

86

Social Structure and Emotion

TABLE 5.2

Correlations among the variables Variables

Variables 1 Ed/Occ 2 White

(1)

(2)

(3)

(4)

(5)

(6)

(7)

1.00 0.05*

1.00

3 Discrepancy

⫺0.16*

⫺0.09*

1.00

4 Anger

⫺0.18*

⫺0.04

0.14*

1.00

5 Depression

⫺0.19*

⫺0.03

0.13*

0.63*

6 Distress

⫺0.19*

⫺0.01

0.13*

0.66*

7 Love

⫺0.18*

0.04

0.04

⫺0.04

1.00 0.69* 0.00

1.00 ⫺0.01

1.00

p ⱕ 0.05.

*

RESULTS I begin with an analysis of the impact of the status characteristics on the level of the discrepancy between a person’s spousal identity standard and how he or she appears in the situation. Table 5.3 shows the results of that analysis and it can be seen that white persons have lower spousal identity discrepancies (more spousal identity verification) than do non-whites, and that the higher a person’s education and occupational status the less the spousal identity discrepancy. Both of these effects are in accord with the hypothesis that status groups, reflecting access to resources, have an impact on the ability of persons to verify their identities. Those in the lower status categories, with fewer resources, have more trouble verifying their spousal role identity. I turn now to look at the consequences of status categories on emotional reactions to identity discrepancy. Table 5.2 shows the zero-order relationships between discrepancy and each of the four emotions of anger, depression, distress, and love. For the negative emotions of anger, depression, and distress, we see that the greater the spousal identity discrepancy, the greater are these negative emotions. This is consistent with our hypotheses and shows the basic relationship between discrepancy and emotion. There is, however, no zero-order relationship between spousal identity discrepancy and love felt for the spousal partner. While we might expect that feelings of love would be decreased toward the spouse with whom one is having problems verifying their spousal role identity, such a zeroorder relationship does not exist. Adding in the potential moderating effects of status categories, as reported in Table 5.4, shows that a number of the interactions are significant. As a consequence, we know that the relationship between identity discrepancy and the emotional outcomes varies among the different status groups. Considering anger first, the table shows that for every increase of one standard deviation in identity

CH05-P374095.indd 86

5/26/2008 2:34:18 PM

87

Identity, Social Status, and Emotion Standardized regression coefficients for the effects of race and Ed/Occ on identity discrepancy

TABLE 5.3

Status variable

Discrepancy

White

⫺0.09*

Ed/Occ

⫺0.12*

*

p ⱕ 0.05.

Standardized regression coefficients for the effects of identity discrepancy, race, Ed/Occ, and their interactions on emotions.

TABLE 5.4

Emotion outcomes Variables

Anger

Depression

Distress

Anger depression and distress constrained estimates1

Love

⫺0.05*

Discrepancy

0.09*

0.08*

0.07*

0.10*

White

0.00

0.00

0.01

0.00

0.04

Ed/Occ

⫺0.15*

⫺0.13*

⫺0.16*

⫺0.17*

⫺0.15*

Discrepancy X White

⫺0.06*

⫺0.03

0.01

⫺0.04*

0.04*

Discrepancy X E/O

⫺0.04

⫺0.06*

⫺0.04*

⫺0.04*

0.04*

0.06

0.06

R2

0.06

0.05

1 Estimates of coefficients for anger, depression, and distress constrained to be the same across emotions. * p ⱕ 0.05.

discrepancy, the amount of anger increases by 0.09 standard deviations for nonwhites with average educational and occupational status. The significant interaction for whites indicates that the relationship for whites is ⫺0.06 weaker than for non-whites. Thus, for whites with average educational and occupational status, the impact of identity discrepancy on anger is reduced to 0.03. The protective cloak of being in the high status group has considerably reduced the amount of impact on anger for identity verification failure. The direct effects of educational and occupational status show that there is less of an anger reaction for persons with higher status overall, independent of other things, however, we do not see any significant moderator effects for this status variable as it might impact the consequences of identity discrepancy. Considering next the emotion of depression, the results in Table 5.4 show that, again, for non-whites, an increase in discrepancy leads to an increase in depression

CH05-P374095.indd 87

5/26/2008 2:34:19 PM

88

Social Structure and Emotion

( ⫽ 0.08). Also, having higher educational and occupational status reduces the amount of depression overall, but this also interacts with the effects of identity discrepancy so that persons with higher status do not feel the depressive effects of identity discrepancy to the extent that persons with lower status do. For nonwhites who are one standard deviation above the mean of educational and occupational status, the effects of identity discrepancy on depression are reduced to almost zero. Conversely, if a person is one standard deviation below the average on educational and occupational status, the impact of identity discrepancy on depression is strongly increased, almost doubling. With respect to the consequences of being non-white, Table 5.4 shows no significant variation in the consequences of identity discrepancy on emotional response that depends upon being white or non-white. The effects for the emotional response of distress are similar to the effects for depression. The greater the identity discrepancy, the greater is the level of distress reported for non-whites ( ⫽ 0.07), but this basic effect on distress is decreased for persons with higher than average levels of educational and occupational status. Racial category does not significantly moderate the consequences of identity discrepancy on distress. Because the effects of identity discrepancy for the emotions of anger, depression, and distress are all very similar in magnitude (though not in meaning), I tested the significance of the differences among the different results. The comparable coefficients across the different emotional outcomes are not significantly different (F12, 1498 ⫽ 0.37, p ⬎ 0.95). Using constraints to keep the estimates the same across equations provides more powerful estimates of the effects. These are presented in the fourth column in Table 5.4 labeled “constrained estimates.” For these estimates, interaction effects for the negative emotions are significant and suggest that the effects of identity discrepancy on these negative emotions are reduced for persons in the higher status group of whites and for the higher levels of educational and occupational status. These results suggest that although the emotions are different and felt to different extents by different people, identity discrepancy has comparable standardized effects on all of these emotions and these effects are moderated by status for all three emotions. Finally, turning to the emotion of love, Table 5.4 shows that the baseline effect of increasing discrepancy for non-whites with average levels of educational and occupational status is to decrease the level of love felt for the partner, which is consistent with the hypothesis. Like the effects for the negative emotions, Table 5.4 shows that there are also moderating effects for both racial status as well as educational and occupational status on the consequences of identity discrepancy for the emotion of love. Being in the higher status white group reduces the impact of the discrepancy to almost zero, and moving up one standard deviation in educational and occupational status also significantly reduces the negative effects of identity discrepancy. The strongest negative effects are for non-whites with low educational and occupational status.

CH05-P374095.indd 88

5/26/2008 2:34:19 PM

Identity, Social Status, and Emotion

89

DISCUSSION Emotion is one of the concomitants of the identity verification process. Emotion is not just an outcome, because it also helps serve to motivate and guide the verification process of keeping perceptions of self-relevant meanings in the situation in alignment with the meanings in the identity standard as suggested by Stets and Burke (2005). The stronger the discrepancy between perceptions and identity standard, the more are positive emotions reduced while negative emotions are increased in strength and persistence (Burke 1991; Burke and Harrod 2005) and the more motivated is the person to counteract the disturbance that is the source of the discrepancy. Emotions also vary by the social context and the nature of the discrepancy, with different emotions leading to different restorative behaviors, with anger, for example, focusing one outward and sadness and depression focusing one inward (Stets and Burke 2005). According to identity theory, resources are intimately tied up in the identity verification process and hence with the emotions that are produced in that process. The self-relevant meanings that are perceived in the situation are directly tied to actual and potential resources through signs and symbols. Controlling the perceived meanings is accomplished by controlling the resources to which the meanings are attached. Bringing the perceived self-relevant meanings into alignment with the meanings in the identity standard brings the resource flows into alignment with the dictates and needs of the role or group attached to the identity, and thereby maintains the connected resource flows of the social structure. In addition, it takes resources to manipulate the meanings/resources in the situation. With more resources at one’s disposal, one has greater control over the resources that are under the jurisdiction of one’s identity. And, with greater control over the resources, there is less likely to be any discrepancy between perceived self-relevant meanings and identity standard meanings. The discrepancies that do occur also should be smaller and of shorter duration. Hence, the negative emotions that arise when discrepancies are large or increasing should be less strong for persons who have more resources. Status is a social marker indicating who has control of resources. With greater access to and control of resources comes greater status and the social marker of status may be more visible to others than the degree of access to resources. We thus often rely on status to tell us who is likely to control important resources. It was in this sense that I examined the relationship between status, the degree of identity verification and the emotions that arise when identities are not verified. What I found was that persons who are in higher status positions, either in terms of education and occupational status or in terms of being a member of the higher status white as opposed to a non-white racial group, had both a greater degree of verification of the spousal identity and were less likely to experience strong negative emotions or a reduction of positive emotions when their spousal identity was not verified. Three negative emotions were examined: anger, which is focused outward and helps orient a person toward removing a perceived external barrier to identity

CH05-P374095.indd 89

5/26/2008 2:34:19 PM

90

Social Structure and Emotion

verification; depression, which is focused more inwards and helps a person attempt to make internal adjustments to their identity standards with the ultimate consequence of increasing identity verification; and distress, which signals that important matters are out of control. From the results of the present study, each of these emotions arose when the discrepancy between perceived meanings and identity meanings increased, especially among those who had lower status. The zero-order correlations among these different negative emotions were in the moderate range of 0.6 meaning that there was an overlap of about 40% in terms of persons who felt one emotion more strongly also feeling another emotion more strongly. On the other hand, there is clearly a lot of independence among the three emotions. People who feel one are not necessarily the people who feel another, which suggests that while identity discrepancy is increasing each of these negative emotions, it is not necessarily the case that any person feels all of them as a result. Rather, persons are likely to feel one or another depending upon other circumstances (Stets and Burke 2005). Indeed, classifying people as low or high on each of the three negative emotions (using median splits) shows that about 42% are low on all three negative emotions (primarily the persons whose spousal identities are better verified). Another 19% are high on all three, and the rest feel various one or two-way combinations. For example, if a person were experiencing a discrepancy in their spousal identity between the meanings in their spousal identity standard and how those meanings were perceived in the situation, and if the person felt that the discrepancy was the result of how another was behaving to disturb the meanings they were trying to control, the person might feel anger. On the other hand, if the person felt that he or she was not able to control the meanings in the manner they desired, then the person may feel depressed. Finally, if the situational meanings were disturbed away from the standard and it was not clear what could be done about it, the person may simply feel distressed. And, of course, combinations of these perceptions and situations may exist resulting combinations of negative feelings. In each case, whatever negative emotion or combination the person felt as a result of the identity discrepancy, he or she would experience less of that emotion or combination to the extent that his or her position in the social structure provided more access to resources, represented in higher status. The effect of identity discrepancy on the emotion of love operates in a very similar fashion. For lower status persons without resources, spousal identity discrepancy decreases love of the spouse, that is, changes it in the negative direction. As the status of a person increases, the negative impact of identity discrepancy on love is diminished, once again showing the protective character of having resources as indicated by having higher status. For all four emotions, then, an increase in identity discrepancy results in a change in the negative direction of the emotion, and an increase in status reduces this impact of identity discrepancy. Overall, then, it is clear from the current results that emotional reactions are related to one’s position in the social structure. Persons in status positions that

CH05-P374095.indd 90

5/26/2008 2:34:19 PM

Identity, Social Status, and Emotion

91

provide access to more resources are better able to verify their identities and with this knowledge of their general ability to verify their identities, they are less likely to have strong emotional reactions in those situations where they have some problems verifying their identities. This was the first- and second-order problem issue. Knowing that you can generally verify your identity (first-order problems solved because you have the resources) prevents you from getting upset when situational disturbances (second-order problems) create temporary discrepancies.6 Although this effect was present for all four emotions investigated, the lack of a high correlation among the different emotions suggests that different people respond emotionally in different ways to identity discrepancy. Stets and Burke (2005) provide some hypotheses suggesting that it is additional meanings in the situation which cue the different emotions. Unfortunately, the present study does not contain data relevant to such tests. Therefore, it remains to future research to begin to sort out the conditions under which different emotions will be generated in response to identity discrepancy. Finally, it is of interest to note that those positions with the most resources allowing the greatest verification of identities are likely the positions that are crucial in the social structure in the sense of being at the hub of the transfer of large amount of (potential) resources that ultimately (as active resources) help maintain the social structure itself. The fact that these central identities are better able to verify themselves (than those in lower status position) means that these central identities are better able to counteract disturbances to the general social structure. Key positions are better protected, and the social structure is conserved.

ACKNOWLEDGEMENT The research reported in this paper is from a longitudinal study of firstmarried couples, “Socialization into Marital Roles,” funded by NIMH grant MH46828, under the direction of Irving Tallman, Peter J. Burke, and Viktor Gecas. I would like to thank the participants of the social psychology seminar at UC Riverside for their comments on an earlier draft.

REFERENCES Baltagi, B. H. 2001. Economic Analysis of Panel Data. New York: John Wiley and Sons. Burke, Peter J. 1966. “Authority Relations and Disruptive Behavior in Small Discussion Groups.” Sociometry: 237–250.

6 There may be a limit to the argument that higher status persons, with more resources, can generally erify their identities and therefore react less strongly to disturbances. If the discrepancy caused by a disturbance is larger than normally experienced, or more persistent than normal, higher-status persons may react more strongly than their lower-status counterparts (cf. Cast and Burke 2002).

CH05-P374095.indd 91

5/26/2008 2:34:19 PM

92

Social Structure and Emotion

Burke, Peter J. 1991. “Identity Processes and Social Stress.” American Sociological Review 56:836–849. Burke, Peter J. 2004a. “Identities and Social Structure: The 2003 Cooley-Mead Award Address.” Social Psychology Quarterly 67:5–15. Burke, Peter J. 2004b. “Identities, Events, and Moods.” Advances in Group Processes 21:25–49. Burke, Peter J. and Michael M. Harrod. 2005. “Too Much of a Good Thing?” Social Psychology Quarterly 68:359–374. Burke, Peter J. and Jan E. Stets. 1999. “Trust and Commitment Through Self-Verification.” Social Psychology Quarterly 62:347–366. Cast, Alicia D. and Peter J. Burke. 2002. “A Theory of Self-Esteem.” Social Forces 80:1041–1068. Cast, Alicia D., Jan E. Stets, and Peter J. Burke. 1999. “Does the Self Conform to the Views of Others?” Social Psychology Quarterly 62:68–82. Derogatis, Leonard R. 1977. Scl-90: Administration, Scoring and Procedures Manual for the Revised Version and Other Instruments of the Psychopathology Rating Scale Series. Duncan, Otis Dudly. 1961. “A Socioeconomic Index for All Occupations.” In: A. Reiss (ed.), Occupations and Social Status. New York: Free Press. Freese, Lee and Peter J. Burke. 1994. “Persons, Identities, and Social Interaction.” Advances in Group Processes 11:1–24. Greene, William H. 1990. Econometric Analysis. New York: Macmillan. James, William. 1890. Principles of Psychology. New York: Holt Rinehart and Winston. Kemper, Theodore D. 1987. “How Many Emotions Are There?” American Journal of Sociology 93:263–289. Kemper, Theodore D. 1989. “Love and Like and Love and Love.” In: D. D. Franks and E. D. McCarthy (eds), The Sociology of Emotions: A Structural Approach Greenwich, pp. 249–268. Connecticut: JAI Press. Kemper, Theodore D. 1991. “Predicting Emotions from Social Relations.” Social Psychology Quarterly 54:330–342. Kinch, John W. 1963. “A Formalized Theory of the Self-Concept.” American Journal of Sociology 68:481–486. Lindesmith, Alfred R. and Anselm L. Strauss. 1956. Social Psychology. New York: Holt Rinehart and Winston. Osgood, Charles E., William H. May, and Murray S. Miron. 1975. Cross-cultural Universals of Affective Meaning. Urbana: University of Illinois Press. Osgood, Charles E., George J. Suci, and Percy H. Tannenbaum. 1957. The Measurement of Meaning. Urbana: University of Illinois Press. Radloff, Lenore S. 1977. “The CES-D Scale: A Self-Report Depression Scale for Research in the General Population.” Applied Psychological Measurement 1:385–401. Ridgeway, Cecilia L. 2006. “Status Construction Theory.” In: Peter J. Burke (ed.), Contemporary Social Psychological Theories. Stanford, CA: Stanford University Press. Rubin, Zick. 1973. Liking and Loving: An Invitation to Social Psychology. New York: Holt Rinehart and Winston. Stets, Jan E. and Peter J. Burke. 2005. “New Directions in Identity Control Theory.” Advances in Group Processes 22:43–64. Stets, Jan E. and Alicia D. Cast 2005. “Verifying and Maintaining Social Actors: The Role of Resources.” Presented at the annual meetings of the American Sociological Association, Philadelphia. Stets, Jan E. and Michael M. Harrod. 2004. “Verification across Multiple Identities: The Role of Status.” Social Psychology Quarterly 67:155–171. Stevens, Gillian and Joo Hyun Cho. 1985. “Socioeconomic Indexes and the New 1980 Census Occupational Classification Scheme.” Social Science Research 14:142–168. Stryker, Sheldon. [1980] 2002. Symbolic Interactionism: A Social Structural Version. Caldwell, NJ: The Blackburn Press.

CH05-P374095.indd 92

5/26/2008 2:34:19 PM

Identity, Social Status, and Emotion

93

Swann, William B., Jr., Chris De La Ronde, and J. Gregory Hixon. 1994. “Authenticity and Positivity Strivings in Marriage and Courtship.” Journal of Personality and Social Psychology 66:857–869. Swann, William B., Jr., J. Gregory Hixon, and Chris De La Ronde. 1992. “Embracing the Bitter Truth: Negative Self-Concepts and Marital Commitment.” Psychological Science 3:118–121. Tallman, Irving, Peter J. Burke, and Viktor Gecas. 1998. “Socialization into Marital Roles: Testing a Contextual, Developmental Model of Marital Functioning.” In: T. N. Bradbury (ed.), The Developmental Course of Marital Dysfunction, pp. 312–342. Cambridge, UK: Cambridge University Press. Turner, Jonathan H. and Jan E. Stets. 2005. The Sociology of Emotions. New York: Cambridge University Press.

CH05-P374095.indd 93

5/26/2008 2:34:19 PM

6 A Strategy for Incorporating Social Intelligence in the Design of Virtual Environments Lisa Troyer University of Connecticut, Storrs, CT, USA

ABSTRACT The purpose of this chapter is to examine factors that affect the behavior of human actors in virtual environments. It addresses a longstanding question among researchers and users of virtual reality technologies: Under what conditions do a human actor’s interactions with a virtual actor (i.e., “avatar” or “agent”1) approximate face-to-face interactions among human actors? I integrate elements of social psychological theories: affect control theory [ACT; Heise, R., Journal of Mathematical Sociology 13 (1987) 1] and identity theory [Stryker, S. and P. J. Burke, Social Psychology Quarterly 63 (2000) 284] with insights from a social risk framework to develop a theoretical framework that addresses 1 These terms (avatar, agent) refer to computer-rendered representations of an actor in computermediated environments (such as virtual reality). The term “avatar” generally refers to a computerrendered representation of a human being, the behavior of which is primarily controlled by a human actor. The term “agent” generally refers to a computer-rendered representation of an actor, the behavior of which is controlled by computer scripting and programming. In this paper, I use the term “virtual actor” to represent both of these types of virtual, computer-rendered actors. The project described, however, is primarily directed at facilitating the development of more socially compelling (and hence “accurate”) agents.

Social Structure and Emotion

CH06-P374095.indd 95

95

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 2:36:13 PM

96

Social Structure and Emotion

this question. A critical element in this framework is the role that affect plays in triggering socially appropriate responses in virtual actors. This framework provides a foundation for incorporating “social intelligence” in the design of virtual environments.

INTRODUCTION The question of the conditions under which interaction in a virtual environment approximates interaction in a face-to-face setting represents an issue that is becoming increasingly important across both work and leisure domains of social life. For example, video gaming technologies have evolved into sophisticated virtual environments, in which human actors engage in a range of competitive and social interactions (e.g., sports contests, role-playing games) with both other human actors (represented as avatars) and non-human actors (represented as agents). With increased speed and reliability of computer networks, new communication tools have emerged that allow human actors to enter virtual spaces to interact with one another and digitally rendered and programmed others. In the realm of worklife, virtual environments are emerging as a popular training tool. Examples of such training tools range from the often-disparaged Microsoft agent, “Clippy” (the office assistant packaged with Microsoft Office products to offer ostensibly helpful tips), to highly realistic three-dimensional simulators used in corporate and federal training environments (e.g., the National Advanced Driving Simulator (NADS) and the Virtual Soldier Project, both housed at the University of Iowa). Considerable research attention has been paid to the accurate visual rendering of virtual actors and objects. Computer scientists and engineers have made important advances that allow them to render and manipulate virtual actors and objects in virtual environments in ways that very closely (if not completely) approximate how the actors and objects appear and move in physical environments. In contrast, the social representation of virtual actors has received less attention. Advances in artificial intelligence (particularly as it relates to cognition) have facilitated the development of avatars and agents that can respond to information inputs from human actors (e.g., queries, instructions). Yet, our pilot research shows notable differentials in human actors’ acceptance of and reaction to avatar- and agent-generated information/behavior when appropriate nonverbal responses (e.g., expressions, gesturing) from the avatar/agent are absent, compared to their acceptance of and reaction to the same information and behavior generated by another human actor who displays appropriate non-verbal responses. I propose that these and other differentials in how humans respond to and interact with virtual actors reflect a dearth of attention to modeling the social intelligence of actors in the design of virtual actors and virtual environments. While considerable advances have been made in the three-dimensional rendering

CH06-P374095.indd 96

5/26/2008 2:36:13 PM

A Strategy for Incorporating Social Intelligence

97

and artificial intelligence of virtual actors in virtual environments, very little systematic attention has been paid to modeling social intelligence and infusing virtual environments with social intelligence. I adopt and refine Thorndike’s (1920) definition of “social intelligence” as, “the ability to perceive one’s own and others’ internal states, motives, and behaviors and to act toward them optimally on the basis of that information” (see also Sternberg 1985). In the next section of this proposal, I step through a theoretical strategy for the development of a model of social intelligence in virtual reality. I start with an overview of social intelligence, which suggests that emotional intelligence is an important factor in individual’s abilities to successfully navigate social life. Next I offer brief reviews of two social psychological theories, ACT, and identity theory, which offer dynamic models of human social behavior. An important element of these theories is the role that affect plays in facilitating social interaction – sentiments act as barometers through which one’s self, other actors, objects, and behaviors are evaluated in social settings. I integrate these theories with a conceptualization of social risk, which provides the basis for incorporating social intelligence in the design of virtual environments and actors.

THEORETICAL FRAMEWORK In this section, I examine three lines of theory and research, which provide insight to facilitate more accurate modeling of virtual actors: (1) social intelligence, (2) ACT, and (3) identity theory. Research on social intelligence has primarily fallen under the purview of psychologists and, more recently, neurologists. In contrast, ACT and identity theory have emerged as perhaps the most succinct and well-developed theories of social identity and social action in the domain of social psychology. They integrate both psychological and sociological perspectives on identity and action. I begin with a discussion of social intelligence, which suggests the key role that emotions play in judgmental decisionmaking. ACT provides a dynamic formal model of emotion in social action, while identity theory contributes insight on factors that affect the extent to which an actor might become invested in a social role (and hence the extent to which the formal model may be borne out in virtual environments). SOCIAL INTELLIGENCE

The concept of “social intelligence” was first articulated by Thorndike (1920), who proposed that intelligence can be conceptualized as the ability to understand and mange three different elements: (1) ideas (abstract intelligence), (2) concrete objects (mechanical intelligence), and (3) people (social intelligence). Attempts to measure social intelligence soon followed (for a review, see Khilstrom and Cantor 2000). As psychometric instruments designed to operationalize social intelligence evolved, it became clear that emotions play a central role in social

CH06-P374095.indd 97

5/26/2008 2:36:13 PM

98

Social Structure and Emotion

intelligence. That is, researchers found that an important dimension of social intelligence involves attention to and interpretation of one’s own and others’ feeling states. Indeed, some researchers have argued that “emotional intelligence” is itself distinct from social intelligence. Yet, attempts to independently measure social and emotional intelligence have been elusive (Bar-On 2000). Path-breaking research by Damasio (1994) established the neurological basis of emotional intelligence and its role in decision-making, through the articulation of the somatic marker hypothesis. This hypothesis, which links emotion to neurological processes, suggests that feeling states trigger somatic states, which generate brain signals that affect both conscious and subconscious decisions and actions. The somatic hypothesis has led researchers to posit that cognitive intelligence is largely the domain of cortical function, while emotional and/or social intelligence falls into the domain of limbic function (Goleman 1995; Stein and Book 2000). A recent report by Bar-On et al. (2003) provides empirical support for this assertion. They compared brain-injured patients with bilateral lesions in the somatic circuitry to those with lesions outside this circuitry. Those with lesions in the somatic circuitry evidenced significantly lower emotional and social intelligence (despite having normal levels of cognitive intelligence) compared to those with lesions outside this circuitry. Consistent with Damasio’s somatic marker hypotheses, these patients also exhibited poor judgment in decision-making. While the debate over the independence of social and emotional intelligence (or whether they represent a single construct) has not been resolved, the purpose of this project is not to join in on the debate, much less resolve it. Rather, I will treat “social intelligence” as embodying emotional intelligence. What remains clear is that emotion is a critical and dynamic element of intelligence. Consequently, accurately modeling social actors requires a dynamic representation of social intelligence that incorporates emotion. In the next section, I describe such a theory, which synthesizes insights from sociological and psychological social psychology to generate a dynamic model of social interaction. AFFECT CONTROL THEORY2

As the above discussion indicates, a key requirement for a valid model of social intelligence is that it must dynamically represent emotion. That is, as social interaction unfolds, the model must account for shifts in the meanings an actor assigns to others, objects, and behaviors and the transitions in the feelings the actor experiences toward others, objects, and behaviors. Furthermore, quantifiable, algorithmic models are preferable to discursive models for the purposes of computer rendering. Within the field of social psychology, Heise (1987) has developed just such a model: ACT. This model portrays individuals as continually assessing meaning in social interaction. Moreover, according to the theory, actors seek consistency across interactions and that consistency is emotionally experienced. By mathematically modeling the dynamics of interaction, 2

This section draws heavily on Troyer and Robinson (2006).

CH06-P374095.indd 98

5/26/2008 2:36:13 PM

A Strategy for Incorporating Social Intelligence

99

ACT provides the machinery for generating concise predictions regarding current and future emotions and behaviors elicited by actors. To date, ACT has been applied to human actors and actions. Less attention has been paid by the theory to interaction with non-human objects and agents (although for recent exceptions, see Heise 2004; Troyer 2004). In this section, I summarize ACT, arguing that it provides a useful foundation for more accurate representations of social intelligence in virtual environments. Three important assumptions form the theoretical foundation for ACT: (1) individuals seek to enact behaviors that will generate feelings that are appropriate to the situation; (2) if individuals cannot maintain situation-appropriate feelings through their behavior, then they will adjust their perceptions of the situation; and (3) individual’s emotions signal the relationship between their experiences and their definitions of the situation. The theory portrays individuals as engaged in ongoing assessment of the meaning of the social interactions in which they are embedded and as agents who seek to maintain transituational consistency in the meanings that interactions evoke for them. I extend these basic premises of social interaction to understanding virtual interaction as a social situation. At the onset of interaction, actors assign elements of interaction to categories. For instance, humans are categorized into social roles (e.g., Army Enlistee), with shared expectations regarding actions appropriate for the role. Actions that we take toward and to one another are the defining markers of “social events.” Thus, ACT conceptualizes social events in terms of their constituent elements: (1) Actors who assume social identities (e.g., Army Enlistee), (2) the Behaviors of the actors (e.g., Attack), and (3) the Object-others to whom the actions are directed, which are also social identities (e.g., Enemy). These elements correspond to the linguistic structure we use to describe and understand situations, for instance, “Army Enlistee Attacks Enemy.” This case–grammar structure of events takes the form of Actor–Behavior–Object (ABO). Each element of the ABO structure has meaning, which is independent of any particular event and which is defined in three dimensions: Evaluation, Potency, and Activity (EPA). Evaluation reflects contrasts such as good versus bad. Potency reflects contrasts like powerful versus powerless. Activity reflects contrasts like lively versus quiet. Each dimension is measured on a scale from 4.5 to 4.5, with zero reflecting a neutral perception of the element on the dimension. Using evaluation as an example, a value of 1.0 represents an element that is slightly good, 2.0 one that is quite good, 3.0 is extremely good, and 4.5 represents something that is infinitely good. On the negative side of the scale, 1.0 reflects something that is slightly bad, 2.0 something that is quite bad, 3.0 is extremely bad, and 4.5 represents something infinitely bad. Thus, for example, using survey methods, researchers have found that “Army Enlistee” has an EPA rating of 1.39 (Evaluation), 0.88 (Potency), and 0.96 (Activity).3 This corresponds to the notion that “Army Enlistee” means something that is 3

The EPA ratings and other quantitative values we report were drawn from Heise’s software program, Java Interact 2 (Heise 2001), available at: http://www.indiana.edu/~socpsy/ACT’/.

CH06-P374095.indd 99

5/26/2008 2:36:13 PM

100

Social Structure and Emotion

slightly good, slightly powerful, and very slightly lively. The object “Enemy” has an EPA rating of 2.11, 0.75, and 0.19. These EPA values mean something is quite bad, slightly powerful, and neither active nor inactive. The action of “Attack” has an EPA rating of 2.6, 1.5, and 2.0, corresponding to an act that is quite-to-extremely bad, slightly-to-quite powerful, and quite lively. The independent EPA ratings of each element are called “fundamental sentiments.” The fundamental sentiments for an actor, behavior, or object reflect our affective associations or the meaning of the element to us and are quantified in the EPA dimensions. The independent elements Army Enlistee, Enemy, and Attack may combine as an event, “Army Enlistee Attacks Enemy.” The event elicits feelings toward the actor (Army Enlistee), action (Attacking), and object of action (Enemy). Since Attacking is quite bad and very lively, it elicits a negative impression of the actor. Yet, this negativity is somewhat expected because it is being inflicted upon a quite bad, slightly powerful person (Enemy). That is, while we may not generally expect good people to engage in bad actions, we do expect bad people to be the recipients of bad actions. In contrast, consider the event “Army Enlistee Helps Enemy.” This event generates a positive impression of the actor. Yet, it is unexpected, since bad people are generally not the recipients of good actions. Smith-Lovin (1987) developed mathematical models (referred to as impressionformation equations) to predict the feelings evoked toward ABO elements as the result of an event. These models mathematically combine the independent EPA ratings of elements comprising an event (referred to as “out-of-context” ratings) to estimate revised EPA ratings for each element when they are combined in an event (referred to as “in-context” ratings). An impression-formation equation predicting the new evaluation (i.e., goodness versus badness) of an actor following his or her participation in an event is given below:

Ae  c1 Ae  c2 Ap  c3 Aa  c4 Be  c5 Bp  c6 Ba  c7Oe  c8O p  c9Oa  c10 Aa Be  c11 Ba Oe  c12 Ap Bp  c13 BpO p  c14 Aa Ba  c15 Ae Bp  c16 Ae Ba  c17 Ap Be  c18 ApOa  c19 BeO p  c20 BpOe  c21 BpOa  c22 Ba Oe  c23 Ba O p  c24 Ae BeOe  c25 Ap BpO p  c26 Aa Ba Oa  c27 Ae BpO p  c28 Ap BpOa  ee

(6.1)

There are nine impression-formation equations of the form of Eq. (6.1) (one predicting each E, P, A rating for each A, B, O element). Thus, for example, the impression-formation equations predict that the EPA ratings for “Army Enlistee” in the event “Army Enlistee Attacks Enemy” become 1.82, 0.20, and 1.08. That is, the meaning attached to the enlistee who attacks enemy shifts to slightly bad, very slightly powerful, and slightly lively. These feelings are referred to as “transient impressions” by ACT. In contrast, the transient impression for the enlistee who helps an enemy shifts to the EPA profile of 1.06, 2.23, and 1.37 – a shift to slightly good, quite powerful, and slightly active.

CH06-P374095.indd 100

5/26/2008 2:36:13 PM

A Strategy for Incorporating Social Intelligence

101

Using these equations, ACT predicts how meanings shift as social interaction unfolds and, importantly, likely actions in response to the new meanings. Recall that for each event, there is a fundamental sentiment associated with actors and behaviors (out-of-context EPA ratings) and transient impression that results from the event (in-context EPA ratings). The difference between the fundamental sentiment and the transient sentiment is referred to as “deflection” in ACT. Deflection is operationalized as the sum of the squared differences between the fundamental sentiments associated with the elements (ABO) of the event and the transient impressions resulting from the event. Deflection corresponds to the perceived likelihood of an event (Heise and MacKinnon 1987). A deflection of 8.0 or less seems common and expected (“Army Enlistee Attacks Enemy” elicits a deflection of 6.0). Events generating deflection values of 9.0–16.0 seem unusual (the deflection for the event “Army Enlistee Helps Enemy” is about 9.0); those with deflection values of 17.0–24.0 seem extraordinary. Experiencing an event with a deflection of 25.0–44.0 seems incredible, and those with deflections of 45.0 and higher seem so unbelievable that they elicit feelings commensurate with a supernatural experience. Two features of ACT allow it to predict future events. First, the theory assumes that actors seek stability in meanings and take actions to generate such stability; they look for ways to minimize deflection. Second, three decades of ACT research have yielded a “dictionary” of EPA ratings for thousands of ABO elements. The assumptions, quantitative calculus operationalizing interaction, and dictionary of EPA ratings allow researchers to predict both how actors will emotionally respond in social interactions and their subsequent reactions. The most likely reactions (and hence future events) are those that reduce deflection arising from the current event. As this suggests, social interaction is itself a dynamic process and emotion is its compass – it is the appraisal of one event that determines subsequent events. I propose that this framework can also be applied to modeling social intelligence in virtual environments. Dictionaries and mathematical models of ACT afford a high degree of precision and predictive capability. The models and dictionaries are combined in a software program, Interact (Heise 2001). This software allows researchers to simulate social interaction events and generate testable predictions regarding the sequences of events and the feelings actors will experience. Yet, ACT’s focus is largely limited to interpersonal (i.e., humanto-human interaction). Recently, Heise (2004) has described how ACT’s arguments on the emotional responses of actors can be incorporated into the design of computer agents. In another paper, I describe how the theory can also be used to facilitate the design of computer interfaces and software (Troyer 2004). These initial statements (while they remain to be thoroughly tested) are suggestive of the broad applicability of ACT to social settings, whether they are physical or virtual. While ACT provides the machinery for dynamically modeling social intelligence, further theorizing in social psychology suggests that the extent to which

CH06-P374095.indd 101

5/26/2008 2:36:13 PM

102

Social Structure and Emotion

such social intelligence may be called and acted upon by an actor depends on the salience of any particular identity she or he holds in the social context. While ACT recognizes that individuals are capable and do “wear many hats” in their daily lives (i.e., take on different role-identities, such as mother, sister, teacher, soldier), the formal model begins with the assumption that an actor does hold a particular identity at the onset of interaction. The theory does not address the fact that individuals may be more or less invested in a particular identity at any moment in time (nor does it address factors which may affect such investments). This is not a shortfall of the theory (it is unreasonable to expect any single theory to address all of social existence). Rather it represents an opportunity for expanding the theory, while also being an important consideration for the application of ACT to virtual environments, insofar as actors may vary in their investments in the roles they assume in virtual environments as they transition into and out of them. Identity theory (Stryker and Burke 2000) provides important insights on the investments that actors make in the role-identities that they assume. I outline this theory in the next section. IDENTITY THEORY

As noted above, ACT posits that actors take on an identity in the context of a social event. This identity, however, is malleable – it can change as the interaction unfolds. The precision of the theory’s model of the dynamics of social interaction, however, may hinge on the actors holding some minimal investment in the role-identity. What factors affect such an investment? Identity theory (Stryker and Burke 2000) provides insight to help us address this question. Identity theory posits that an actor’s sense of self is a function of the identities she or he assumes in social interaction (Stryker and Burke 2000). As in ACT, identities are conceptualized as positions in a social system to which unique constellations of expectations for behavior and attitudes are assigned. For instance, the identity of “mother” calls up expectations of a caring, nurturing female. The identity of “teacher” calls up expectations of an individual who is wise and assists others. A person’s identity varies, depending on the interaction context and importantly, the others with whom one is interacting in the context. Also similar to ACT, identity theory explicitly recognizes that individuals can assume different identities at different points in social interaction. That is, actors possess a kind of “toolkit” of identities – different hats that can be worn, depending on the situation. Unlike ACT, however, identity theory offers insight into the factors that lead actors to enact particular identities. This is particularly useful for the design and implementation of virtual environments. According to identity theory, actors organize their identities in a hierarchical fashion. Identities that are at the top of the hierarchy are more salient – they are more salient to the actor and more likely to be enacted. Furthermore, conforming to the expectations associated with an identity becomes more important, the more salient the identity is. Salience (or ranking of an identity in the

CH06-P374095.indd 102

5/26/2008 2:36:13 PM

A Strategy for Incorporating Social Intelligence

103

identity hierarchy), in turn, depends on an actor’s commitment to the identity. Commitment refers to an actor’s investment in a particular identity, which results from relationships an actor has with others that rely on the identity. For example, if an individual has many interactions with scientific collaborators and highly values those individuals and interacts with them, then the identity “researcher” is likely to occupy a high place in the individual’s identity hierarchy. Because of this, it is likely to be very salient to the individual – the expectations and attitudes associated with the identity will be easily accessible and, thus more likely to be enacted, than less accessible and salient repertoires of behavior associated with other identities. Thus, identity theory suggests that the salience of a particular identity (and hence the extent to which the expectations associated with it will be enacted) is a function of an individual’s commitment to it. Commitment, in turn, is a function of the importance of the social relations one has with others that depend on the particular identity. This suggests the importance of designing virtual environments that take into account identity theory’s insights on salience and commitment. The next section develops this point further. IDENTITY COMMITMENT, SOCIAL RISK, AND SOCIAL REALISM

Thus far, I have proposed that ACT represents a mathematical foundation for more accurately modeling virtual actors. Furthermore, drawing on identity theory, I have proposed that some minimal level of commitment to the identities’ actors assume in virtual interactions is required for such an ACT model. As noted above, this requires examining the social relations (and their importance) an actor has with others and ensuring that the same relations (and importance) are approximated in the virtual setting. An important element of identity theory’s notion of commitment is the idea that there is a cost to violating the expectations associated with a role-identity. As Burke (1991) has noted, that cost is manifested in the form of stress, manifested, in part, through negative affective states. Burke’s elaboration of identity theory, identity control theory, suggests that actors scan their social environment for cues regarding whether they are meeting the expectations others hold for them as they enact their social roles. The greater an actor’s commitment to a social role, the greater the stress the actor experiences when she or he receives indications of expectation violations. Returning to the foundational work of Stryker (Stryker and Burke 2000), this suggests that virtual environments placing actors at risk of jeopardizing relations with important others may enhance the extent to which virtual interaction approximates face-to-face interaction. That is, for example, if actors believe that their performance in the virtual environment carries a social risk – the potential to compromise or be stripped of an important identity this is also held outside the virtual environment – then they will be more inclined to attend and respond to the cues of virtual actors within the

CH06-P374095.indd 103

5/26/2008 2:36:13 PM

104

Social Structure and Emotion

virtual environment. Thus, not only the content, but also the execution of virtual environments is important to consider when attempting to emulate behavior in face-to-face settings. Additionally, however, the social realism of the environment plays a key role. According to Burke’s formulation (1991), responses (verbal, non-verbal, and paraverbal) from others to an individual’s social actions provide an important gauge that actors use to assess whether their actions are aligned with others’ expectations or are in violation of them. Facial expressions that actors give off represent a critical type of response. Emotional expressions of surprise, frustration, or anger may signal expectation violations, while nods and smiles signal conformity. This is where social realism takes on a profound importance in the design of virtual environments. To date, considerable attention has been paid to rendering three-dimensional actors whose physical movements are fluid and correspond to actual movements of human actors. Researchers have only recently begun to attend to the role that expressions play in achieving social realism. Blascovich and colleagues (Blascovich et al. 2002) have conducted pioneering work on the utility of virtual environments for exploring social psychological processes. They propose that the virtual environment technologies offer a promising tool for social psychological research insofar as they allow researchers to enhance the mundane realism of experiments, without sacrificing control. Perhaps more importantly, however, Blascovich et al. document critical capabilities of the technology that have only recently made social psychological research using this technology possible. Chief among these capabilities is the ability to render high-resolution photo-realistic virtual actors. Software developed in the laboratories of Blascovich and colleagues capitalizes on these capabilities and their potential to enhance social psychological research (http://www.worldviz. com). The software allows researchers to realistically model and manipulate human facial expressions (in addition to motor movements of the lower and upper extremities). This technical capability to render virtual actors that are highly realistic in visual appearances complements the theoretical work of Blascovich et al. (2002). They draw on work by Heeter (1992) to conceptualize the social realism of virtual reality in three social psychological dimensions: social presence, personal presence, and environmental presence. Social presence involves the extent to which one believes she or he is in the presence of other human actors. Personal presence refers to the degree to which a human actor believes that she or he exists in the virtual environment. Environmental presence refers to the extent to which the human actor believes that the virtual environment (including other virtual actors) is responsive to the behavior of the human actor. Blascovich et al. focus on social presence as a key dimension and capture environmental presence through the concept of behavioral realism – the extent to which virtual actors and objects in virtual realities behave as they would in a face-to-face world of human actors and physical objects. They propose that social psychological processes in virtual environments will approximate those of face-to-face environments to

CH06-P374095.indd 104

5/26/2008 2:36:13 PM

A Strategy for Incorporating Social Intelligence

105

the extent that both behavioral realism and social presence are high. Thus, the combination of behavioral realism and social presence, I propose, determines the “social realism” of the environment – the extent to which actors perceive the setting as approximating a face-to-face setting. As I have suggested above, social realism affects the extent to which the emotional feedback from others (expressed through verbal, paraverbal, and non-verbal channels) becomes a signal of conformity to or deviance from the expectations associated with a social identity. As such, it exacerbates the social risk that is entailed in social interaction: Interaction under conditions of greater social realism carries greater social risk (i.e., threat of violation and loss of identity) compared to interaction under conditions of less social realism. Affective responses will be stronger when social risk is heightened. Thus, an important factor in triggering the affective states that engender normative and the realistic social reactions (and hence, behavioral reactions) in human actors depends on social realism embodied in the virtual environment. Thus, the combination of a situation that matches an actor’s commitment to an identity in virtual interaction to the actor’s commitment to the same identity in a comparable face-to-face interaction along with high social realism increases the likelihood that behaviors and attitudes exhibited in the virtual environment will correspond to those exhibited in face-to-face interaction.

CONCLUSION In summary, by incorporating both ACT and identity theory in the design and execution of virtual environments, we can more fully represent and model social intelligence, creating a much closer approximation of virtual interaction to interaction between human actors in face-to-face contexts. As I have described, ACT provides the bases for infusing virtual actors with social intelligence. The mathematical models of ACT are particularly amenable to being incorporated into the design of virtual agents. Identity theory and identity control theory provide insight into the development designing virtual environments in a manner that will take advantage of the growing ability to capitalize on social realism to approximate the roles that commitment and social risk play in determining social action.

REFERENCES Bar-On, R. 2000. “Emotional and Social Intelligence: Insights from the Emotional Quotient Inventory.” In: R. Bar-On and J. D. A. Parker (eds), The Handbook of Emotional Intelligence, pp. 363–388. San Francisco: Jossey-Bass. Bar-On, R., D. Tranel, N. L. Denburg, and A. Bechara. 2003. “Exploring the Neurological Substrate of Emotional and Social Intelligence.” Brain 126:1790–1800. Blascovich, J., J. Loomis, A. C. Beall, K. R. Swinth, C. L. Hoyt, and J. N. Bailenson. 2002. “Immersive Virtual Environment Technology as a Methodological Tool for Social Psychology.” Psychological Inquiry 13:103–124.

CH06-P374095.indd 105

5/26/2008 2:36:13 PM

106

Social Structure and Emotion

Burke, P. J. 1991. “Identity Processes and Social Stress.” American Sociological Review 56:836–849. Damasio, A. R. 1994. Descartes’ Error: Emotion, Reason, and the Human Brain. New York: G.P. Putnam. Farrell, K. and R. T. Marier. 2004. “Optimization-Based Kinematic Models for Human Posture and Motion,” Technical Report #VSR-04.11, University of Iowa Virtual Soldier Research Program. Goleman, D. 1995. Emotional Intelligence. New York: Bantam Books. Heeter, C. 1992. “Being There: The Subjective experience of Presence.” Presence: Teleoperators and Virtual Environments 1:262–271. Heise, D. R. 1987. “Affect Control Theory: Concepts and Model.” Journal of Mathematical Sociology 13:1–33. Heise, D. R. 2004. “Enculturating Agents with Expressive Role Behavior.” In: S. Payer and R. Trappl (eds), Agent Culture: Human–Agent Interaction in a Multicultural World, pp. 127–142. Mahwah, NJ: Lawrence Erlbaum. Khilstrom, J. F. and N. Cantor. 2000. “Social Intelligence.” In: R. J. Sternberg (ed.), Handbook of Intelligence, 2nd edition. pp. 359–379. Cambridge, UK: Cambridge University Press. Smith-Lovin, L. 1987. “Impressions from Events.” Journal of Mathematical Sociology 13:35–70. Stein, S. J. and H. E. Book. 2000. The EQ Edge: Emotional Intelligence and Your Success. Toronto: Stoddart Publishing. Sternberg, R. J. 1985. Beyond IQ: A Triarchic Theory of Human Intelligence. New York: Cambridge University Press. Stryker, S. and P. J. Burke. 2000. “The Past, Present, and Future of an Identity Theory.” Social Psychology Quarterly 63:284–297. Thorndike, E. L. 1920. “Intelligence and Its Uses.” Harper’s Magazine 140:227–235. Troyer, L. 2004. “Affect Control Theory as a Foundation for the Design of Socially Intelligent Systems.” In: Proceedings of 2004 American Association for Artificial Intelligence Symposium on Architectures for Modeling Emotion: Cross-Disciplinary Foundations, pp. 27–30. Menlo Park, CA: AAAI Press. Troyer, L. and D. Robinson. 2006. “Contributions of a Microsociological Perspective on Emotion to the Study of Political Identity and Action.” In: D. Redlawsk and G. R. Boynton (eds), Affect and Cognition in Political Action.

CH06-P374095.indd 106

5/26/2008 2:36:13 PM

7 Emotions on the Job: Supporting and Threatening Face in Work Organizations Clare Francis1 and David R. Heise2 1

University of North Dakota, Grand Forks, ND, USA 2 Indiana University, Bloomington, IN, USA

ABSTRACT We examine the production of emotions in work situations, focusing on two aspects of individual’s face important in the management literature: moral character and job competency. Study 1 acquired respondents’ emotional responses to 54 vignettes in which one or the other aspect of respondent’s face was supported or threatened by an interaction partner of higher, lower, or equal rank, in the context of either a consultancy firm or a hospital. Study 2 acquired ratings of self feelings from respondents participating in simulated activities of a bookselling firm, wherein a considerate, average, or quarrelsome manager supported or threatened one or the other aspect of respondent’s face. Having face supported induced pleasant emotions, and unpleasant emotions resulted from having face threatened. The specific kind of emotion depended on whether moral character or job competency was involved, on the rank of the interaction partner, on organizational context, and on interactions of all these. Key Words: Status, Identity, Face-work

INTRODUCTION The notion of social face was introduced millennia ago both in Asian texts (Hu 1944) and in Greek and Hebrew texts (Botterweck et al. 1974; Friedrich 1967). Social Structure and Emotion

CH07-P374095.indd 107

107

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/27/2008 7:22:48 PM

108

Social Structure and Emotion

Analyses of social face have continued in numerous contemporary disciplines, including sociology (Goffman 1967), management (Tjosvold and Huston 1978), communications (Brown and Levinson 1978), and conflict analysis (TingToomey et al. 1991). The concept of face is richly articulated for management researchers in organizational face theory (Earley 1997). In this study, we adopt Earley’s (1997, p. 43) definition of face: the evaluation of self based on internal and external (to the individual) judgments concerning a person’s adherence to moral rules of conduct [moral character form of face] and position within a given social structure [social status form of face].

This definition allows both positive and negative face, which corresponds to the fact that social encounters present both potential support to and potential threats to face. Additionally Earley’s definition specifies two face forms derived from 4th century BC Chinese culture (Hu 1944), moral character, and social status. Moral character places emphasis on personal values that are reflected in behaviors. Social status emphasizes job competence, success and the accumulation of possessions such as status and wealth.

FACE AND EMOTION Limited research suggests links between face and emotion. In Chinese culture face is an explicit phenomenon, quite salient, and believed to explain much of social behavior (Stover 1974). Redding and Ng (1982) in a study including 102 Hong Kong executives found that 94% felt face was important in many business areas (e.g., sales, publications, employee relations, peer relationships, and negotiation activities), 100% felt that “having face influenced success in business,” and 97% felt that not having face “strongly negatively influenced success.” Participants indicated face-gain feelings included satisfaction, pride, and confidence; and face-loss feelings included shame, unease, and difficulty in concentrating on work. Tjosvold and Huston (1978) found that bargaining subjects who were affronted by the negotiation partner compared to those who were not affronted felt less warmth for the negotiating partner and less satisfaction in their own negotiating performance than subjects who were not affronted. Our own research focuses on the relationship between face and emotion, particularly in contexts of different kinds of organizations and social relationships. Our work is instigated by the caution to managers that “criticism risks causing one to lose face” (Bond and Lee 1981, p. 293) – presumably with accompanying deleterious emotions, yet criticism and accusation often are integral aspects of maintaining organizations. As a step toward clarifying this anomaly, we explored how emotions arise from face-related behavior by others. In particular, we examined how face behaviors – threats to and support for one’s face – interact with social interactional and organizational variables in generating specific kinds of emotions.

CH07-P374095.indd 108

5/27/2008 7:22:48 PM

Supporting and Threatening Face in Work Organizations

109

STUDY 1: VIGNETTES In our first study we presented respondents with a series of vignettes and asked them to indicate what emotion they would feel when confronted with the situation described in the vignette. Our analyses focus on the following questions. Does the kind of face behavior occurring in various situations explain differences in a respondent’s emotional responses to the situations? For a given kind of encounter, does the rank of the other interactant – superordinate, peer, subordinate – influence a respondent’s emotional response? For a given kind of encounter, does organizational setting influence a respondent’s emotional response? PROCEDURES

Design of Vignettes: We composed vignettes to convey realistic work-world experiences, while implementing concepts in the face and organizational literature. A factorial design varied type of organization, valence of the encounter, aspect-of-face involved in the encounter, the relative rank of the other person in the encounter, and the specificity of the other’s identity. Clarity and realism of the stimuli were honed in two preliminary pilot tests with undergraduate students at Indiana University. Each factor contained two levels, except rank which had three. Thus the factorial design generated 48 situations. An additional six neutral situations were included in the study (as noted below under Valence) yielding a total of 54 vignettes for presentation to each respondent. These are listed in Table 7.1. The factors were defined as follows. Organizational Setting: One group of situations was set in a consulting firm by prefacing each situation with the following text: You are Terry Nolan, an honest senior consultant and financial planning expert in a management consulting firm. Your superior, Chris Green, is a manager. Your colleagues, other senior consultants, are often your teammates on projects.

The rest of the situations were set in a hospital. The preface in this case was the following: You are Pat Wheeler, an honest registered nurse on the orthopedics unit at a hospital. You have been there for five years and your subordinates are five nurse aides. Your nurse manager is Robin Smith and you also work with five physicians.

Valence: One set of situations presented a threat to the respondent’s face. For example, following the preface for a consultancy, the next sentence might be “Your clerk accuses you of theft of company communications equipment.” A second set of situations supported the respondent’s face. For example, following the hospital preface, the next sentence might be “Your supervisor congratulates you on your alertness in reviewing medical records.” In the hospital setting, respondents were presented with a third set of situations that were not part of the factorial design. These were neutral, informatory

CH07-P374095.indd 109

5/27/2008 7:22:48 PM

CH07-P374095.indd 110

⫺0.40 ⫺0.26 ⫺0.52 ⫺0.09 ⫺0.52 ⫺0.14 ⫺0.67 ⫺0.44 ⫺0.22 ⫺0.51 ⫺0.34 ⫺0.37 1.48 1.55 1.84 1.89 1.75 1.49 1.97 1.68

⫺1.16 ⫺1.44 ⫺1.17 ⫺1.23 ⫺0.80 ⫺0.53 ⫺1.10 ⫺0.85 ⫺0.80 1.71 1.80 2.31 2.28 2.19 1.63 2.34 2.23

Potency

⫺1.53 ⫺0.98 ⫺1.49

Evaluation

Males, N ⫽ 135

1.24 1.15 1.59 1.27

0.56 ⫺0.93 ⫺0.55 ⫺0.60 ⫺0.66 ⫺0.53 ⫺0.63 1.07 1.16 1.31 1.36

0.64 ⫺0.02

0.54 0.10 0.16

2.80 2.15 2.90 2.76

⫺1.87 ⫺1.92 ⫺1.67 ⫺1.11 ⫺2.06 ⫺1.51 ⫺1.92 2.19 2.60 2.85 2.80

⫺1.93 ⫺2.14

⫺2.14 ⫺1.92 ⫺2.18

2.28 1.88 2.37 2.12

⫺0.05 ⫺1.03 ⫺0.82 ⫺0.51 ⫺0.97 ⫺0.63 ⫺1.04 1.99 2.23 2.34 2.35

⫺0.19 ⫺0.49

⫺0.28 ⫺0.46 ⫺0.46

Potency

1.66 1.36 1.96 1.51

0.86 ⫺0.86 ⫺0.49 ⫺0.38 ⫺0.71 ⫺0.27 ⫺0.75 1.53 1.79 1.71 1.95

0.72 0.38

0.67 0.23 0.35

Activity

Females, N ⫽ 100 Evaluation

Means

Activity

Mean Evaluation, Potency, and Activity Coordinates of Emotional Responses to Work World Vignettes

Your manager accuses you of stealing computer equipment from storage. Your superior discredits you by charging you with understating vacation days taken. Your teammate discredits you by charging you with taking time off without reporting it. Your colleague accuses you of stealing one of the firm’s laptop computers. Your receptionist discredits you by charging you with being dishonest on your time report. Your clerk accuses you of theft of company communications equipment. Your manager opposes you about your conclusions in the report you prepared. Your superior contradicts you about your analysis of the client’s financial results. Your teammate contradicts you about your interpretation of an office policy change. Your colleague opposes you about the information you sent to co-workers. Your receptionist contradicts you about your estimated time to complete the report. Your clerk opposes you about your communication approach for the interview. Your manager commends you for your integrity in dealing with a deviant employee. Your superior compliments you on your ethical behavior with vendors. Your teammate compliments you on your ability to deal honestly with suppliers. Your colleague commends you for your high moral conduct in dealing with contractors. Your receptionist compliments you on your trustworthiness. Your clerk commends you for your integrity in confronting a corrupt vendor. Your manager supports you in your leadership abilities. Your superior backs you in your approach to scheduling projects.

Consultancy

TABLE 7.1

110 Social Structure and Emotion

5/27/2008 7:22:48 PM

CH07-P374095.indd 111

Your manager incriminates you with a charge of unjust treatment of nurse aides. Your supervisor is disappointed with you about the charges against you for the missing narcotics. Your teammate incriminates you with a charge that you diminished employee dignity. Your co-worker is disappointed with you about the charge that you took controlled drugs. Your aide incriminates you with a charge of mistreatment of food service workers. Your subordinate is disappointed with you about the charges against you for theft of controlled drugs. Your manager criticizes you for the clinical analysis you prepared. Your supervisor blames you for your units analysis errors due to your lack of attention. Your teammate criticizes you about documentation problems she feels you missed in clinical procedures. Your co-worker blames you for misstatement of your group’s clinical report due to your lack of skills. Your aide criticizes you about the tone of your report. Your subordinate blames you for your team’s poor patient outcomes due to your lack of ability. Your manager is pleased with you for your integrity in a controversial issue. Your supervisor praises you for your honesty about safety issues and hazardous materials. Your teammate is pleased with you for your moral character in dealing with unethical vendors. Your co-worker praises you for your truthfulness in a difficult patient matter.

Hospital

Your teammate backs you about your design for developing training programs. Your colleague supports you in your talent in building client relations. Your receptionist backs you in your steps to motivate the staff. Your clerk supports you in your computer technology skills.

⫺0.82 ⫺0.50 ⫺0.84 ⫺0.69 ⫺0.89 ⫺0.69 ⫺0.64 ⫺0.41 ⫺0.66 1.85 1.90 1.75

⫺1.54 ⫺1.56 ⫺1.59 ⫺1.29 ⫺1.50 ⫺1.29 ⫺1.51 ⫺0.97 ⫺1.52 2.30 2.33 2.16

1.83

⫺0.46

⫺1.56

2.30

⫺0.54 ⫺0.86

1.49 1.68 1.64 1.60

⫺1.47 ⫺1.64

1.90 2.03 2.12 2.10

1.30

1.26

1.35 1.40

⫺0.68 ⫺0.29

⫺0.19

⫺0.89

⫺0.74 ⫺0.66

⫺0.01 ⫺0.43

⫺0.52

0.16

0.03 ⫺0.43

1.08 1.24 1.06 1.09

2.83

2.76

2.94 2.84

⫺1.69 ⫺2.24

⫺2.19

⫺1.92

⫺2.03 ⫺2.33

⫺2.16 ⫺2.32

⫺2.20

⫺2.16

⫺2.17 ⫺2.22

2.54 2.75 2.66 2.71

2.26

2.27

2.44 2.35

⫺0.82 ⫺0.84

⫺0.74

⫺0.97

⫺1.11 ⫺1.27

⫺0.51 ⫺1.07

⫺0.99

⫺0.48

⫺0.50 ⫺1.02

2.08 2.22 2.07 2.04

(Continues)

1.64

1.75

1.88 1.85

⫺0.69 ⫺0.35

⫺0.19

⫺0.69

⫺0.99 ⫺0.97

0.23 ⫺0.79

⫺0.51

0.34

0.25 ⫺0.67

1.69 1.68 1.50 1.33

Supporting and Threatening Face in Work Organizations

111

5/27/2008 7:22:49 PM

CH07-P374095.indd 112

(Continued)

1.69 1.78 1.55 1.96 1.38 1.84 1.34 1.85 0.81 0.48 0.85 0.82 0.80 0.58

1.91 2.31 1.27 0.75 1.29 1.26 1.23 0.97

Potency

2.17 2.26 2.10 2.31 1.89 2.29

Evaluation

Males, N ⫽ 135

1.35 ⫺0.10 ⫺0.65 ⫺0.11 ⫺0.34 ⫺0.32 ⫺0.52

0.57

1.16 1.17 0.99 1.54 0.74 1.41

Activity

2.84 1.21 0.91 1.68 1.65 1.58 1.64

2.51

2.77 2.83 2.72 2.90 2.65 2.91

2.28 0.87 0.52 1.13 0.99 1.02 1.01

1.82

2.18 2.25 2.11 2.42 1.99 2.36

Potency

1.75 0.06 ⫺0.62 0.04 ⫺0.43 ⫺0.22 ⫺0.45

1.05

1.62 1.70 1.50 1.94 1.28 1.83

Activity

Females, N ⫽ 100 Evaluation

Means

Note: Within each setting the first 12 events are threatening, the next 12 supportive; the final 6 hospital events are informatory. Within threatening and supportive blocks, the first 6 relate to moral character, the second 6 to competence. Within each of the six-item blocks, the first two involve alters who are superordinates, the middle two peers, and the final two subordinates. Each pair of alters is represented by a relatively specific role identity first, then a relatively generic role identity.

Your aide is pleased with you for your truthfulness in a difficult situation. Your subordinate praises you for your honesty about safety issues. Your manager agrees with you about your interpretation of a new regulation. Your supervisor congratulates you on your alertness in reviewing medical records. Your teammate agrees with you about your views on the outcome of new rules. Your co-worker congratulates you on your attention to detail in patient charting procedures. Your aide agrees with you about your approach to communicate the technical bulletin. Your subordinate congratulates you on your analysis of patient chart formats. Your manager listens to you as you describe the schedule changes. Your supervisor talks to you about having the meeting minutes distributed. Your teammate listens to you talk about the work shift revisions. Your co-worker talks with after the meeting about the project schedule. Your aide listens to your description of the new rules for nurses. Your subordinate talks to you about the agenda before the conference.

TABLE 7.1

112 Social Structure and Emotion

5/27/2008 7:22:49 PM

Supporting and Threatening Face in Work Organizations

113

situations that were neither threatening nor supportive – for example, “Your co-worker talks with you after the meeting about the project schedule.” Aspect of Face: Two different aspects of the respondent’s face were threatened or supported in the factorial-design vignettes. The respondent’s moral character was implicated by references to alleged illegal or unethical activities by the respondent or by references to the respondent’s honesty and integrity. Job competency (or the social status aspect of face as it is called in the management literature) was evoked by reference to the respondent’s work activity in the organization – reporting, analyzing, interpreting, designing, leading, applying technology. Alter’s Rank: The other person involved in the interaction had one of three ranks, relative to the respondent: superordinate, peer, or subordinate. Alter’s Specificity: The other person in the interaction had a fairly specific job title, or a more generic one. The specificity factor is not of prime interest in this study, and was included as a means of producing replications within other cells of the design. INSTRUMENTATION

Data were collected by means of a Java program running on a World Wide Web page, visited by respondents via their Internet browser. Respondents interacted with the program until they completed the questionnaire, whereupon their data were stored via the Internet on a university computer. The Java program, written especially for this project, employed an emotion spiral (Heise and Calhan 1995) to collect data about feelings, and multiple-choice questions to collect demographic data. Emotion Spiral: At the end of each vignette’s text, the respondent was asked, “How do you feel at that moment?” Respondents recorded their emotions on the emotion spiral displayed in Figure 7.1. The shaded box in the middle of the spiral presents the following explanation of the emotion spiral when the spiral first appears on respondents’ computer screens. This box will describe various situations. Click on the emotion closest to your feeling when imagining each situation. You’ll find pleasant emotions at the top, active emotions at the right, and dominant emotions on the outside of the spiral. “No emotion” is just above this box. You also can indicate low or high emotional intensity by using the pull down menu above.

The 25 emotional states given as choices were annoyed, ashamed, at ease, bitter, calm, contented, crushed, depressed, disgusted, embarrassed, excited, flustered, furious, happy, joyless, mad, nervous, no emotion, outraged, overjoyed, pleased, proud, scared, thrilled and unhappy. These constitute a representative sample of the approximately 100 named emotional states in English. Their arrangement along the spiral approximately represents the average Evaluation, Potency, and Activity (EPA) ratings of the emotions by respondents in 1980s studies. Mood: The explanation of the emotion spiral was followed by the following instruction: “Practice the procedure by indicating the emotion you feel right now.

CH07-P374095.indd 113

5/27/2008 7:22:49 PM

114

Social Structure and Emotion

Happy

Proud

Pleased

Contented

Overjoyed Average intensity

At ease

Calm

Joyless

Unhappy

Thrilled

No emotion This box will describe various situations. Click on the emotion closest to your feeling when imagining each situation. You’ll find pleasant emotions at the top, active emotions on the right, and dominant emotions on the outside of the spiral. “No emotion” is just above this box. You also can indicate low or high emotional intensity by using the pull down menu above.

Excited

Flustered

Done

Nervous

Practice the procedure by indicating the emotion you feel right now. Click the Done button when you are ready to go on.

Scared Outraged

Depressed Crushed

Embarrassed Ashamed

Bitter

Furious Disgusted

Mad Annoyed

FIGURE 7.1

Emotion spiral measuring instrument.

Click the Done button when you are ready to go on.” Responses to this instruction assess the respondent’s initial affective state. The emotion spiral presented the following verbal bridge in its text box as a transition from consultancy situations to hospital situations. The next set of scenarios is based on a hospital setting. Again, click on the emotion closest to your feeling when imagining each situation. Use the pull down menu to indicate emotional intensity, and click the Done button when you are ready to go on. To continue, indicate the emotion you feel right now, and click the Done button.

We used Responses to this instruction to assess the respondent’s midway affective state. We converted the respondent’s initial and midway affective states to Evaluation–Potency–Activity coordinates – as detailed below, and averaged the two EPA profiles to obtain a quantitative measure of the respondent’s overall mood while engaged in the research task. PARTICIPANTS

Two hundred thirty five undergraduate students in a junior level business course at Indiana University participated in the study to satisfy a course research requirement. Demographics were: 42% females and 58% males; 54% 20 years

CH07-P374095.indd 114

5/27/2008 7:22:49 PM

Supporting and Threatening Face in Work Organizations

115

old or less, 44% 21 to 25 years, and 2% 25 to 30 years old; 84% Caucasian, 12% Asian or Asian-American, and 4% other minorities. EMOTION COORDINATES

Respondents specified their emotions qualitatively in terms of the 25 emotion options in the emotion spiral. We quantified their choices by substituting for each emotion word the mean rating of the emotion on EPA scales. The EPA ratings were obtained in another study at the same university during the years 2002–2003, employing technology described by Heise (2001), and the mean ratings are publicly available in a Java program running on the World Wide Web (Heise 1997). We used means of male ratings to quantify male emotion choices, and female means to quantify female choices. Female and male means correlate highly: 0.98, 0.97, 0.97 for Evaluation, Potency, and Activity, respectively. However, female means are somewhat more extreme than male means: the female Evaluation means being 1.2 the size of male means on the average, while female Potency and Activity means are 1.1 the size of male means. To illustrate our procedure, if a male respondent chose the emotion “pleased,” with average intensity, as a response to a situation, then we recorded the three numbers 2.63, 1.71, 1.12 as the respondent’s data for that situation, these numbers being males’ mean EPA of “feeling pleased.” For female respondents, we recorded the numbers 3.17, 2.42, 1.43, which are females’ mean EPA ratings of “feeling pleased.” Respondents’ responses were 86% average intensity, 5% were low intensity, and 9% were high intensity. When respondents indicated low intensity, we halved the numbers for the emotion’s EPA profile. For example, a male response of “pleased” with low intensity was recorded as 1.315, 0.855, 0.560. When respondents selected high intensity, we recorded one and a half times the emotion’s EPA profile. For example, a male response of “pleased” with high intensity was recorded as 3.945, 2.565, 1.680. This implementation of the intensity variable derives from viewing emotions as directions in the EPA space. To have a low intensity emotion is to go only a little way in the emotion’s direction; to have a high intensity experience is to go extra far in the emotion’s direction. RESULTS

We employed within-subjects analyses of variance (ANOVA) to analyze the 48 situations of the factorial design. This form of analysis is possible because each respondent recorded an emotional response for every situation. Therefore, an EPA component of emotional response can be considered a repeated measure, where the repetition occurs across cells of the factorial design. Mood varied substantially across respondents participating in the study: 23% were in unpleasant states on the average, 29% neutral, and 48% in pleasant states. In preliminary analyses, we assessed the impact of respondents’ moods

CH07-P374095.indd 115

5/27/2008 7:22:49 PM

116

Social Structure and Emotion

on respondents’ emotion choices. The EPA components of mood were entered as covariates into ANOVAs of respondents’ emotion choices, the mood Evaluation being the covariate in an analysis of emotion Evaluation, the mood Potency in an ANOVA of emotion Potency, the mood Activity in an ANOVA of emotion Activity. No main effect or interaction involving mood was significant at the 0.05 level in the Evaluation or Potency analyses, and just one interaction was significant in the Activity analysis. Since this number of significant outcomes is no greater than would be expected by chance, we conclude that mood did not affect respondents’ choices in this study, and subsequent analyses forego consideration of respondents’ moods. Table 7.1 presents the male and female means of the EPA components of emotional response for each of the 54 situations presented to respondents. The top 48 rows in this table can be examined to investigate the highest order interaction in our design – setting * valence * aspect of face * rank of alter * specificity * sex. The organization of the table in terms of the factors is described in the table’s footnote. To allow for possible correlations among the EPA components of emotional response, we analyzed all three components simultaneously in a multivariate repeated measures design (alternatively called a doubly multivariate withinsubject design). Sex was included as a between-subjects factor. The two levels of alter’s specificity were treated as replications. Multivariate tests of significance for main effects and interactions are reported in Table 7.2. We display tests based on Pillai’s Trace, but significance levels are nearly identical using other criteria reported by the software used in this analysis – SPSS 13. All of the main effects and many of the interaction effects are significant at the 0.05 level. The highest level interaction – setting * valence * aspect of face * rank of alter * sex – is not significant ( p ⬍ 0.27). However, the corresponding interaction without sex is significant ( p ⬍ 0.001), and therefore our interpretation focuses on the constituents of this interaction. The means of cells involved in this interaction, and the 95% confidence intervals for the means, are displayed in Table 7.3. In our interpretations, we report two emotions closest to the EPA profiles in Table 7.3, restricting choices to emotions in the emotion spiral. Distances between Table 7.3 profiles and emotion profiles are based on ratings of emotions from our 2002–2003 survey (rather than the 1980s ratings used to position the emotions on the spiral), with EPA profiles of emotions averaged across sexes. INTERACTIVE EFFECTS

Table 7.3 shows that threats to moral character result in similar emotions regardless of the rank of the threatener. In the context of the consulting firm, such threats lead individuals to feel flustered or mad. In the context of the hospital, such threats generally lead individuals to feel annoyed or bitter. Threats to one’s reputation for competency create unpleasant emotions of various kinds, and these feelings are most pronounced when the threats come from

CH07-P374095.indd 116

5/27/2008 7:22:49 PM

117

Supporting and Threatening Face in Work Organizations

Repeated-Measures Analysis of Variance of Evaluation, Potency, and Activity Coordinates of Emotional Responses to Work World Vignettes, With Two Replications Per Cell. Statistics Based on Pillai’s Trace

TABLE 7.2

Hypothesis df

Error df

Significance

46.539

3

231

0.000

23.942 2.188 3163.590 97.841 121.995 22.095 3.482 1.269 16.878 3.823 49.074 6.205 2.278 0.115 60.837 17.459 13.135 0.488 1.962 1.187 35.193 3.743 2.840 1.595 2.713 0.181 9.436 0.238 4.647

3 3 3 3 3 3 6 6 3 3 3 3 6 6 3 3 6 6 6 6 3 3 6 6 6 6 6 6 6

231 231 231 231 231 231 228 228 231 231 231 231 228 228 231 231 228 228 228 228 231 231 228 228 228 228 228 228 228

0.000 0.090 0.000 0.000 0.000 0.000 0.003 0.273 0.000 0.011 0.000 0.000 0.037 0.995 0.000 0.000 0.000 0.817 0.072 0.314 0.000 0.012 0.011 0.150 0.015 0.982 0.000 0.964 0.000

1.277

6

228

0.269

F Between Subjects Sex Within Subjects Setting Setting * sex Valence Valence * sex Aspect of face Aspect of face * sex Rank of alter Rank of alter * sex Setting * valence Setting * valence * sex Setting * aspect of face Setting * aspect of face * sex Setting * rank of alter Setting * rank of alter * sex Valence * aspect of face Valence * aspect of face * sex Valence * rank of alter Valence * rank of alter * sex Aspect of face * rank of alter Aspect of face * rank of alter * sex Setting * valence * aspect of face Setting * valence * aspect of face * sex Setting * valence * rank of alter Setting * valence * rank of alter * sex Setting * aspect of face * rank of alter Setting * aspect of face * rank of alter * sex Valence * aspect of face * rank of alter Valence * aspect of face * rank of alter * sex Setting * valence * aspect of face * rank of alter Setting * valence * aspect of face * rank of alter * sex

a superordinate. In the consultancy, the closest emotions are feeling disgusted or nervous, with the order reversed in the case of peers or subordinates. In the hospital, the closest emotions are disgusted or embarrassed when the threat comes from superordinates, disgusted or annoyed when the threat comes from peers, and annoyed or disgusted when the threat comes from subordinates.

CH07-P374095.indd 117

5/27/2008 7:22:50 PM

CH07-P374095.indd 118

Hospital

Threatening

Consultancy

Supportive

Threatening

Supportive

Valence

Superordinate Peer Subordinate Superordinate Peer Subordinate

Job competency

Superordinate Peer Subordinate

Job competency

Moral character

Superordinate Peer Subordinate

Superordinate Peer Subordinate

Job competency

Moral character

Superordinate Peer Subordinate

Superordinate Peer Subordinate

Job competency

Moral character

Superordinate Peer Subordinate

Rank of alter

Moral character

Aspect of face

⫺1.89 ⫺1.83 ⫺1.72

⫺1.79 ⫺1.73 ⫺1.60

2.51 2.43 2.39

2.43 2.35 2.30

2.54 2.44 2.42

⫺1.94 ⫺1.92 ⫺1.97

⫺1.87 ⫺1.87 ⫺1.91

2.60 2.51 2.51

2.48 2.14 2.31

1.92 2.49 2.08

2.56 2.30 2.40

2.07 2.56 2.19

⫺1.54 ⫺1.35 ⫺1.41

2.59 2.52 2.49

2.66 2.59 2.59

⫺1.69 ⫺1.63 ⫺1.49

⫺1.81 ⫺1.81 ⫺1.84

2.63 2.46 2.49

2.23 2.63 2.31

⫺1.26 ⫺1.04 ⫺1.13

2.01 1.89 1.82

2.13 2.03 1.98

⫺0.99 ⫺0.76 ⫺0.68

⫺0.73 ⫺0.69 ⫺0.73

2.04 1.87 1.84

1.81 2.11 1.85

⫺0.74 ⫺0.55 ⫺0.60

1.95 1.82 1.74

2.08 1.96 1.90

⫺1.07 ⫺0.85 ⫺0.76

⫺0.81 ⫺0.76 ⫺0.81

1.97 1.75 1.76

1.71 2.04 1.77

⫺0.83 ⫺0.65 ⫺0.68

⫺0.44 ⫺0.40 ⫺0.38

⫺1.40 ⫺1.20 ⫺1.27

⫺0.35 ⫺0.31 ⫺0.30

⫺1.74 ⫺1.78 ⫺1.74

⫺1.65 ⫺1.69 ⫺1.66

⫺1.55 ⫺1.60 ⫺1.57

Lower bound

Mean

Lower bound

Mean

Upper bound

Potency 95% confidence interval

Evaluation 95% confidence interval

2.08 1.96 1.90

2.19 2.10 2.05

⫺0.91 ⫺0.68 ⫺0.60

⫺0.65 ⫺0.61 ⫺0.65

2.10 1.98 1.91

1.92 2.17 1.93

⫺0.65 ⫺0.45 ⫺0.52

⫺0.26 ⫺0.23 ⫺0.22

Upper bound

1.49 1.32 1.18

1.62 1.49 1.41

⫺0.84 ⫺0.49 ⫺0.50

⫺0.21 ⫺0.13 ⫺0.25

1.58 1.42 1.25

1.39 1.58 1.35

⫺0.71 ⫺0.59 ⫺0.54

0.39 0.47 0.44

Mean

Means and Confidence Intervals for Evaluation, Potency, and Activity of Emotional Responses to Work World Vignettes

Setting

TABLE 7.3

1.40 1.23 1.07

1.56 1.40 1.32

⫺0.96 ⫺0.63 ⫺0.63

⫺0.35 ⫺0.27 ⫺0.39

1.50 1.32 1.15

1.30 1.50 1.26

⫺0.82 ⫺0.70 ⫺0.65

0.25 0.32 0.30

Lower bound

Activity 95% confidence interval

1.58 1.41 1.29

1.68 1.57 1.50

⫺0.72 ⫺0.36 ⫺0.37

⫺0.06 0.01 ⫺0.11

1.66 1.53 1.35

1.48 1.66 1.45

⫺0.60 ⫺0.47 ⫺0.44

0.52 0.61 0.58

Upper bound

118 Social Structure and Emotion

5/27/2008 7:22:50 PM

Supporting and Threatening Face in Work Organizations

119

Support of moral character produces pleasant, secure, and activated emotions. In the consultancy, such feelings are most extreme when the support comes from a peer, with the closest emotions being proud or pleased; support from superordinates or subordinates produces feelings of being overjoyed or proud. In the case of the hospital, emotional pleasantness, security, and activation are most pronounced when support of moral character comes from a superordinate, and somewhat less activating when support comes from a subordinate. The closest emotions in this context are proud or pleased, with the order reversed in the case of a subordinate. Support of job competency similarly produces pleasant, secure, activated emotions, and this is especially true when the supporter is a superordinate, with the closest emotions being proud or pleased in this case, for both the consultancy and hospital. Otherwise in a consultancy the closest emotions are overjoyed– proud when support comes from a peer, and overjoyed–pleased in the case of a subordinate. In a hospital the closest emotions are overjoyed or pleased when the support comes from either a peer or subordinate. Events involving neutral face behaviors in a hospital are not in Table 7.3 because they are not part of our factorial design. However, the EPA means for emotional responses to the neutral behaviors are as follows: neutral behavior from a superordinate, 1.03, 0.67, ⫺0.33; from a peer, 1.47, 0.95, ⫺0.21; from a subordinate, 1.36, 0.85, ⫺0.38. These correspond approximately to the emotion “contented” (EPA profile 1.52, 0.91, 0.63) except the emotional responses to face-neutral activity are less activated than contentment. While sex is not a significant factor in the highest order interaction, it does become significant in the interaction of setting * valence * aspect of face * sex. Our examination of a table of means for this interaction (not presented here) reveals the following sex differences. Females respond to threat with more unpleasant emotions than males – especially in the consultancy, and to support with more pleasant emotions. Additionally, females respond to any kind of support in either setting with more dominant and activated emotions than males.

STUDY 2: SIMULATED JOB Our second study was presented to participants as an organizational communications project. Undergraduate students imagined themselves as team leaders processing a package of work memos. Experimental manipulations were operationalized through the background information in packets that the participants received and via the memos they had to process. PROCEDURE

Each session began by distributing folders randomly to the participants. The pre-experiment survey of demographics and other personal information was the first document in the folder.

CH07-P374095.indd 119

5/27/2008 7:22:50 PM

120

Social Structure and Emotion

After participants completed the survey, the researcher verbally described the study as oriented toward developing management training programs. The simulated work setting was described as an office within a large bookselling firm. Participants were asked to assume the identity of Terry Norman, “a good and honest team leader,” and they were told to imagine that it was 8 A.M., the start of a workday. The first task each morning was to read and respond to inter-office memos, thereby enabling team leaders to direct information to team members early in the day. Verbal and written instructions were given to demonstrate memo processing. Participants were given a sample memo and response form to complete during the instruction period. The second item in the folder was the memo packet, with a cover sheet that listed eleven points of background information. The first and fourth points included reference to the manager’s face – quarrelsome, middle-aged, or considerate. The sixth point was the first exposure to the manager’s behavior toward the participant – a behavior manipulation that was repeated in Memo 2. Participants recorded responses to eleven memos. Participants indicated an emotion response to each memo by selecting one of twelve emotions (furious, angry, annoyed, fearful, nervous, embarrassed, calm, contented, happy, proud, excited, or thrilled). Each memo response form also included a prompt like the following: “Manager Pat O’Neal Memo 2. Indicate how you feel about yourself when thinking about this memo from Pat.” Here the participant was presented with 35-position rating scales with the following bi-polar anchors; Powerful– Powerless; Inactive–Active; Bad–Good. These ratings henceforth are referred to as the Potency, Activity, and Evaluation self ratings. The form also included three communications questions to give realism to the study’s stated purpose. A post-experiment survey included manipulation checks. Design: The participants were randomly assigned to the 15 experimental conditions of a 3 ⫻ 5 factorial design. The experimental factors were embedded within the memo packets received by participants. Their imagined manager, Pat O’Neil, was presented to different participants in three different ways: as a good manager – “considerate, high status, and skills;” as a neutral manager – “middle-aged, average status, and skills;” and as a bad manager – “quarrelsome, low status, and skills.” According to program Interact (Heise 1995) a “considerate manager” has an Evaluation–Potency–Activity profile of 1.52, 1.77, 1.29, a “middle-aged manager” has EPA of 0.10, 1.49, 0.87, and a “quarrelsome manager” is 1.17, 1.21, 1.61. Thus these three types of manager are of descending worthiness in terms of evaluation and potency scores. The manager face condition manipulation – good, neutral, or bad – was presented on the cover page of respondents’ packets. For example, in the good manager condition the face page included the statement: “In recent months, your considerate manager, Pat, seen as having high status and skills, has had increased interactions with you.” Five face behavior manipulations by the manager completed the design: character threat to employee – an accusation of theft; status threat – a criticism of job performance; neutral behavior – observation of employee’s software installation;

CH07-P374095.indd 120

5/27/2008 7:22:50 PM

Supporting and Threatening Face in Work Organizations

121

status support – a compliment on job performance; character support – praise of integrity. Face behavior items were derived from the vignettes in Study 1. Manager face behavior to employee was established in the background information, and also in the second memo to be processed. For example, in the character-threat condition, the cover page included the statement “Yesterday your manager, Pat, came to your workspace and accused you of stealing science fiction book inventory based on inventory shortages and awareness of your science fiction video collection;” and memo 2 said, “To: Terry Norman, Team Leader, Order Processing and Shipping; From: Pat O’Neil, Manager, Order Processing and Shipping; In regard to your suspected theft of book inventory, an explanation of science fiction inventory shortages is needed. Meet me in my office tomorrow at 2 P.M. P.O.” Participants’ packages were identical across experimental cells, other than the two independent variable manipulations. Cell sizes across the 15 experimental conditions ranged from 24 to 35. Average elapsed time of participation was approximately 50 minutes. Participants: Undergraduate business and psychology students at Indiana University participated in the experiment in exchange for course credit. Of 452 participants, 90.5% provided adequate data, resulting in a final sample size of 409. Their demographics were: 58% men; 49% 20 years old or less and 49% between 21 and 25; 85% white Americans, 2.7% Asian-Americans, 6.8% foreign born, and the remaining 5.5% consisting of other American minorities. Most students were currently employed, 62% part-time, 33% full-time. Manipulation Checks: The post-experiment questionnaire included several items measuring participant impressions of the manager, with response options ranging from strongly disagree (1) to strongly agree (5): “My manager, Pat: was pleased with me; treated me fairly; treated me with respect”; “I like my manager, Pat”; and “I am satisfied with my manager Pat.” Cronbach’s alpha for a scale made by averaging the five items was 0.85. The mean scale responses for the behavior conditions were: character threat, 2.6; status threat, 2.9; neutral, 3.7; status support, 4.1; and character support 4.1. An ANOVA showed that these differences were significant: F(4,404) ⫽ 47.9, p ⬍ 0.001. Thus, the manipulations successfully conveyed face behavior to subjects. Participants also rated “my manager, Pat O’Neil” on EPA rating scales. ANOVA conducted across all participants indicated means on each dimension were significantly different between the three manager face conditions, good, neutral, and bad. The results for evaluation were F(2,395) ⫽ 34.1, p ⬍ 0.001; for potency F(2,395) ⫽10.7, p ⬍ 0.001, for activity F(2,395) ⫽ 25.5, p ⬍ 0.001. Thus, the manipulations successfully conveyed the personality and competency of the manager. RESULTS

Frequencies of emotions selected by subjects in the 15 experimental conditions are displayed at Table 7.4. Overwhelmingly, unpleasant emotions were associated with threat conditions and pleasant emotions with support conditions.

CH07-P374095.indd 121

5/27/2008 7:22:50 PM

122 TABLE 7.4

Social Structure and Emotion Empoyee Emotion Frequencies Associated With Manager’s Face Behavior in the

Experiment Manager face behavior Employee perception of manager

Character threat

Status threat

Neutral behavior

Character Support

Status Support

Considerate, high status manager

Angry 33% Nervous 29% Annoyed 17%

Embarrassed 46% Annoyed 27% Angry 15%

Calm 39% Content 26% Annoyed 13%

Proud 60% Happy 28%

Proud 50% Happy 27% Thrilled 12%

Middle-aged, average status manager

Angry 25% Annoyed 21% Embarrassed 12%

Annoyed 61% Angry 18% Embarrassed 18%

Calm 48% Content 30% Annoyed 18%

Proud 65% Happy 12%

Proud 43% Happy 26% Thrilled 20%

Quarrelsome, low status manager

Angry 32% Annoyed 28% Nervous 16%

Annoyed 60% Embarrassed 16% Furious 12%

Calm 30% Content 30% Annoyed 30%

Proud 46% Happy 21% Excited 10%

Happy 29% Proud 29% Excited 18%

Converting emotion responses to Evaluation–Potency–Activity profiles yielded data that were too rough for variance analyses, because each dimension contained only 12 measurement points. (Study 1 provided 73 measurement points – no emotion plus 24 emotions at three levels of intensity.) However, respondents recorded their feelings directly on EPA scales immediately after reading the memo that provided the experimental manipulation. The means of self-rated feelings are displayed in Table 7.5. A multivariate ANOVA of the part of the data fitting a factorial design is shown in Table 7.6. This analysis indicates that valence, aspect of face, and their interaction are the only significant effects in this study. We checked that this outcome was not just a result of analyzing self ratings by computing a similar ANOVA for emotion responses converted to Evaluation–Potency–Activity profiles (even though the homogeneity of variance assumption is seriously violated): valence, aspect of face, and their interaction were the only significant effects in that analysis also. Broken down by dimension, the significant effects are as follows. Valence: participants who received supportive face behavior felt significantly nicer, more potent, and more active than participants whose faces were threatened. Aspect of face: participants whose job competency was addressed felt significantly nicer and more potent than participants whose moral character was at issue. Interaction: the difference in being activated by support versus being deactivated by threat was significantly greater for participants whose competence was in play than for participants whose moral character was at issue.

DISCUSSION In both studies face-supportive actions caused the recipient to experience pleasant, secure, and relatively activated emotions, and face-threatening actions

CH07-P374095.indd 122

5/27/2008 7:22:50 PM

CH07-P374095.indd 123

Neutral

Moral character

Threatening

Job competency

Aspect of face

Bad manager

Neutral manager

Good manager

Bad manager

Neutral manager

Good manager

Bad manager

Neutral manager

Good Manager

Manager standing

Male Female Male Female Male Female

Male Female Male Female Male Female Male Female Male Female Male Female

Sex

0.99 1.05 1.07 0.92 1.48 0.9

⫺0.41 ⫺0.46 ⫺0.12 ⫺0.67 0.02 ⫺0.4

0.29 0.3 0.47 0.13 0.75 0.25

⫺0.35 ⫺0.44 ⫺0.59 ⫺0.68 ⫺1.19 ⫺0.84 ⫺0.49 ⫺0.94 ⫺0.63 0.12 0.18 ⫺0.03

⫺1.87 ⫺1.83 ⫺1.94 ⫺2.27 ⫺2.58 ⫺2.29 ⫺1.83 ⫺2.33 ⫺1.79 ⫺1.56 ⫺1.08 ⫺1.62

⫺1.11 ⫺1.13 ⫺1.27 ⫺1.48 ⫺1.88 ⫺1.56 ⫺1.16 ⫺1.63 ⫺1.21 ⫺0.72 ⫺0.45 ⫺0.83

Upper bound

Lower bound

Mean

Evaluation 95% confidence interval

0.25 0.32 0.22 0.14 0.33 0.42

⫺0.86 ⫺1.6 ⫺1.23 ⫺0.83 ⫺1.77 ⫺1.29 ⫺0.38 ⫺1.08 ⫺0.8 ⫺0.69 ⫺0.34 ⫺0.95

Mean

⫺0.46 ⫺0.45 ⫺0.38 ⫺0.71 ⫺0.4 ⫺0.24

⫺1.63 ⫺2.3 ⫺1.91 ⫺1.63 ⫺2.48 ⫺2.03 ⫺1.06 ⫺1.78 ⫺1.39 ⫺1.54 ⫺0.98 ⫺1.76

Lower bound

0.96 1.09 0.82 0.99 1.07 1.08

⫺0.09 ⫺0.89 ⫺0.55 ⫺0.02 ⫺1.06 ⫺0.56 0.31 ⫺0.37 ⫺0.22 0.16 0.29 ⫺0.14

Upper bound

Potency 95% confidence interval

0.65 0.93 0.88 0.97 0.96 0.7

0.2 ⫺0.06 0.32 ⫺0.42 ⫺0.4 ⫺0.31 ⫺1 ⫺0.9 ⫺0.99 ⫺0.28 ⫺0.42 ⫺1.2

Mean

⫺0.03 0.19 0.3 0.15 0.25 0.07

⫺0.54 ⫺0.74 ⫺0.33 ⫺1.2 ⫺1.08 ⫺1.02 ⫺1.66 ⫺1.58 ⫺1.55 ⫺1.1 ⫺1.04 ⫺1.98

Lower bound

(Continues)

1.33 1.67 1.45 1.79 1.67 1.33

0.94 0.62 0.98 0.35 0.28 0.4 ⫺0.34 ⫺0.22 ⫺0.42 0.54 0.19 ⫺0.42

Upper bound

Activity 95% confidence interval

Means and Confidence Intervals for Evaluation, Potency, and Activity of Self-Ratings After Face-Behavior Manipulation

Valence

TABLE 7.5

Supporting and Threatening Face in Work Organizations

123

5/27/2008 7:22:50 PM

CH07-P374095.indd 124

Moral character

Supportive

Job competency

Aspect of face

(Continued )

Valence

TABLE 7.5

Bad manager

Neutral manager

Good manager

Bad manager

Neutral manager

Good Manager

Manager standing Male Female Male Female Male Female Male Female Male Female Male Female

Sex 2.17 2 2.36 1.27 1.72 2.18 2.11 2.92 2.81 2.21 1.84 2.5

Mean 1.62 0.87 1.8 0.54 1.11 1.42 1.43 2.22 2.26 1.51 1.17 1.83

Lower bound 2.72 3.13 2.93 2 2.33 2.94 2.78 3.62 3.36 2.91 2.51 3.17

Upper bound

Evaluation 95% confidence interval

1.75 1.4 1.54 1.25 1.21 1.84 1.41 1.94 1.76 1.48 1.63 1.86

Mean 1.19 0.26 0.97 0.51 0.59 1.07 0.73 1.23 1.21 0.77 0.94 1.18

Lower bound 2.31 2.54 2.11 1.99 1.82 2.61 2.09 2.65 2.32 2.19 2.31 2.54

Upper bound

Potency 95% confidence interval

2.01 1.95 2.04 1.48 1.44 2.11 1.89 2.42 2.07 1.63 2.02 1.95

Mean

1.48 0.85 1.49 0.77 0.85 1.37 1.24 1.74 1.54 0.95 1.36 1.29

Lower bound

Activity 95% confidence interval

2.55 3.05 2.59 2.19 2.04 2.85 2.55 3.1 2.61 2.31 2.67 2.6

Upper bound

124 Social Structure and Emotion

5/27/2008 7:22:51 PM

125

Supporting and Threatening Face in Work Organizations

TABLE 7.6

Multivariate Analysis of Variance of Self-Rating EPA. Statistics Based on

Pillai’s Trace Effect

F

Hypothesis df

Error df

Significance

Valence Aspect of face Manager Standing Sex Valence * aspect of face Valence * manager standing Aspect of face * manager standing Valence * aspect of face * manager standing Valence * sex Aspect of face * sex Valence * aspect of face * sex Manager standing * sex Valence * manager standing * sex Aspect of face * manager standing * sex valence * aspect of face * manager standing * sex

159.11 11.34 0.18 0.08 8.98 0.57 0.68 1.32 0.29 0.87 1.6 1.61 1.16 1.46 0.45

3 3 6 3 3 6 6 6 3 3 3 6 6 6 6

303 303 608 303 303 608 608 608 303 303 303 608 608 608 608

0.000 0.000 0.983 0.970 0.000 0.751 0.665 0.244 0.831 0.457 0.189 0.143 0.329 0.188 0.847

caused unpleasant, vulnerable, and relatively deactivated emotions. These are the largest effects in our studies. ASPECTS OF FACE

In general, threats to moral-face produced more unpleasant emotions than threats to status-face (i.e., the aspect of face based partly on competency). In the two business firms we examined (consultancy and book seller), support of moral-face also produced less pleasant emotions than support of status-face. However, in the hospital, support of moral-face produced more pleasant emotions than support of status-face. In the consultancy and hospital, status threat produced more vulnerable feelings than moral threat. Status support also produced less secure feelings than moral support in the hospital, but there was no consistent difference in the consultancy. In the book firm, status threat produced less vulnerable feelings than moral threat, and status support generally produced more secure feelings than moral support. Status threat produced more deactivation than moral threat in all three settings. Status support was less activating than moral support in the consultancy and hospital, but status support was more activating than moral support in the book firm. These complicated outcomes indicate that differences exist in how the two aspects of face affect emotions, but the two aspects interact in complex ways with the setting of the interaction (and with the identity of alter, as discussed below).

CH07-P374095.indd 125

5/27/2008 7:22:51 PM

126

Social Structure and Emotion

ALTER’S IDENTITY

In Study 1 we treated the actor who instigated emotion as a superordinate, a peer, or a subordinate. In Study 2 we varied the actor’s standing from being a positive organizational contributor, to being average, to being a negative in the organization. The different aspects of actor identity in the two studies require considering the studies separately. Differences in the actor’s rank had no impact on the emotions experienced as a result of threats to moral-face. However, in the case of threats to status-face, emotions were more unpleasant, vulnerable, and inactive when the threat came from a superordinate. Additionally, in a hospital, emotions were less unpleasant and vulnerable when a status threat came from a subordinate. The effect of rank on emotions produced by moral support varied by setting. In the consultancy, the most pleasant, secure, and active emotions resulted when moral support came from peers. In the hospital, moral support by a superordinate produced the most positive feelings, and moral support from a subordinate produced attenuated positive feelings. With regard to support of status-face, the most positive feelings always came from superordinates rather than peers or subordinates. In the consultancy, support from a subordinate produced moderated activation. Generally, then, threats or support by a superordinate produced the most extreme emotions, and threats or support by a subordinate produce somewhat attenuated feelings. In Study 2, variations in actor valuation had no impact on emotions experienced. That is, emotions depended on alter’s action but not on whether alter was pleasant and valuable in the organization, as opposed to being unpleasant and deadwood. SETTING

Study 1 had participants imagine themselves in a consultancy or a hospital as they responded to hypothetical events. The simulated work experience in Study 2 emplaced participants in a bookselling firm. These different settings contributed to emotional responses. In the two business settings, support of status yielded more pleasant emotions than support of moral character, whereas in the hospital supporting moral character generated more pleasant feelings than supporting status. Additionally, in the consultancy, peer support of moral-face was most gratifying, whereas in the hospital moral support was most gratifying when it came from superordinates and less gratifying when it came from peers or subordinates. In the Study 1 settings, the least vulnerable, most secure emotions emerged from threat or support of moral-face rather than of status-face. However, in Study 2, emotional responses were more potent for issues of status-face rather than for issues of moral-face. In Study 1 settings, support of moral-face was more activating than support of status-face. However, in Study 2 more activation arose from support of status-face than from support of moral-face. Thus in some respects the hospital was a different emotional environment than the businesses. This might

CH07-P374095.indd 126

5/27/2008 7:22:51 PM

Supporting and Threatening Face in Work Organizations

127

be because hospitals are relatively unfamiliar places for undergraduate students, or perhaps the life-and-death mission of hospitals impinges on emotionality.1 Also, the bookselling firm in Study 2 was a different emotional environment than the organizations in Study 1 in generating more secure and activated emotions from status issues as opposed to moral issues. This may be indicative of something about the kind of firm involved in Study 2, or it may relate to the different kind of participation involved in Study 2 as compared to Study 1 – simulated work, as opposed to imagining scenes. NEUTRAL BEHAVIOR

We included mundane, neutral behaviors in both studies, even though such behaviors could not be included in the factorial analyses because they do not manifestly relate to threatening or supporting face. In Study 1, the neutral behaviors generated emotions that were slightly pleasant and potent and a bit deactivated, which amounts to “contentment” with low intensity. In Study 2, the neutral behaviors created feelings that were a bit pleasant and potent and slightly activated. Low intensity contentment could arise from this also, but such feelings alternatively might transform into self-consciousness or anxiety – perhaps reflecting the enhanced realism of simulated work as opposed to imagined scenes. In any case, mundane everyday work activities were associated with low intensity feelings, generally pleasant, and not with a lack of emotion altogether. SEX

Sex of the participant was not a significant factor in Study 2, where we analyzed direct ratings of feelings. Respondent sex did have some significant effects in Study 1 – essentially, females responded with more intense feeling than males, especially to support. The significant sex differences in Study 1 could be an artifact of our measurement procedures in Study 1. We translated qualitative emotions into quantitative values using average ratings of each emotion by females or by males. On the whole the affective meanings of emotions are more extreme for females than for males, especially for pleasant emotions. Thus the sex differences found in Study 1 could relate to affective meanings of emotions rather than to differences in emotional response in work situations.

CONCLUSION Heise and Calhan (1995) demonstrated emotion norms for 128 dyadic events. An example item was: “Imagine a critic ridicules you. How do you feel 1

Our thanks to Kathryn Lively for this suggestion.

CH07-P374095.indd 127

5/27/2008 7:22:51 PM

128

Social Structure and Emotion

at that moment?” Heise and Calhan proposed that their research results implied that emotionality is normatively regulated in virtually all contemporary social encounters, and in particular “many commercial and professional transactions are expected to be emotionally charged in orderly ways” (1995, p. 239). Our work extends Heise and Calhan’s claim regarding commercial and professional interactions, by considering emotional responses to highly realistic work-world scenarios, by examining emotions of individuals while they worked in a simulated job, and by employing research subjects who were working to support themselves, a third of them in full-time jobs. Our results confirm the Heise and Calhan claim in that all of the scenarios we examined – including those designed to be neutral and unemotional – gave rise to emotions of greater or lesser intensity. Thus work worlds are as much domains of ordered affectivity as they are domains of rational choice and logical procedures. We examined the issue of face in detail, and our results demonstrate that threats to face generate negative emotions, while interactions that support an individual’s face cause the individual to experience positive emotions. Threats to moral-face typically produce the most intense negative emotions, and support of status-face often produces the most intense positive feelings. However, the emotional consequences of bringing an individual’s face into play also are determined by the rank of the interaction partner and by the organizational setting of the encounter. Complications aside, a face-supportive work environment promotes pride and happiness among workers, and a face-threatening work environment generates aggravation and humiliation. We do not have evidence about how these emotional outcomes contribute to productivity. However, Organizational Citizenship Behavior (Organ 1988; 1997) has been shown to account for about 25% of variance in organizational financial indicators and 38% of variance in customer service (Podsakoff, Ahearne and MacKenzie 1997; Walz and Niehoff 1996). Helping in order to solidify and preserve organizational relationships is part of Organizational Citizenship Behavior, and pleasant affective states are significantly related to higher helping levels (George 1989; Isen and Levin 1972). Oddly, so are unpleasant emotions (Lee and Allen 2002). Study 2 was partially designed to resolve this seeming anomaly, and while we do not present the results in detail here, the essential finding was that unpleasant emotions increase helping only when they are generated by a valued manager. The bottom line, then, is that optimizing organizational operations requires a face-supportive environment that produces positive affect and generous helping, or else a cadre of managers with sufficient charisma to overcome the deleterious effects of threatening a subordinate’s face.

REFERENCES Bond, Michael and Peter Wing Ho Lee. 1981. “Face-Saving in Chinese Culture: A Discussion and Experiment of Hong Kong Students.” In: Ambrose Y. C. King and Rance P. L. Lee (eds), Social Life and Development in Hong Kong, pp. 285–305. Hong Kong: Chinese University Press.

CH07-P374095.indd 128

5/27/2008 7:22:51 PM

Supporting and Threatening Face in Work Organizations

129

Botterweck, G. Johannes, Helmer Ringgren, and Heinz-Joseph Fabry. 1974. Theological Dictionary of the Old Testament. David E. Green (Trans). Grand Rapids, MI: William B. Eerdmans Publishing Company. Brown, Penelope and Steven Levinson. 1978. “Universals in Language Usage: Politeness Phenomena.” In: E. Goody (ed.), Questions and Politeness. London: Cambridge University Press. Earley, P. Christopher. 1997. Face, Harmony, and Social Structure: An Analysis of Organizational Behavior Across Cultures. New York: Oxford University Press. Friedrich, Gerhard (ed.). 1967. Theological Dictionary of the New Testament. Geoffrey W. Bromiley (Trans. and ed.). Grand Rapids, MI: Wm. B. Eerdmans. George, Jennifer M. 1989. “Mood and Absence.” Journal of Applied Psychology 74(2):317–324. Goffman, Erving. 1967. Interaction Ritual: Essays on Face-to-Face Behavior. New York: Anchor Books. Heise, David R. 1995. Interact 2: A Computer Program for Studying Cultural Meanings and Social Interaction. Bloomington, IN: Indiana University. Heise, David R. 1997. “Interact On-Line (Java applet)”, Retrieved January 1, 2004 (www.indiana. edu/~socpsy/ACT/interact/JavaInteract.html). Heise, David R. 2001. “Project Magellan: Collecting Cross-Cultural Affective Meanings Via the Internet.” Electronic Journal of Sociology 5(3). Heise, David R. and C. Calhan. 1995. “Emotion Norms in Interpersonal Events.” Social Psychology Quarterly 58:223–240. Hu, Hsien C. 1944. “The Chinese Concept of Face.” American Anthropologist 46:45–64. Isen, Alice M. and Paul F. Levin. 1972. “The Effect of Feeling Good on Helping: Cookies and Kindness.” Journal of Personality and Social Psychology 21:384–388. Organ, Dennis W. 1988. Organizational Citizenship Behavior: The Good Soldier Syndrome. Lexington, MA: Lexington Books. Organ, Dennis W. 1997. “Organizational Citizenship Behavior: It’s Construct Clean-up Time.” Human Performance 10(2):85–97. Podsakoff, Phillip, Michael Ahearne, and Scott B. MacKenzie. 1997. “Organizational Citizenship Behavior and the Quantity and Quality of Work Group Performance.” Journal of Applied Psychology 82:262–270. Redding, S. Gordan and Michael Ng. 1982. “The Role of “Face” in the Organizational Perceptions of Chinese Managers.” Organization Studies 3(3):201–219. Stover, Leon E. 1974. The Cultural Ecology of Chinese Civilization. New York: Mentor. Ting-Toomey, Stella, Ge Gao, Paula Trubisky, Zhizong Yang, Hak Soo Kim, Sung-ling Lin, and Tsukasa Nishida. 1991. “Culture, Face Maintenance, and Styles of Handling Interpersonal Conflict: A Study in Five Cultures.” International Journal of Conflict Management 2(4):275–296. Tjosvold, Dean and Ted Huston. 1978. “Social Face and Resistance to Compromise in Bargaining.” Journal of Social Psychology 104:57–68. Walz, Sandra M. and Brian P. Niehoff. 1996. “Organizational Citizenship Behaviors and Their Effect on Organizational Effectiveness in Limited-Menu Restaurants.” Academy of Management Best Papers Proceedings 23:307–311.

CH07-P374095.indd 129

5/27/2008 7:22:51 PM

PA R T

III Emotions in Exchange Relations

CH08-P374095.indd 131

5/26/2008 2:45:20 PM

8 Commentary James W. Balkwell University of Georgia, Athens, GA, USA

Sociologists generally agree that social structures provide the frameworks within which human interaction takes place. Moreover, the give-and-take of human interaction may (and typically will) help to preserve those social structures, although in less common cases it may serve to weaken or alter them. The early versions of what came to be called social exchange theory (Homans 1958; Thibaut and Kelley 1959; Emerson 1962; Adams 1963; Blau 1964) focused mainly on the exchange process itself, understood as a process of human interaction, emphasizing the hedonistic motives of those engaged in social exchanges, positing that people seek to maximize their profits (rewards minus costs) from their interpersonal relationships, including those in the workplace and those in more personal settings such as dating. While people’s emotional responses played a part in some of those early analyses (e.g., Homans 1961:72–78), the predominant part was given to rational choice, social learning, or a combination of these. Essentially, these analyses posited that individuals judge the attractiveness of their relationships, and then they make choices, based on their judgments. While social structures always were implicitly involved in this theorizing, the main emphasis was upon the interacting individuals, not upon a larger social structure. Setting the stage for an explicit consideration of the larger social structure was Emerson’s (1962) analysis of power and dependency in social exchange, the power of one actor resting on the dependency of the other, which led him to conclude that power is a property of a relation, not of an actor. Such relations usually are embedded in larger structures of relations, as Thibaut and Kelley’s (1959:21–23) early analysis of potential “alternative exchange relations” alluded. As time went on, these larger social structures became the primary focus. A major part of what modern sociologists mean by “social structure” consists of the networks of relations found in groups, organizations, and communities. Social Structure and Emotion

CH08-P374095.indd 133

133

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 2:45:20 PM

134

Social Structure and Emotion

Such networks became explicitly incorporated into social exchange theory in the 1970s and 1980s (Cook and Emerson 1978; Cook et al. 1983; Markovsky, Willer, and Patton 1988). These extensions inevitably led to the identification of further theoretical issues, creating the kinds of lively debates that often go together with progress in any field of scholarship (Molm and Cook 1995; Willer 1986; Willer and Skvoretz 1991). A relatively recent summary of the major controversies involving rival conceptions of network exchange processes can be found in Willer (1999). Even so, there was little controversy about the nature of the entailed human action: those involved in these debates and the research to help settle them continued, for the most part, to understand social exchange as resting principally on rational decision-making, social learning, or a combination of these. This assumption may have been satisfactory early on, but persons who participate regularly in real-world exchange networks (e.g., the ones in their workplaces) know very well that an ongoing exchange relation may have affective consequences, including increased or decreased levels of commitment to the relationship, which in turn may bear upon the continued functioning of that relationship (Lawler and Yoon 1993, 1996, 1998). Emotions (i.e., short-lived reactions to immediate events) surely play a part in creating those affective consequences. Thus, it probably was just a matter of time before theorists and researchers would be trying to incorporate emotional processes into network exchange theory. The conference upon which this book is based provided a wonderful opportunity for those engaged in these recent efforts to share their ideas on social structure and emotion, focusing on the “network exchange” facets of social structure and drawing upon the most pertinent social psychological work on emotional processes. In the remainder of this introduction, I will sketch key parts of the “network exchange” facets of social structure, including some of the most important variables; then I will sketch some equally basic ideas about social emotions and emotional processes; and finally I will describe an orienting perspective for bringing these elements together.

ORGANIZATION OF SOCIAL EXCHANGE While a social exchange relation constitutes a linkage in the larger social structure in which it is embedded, social exchange itself has some structural features, which we may call the organization of the exchange process (reserving the word “structure” for the encompassing networks, so as to avoid too much confusion). The entwined consequences of those features also can be considered as part of that organization, although for some purposes we might want to distinguish between the features themselves and their consequences. Max Weber, Émile Durkheim, and other giants of early sociology theorized near the beginning of the last century that to function fairly effectively and to endure over any length of time, a society must be constituted or structured so

CH08-P374095.indd 134

5/26/2008 2:45:20 PM

135

Commentary

as to foster a sense of trust, reasonable certainty in anticipating others actions, a sense of fairness, and related outcomes. Specifying these broad ideas more precisely, particularly as they apply to social exchange, the type of social exchange and the partners’ behaviors constitute key independent variables, while the factors upon which social order rests (such as trust, an absence of too much uncertainty, affective regard, a sense of fairness, and kindred factors) constitute key dependent variables. Essentially, the social structure is posited to foster (or in some instances fail to foster) a well-functioning social setting. The question is, “How does this work?” Molm’s formulation of social exchange theory is a notable attempt to answer this question (Molm, Peterson, and Takahashi 1999; Molm 2003). In her theory of reciprocity, the ultimate independent variables are the structure of reciprocity, or features of it, which differ between (1) social exchange that is direct and negotiated, (2) that which is direct and reciprocal in a more probabilistic and time-lagged sense, with responses coming later if at all, and (3) that which involves giving to and getting from multiple others, the social exchanges being much less direct in character. Regarding the latter, much of the “giving” and “receiving” that constitutes normal social life can be understood as social exchange, but the exchange partner to which a certain actor “gives” and the one from which he or she “receives” may be different persons. Still, these “generalized social exchanges” may be equitable or otherwise, depending on the rewards and costs to the individuals involved. In this theory, the ultimate dependent variables are integrative bonds – a fabric of social ties – understood as resting on trust, affective regard, feelings of solidarity, and feelings of fairness. It is clear that Molm’s dependent variables have affective components, representing widely shared evaluations, but Molm is careful to distinguish between emotion, on the one hand, and affect, on the other, the first referring to an intense but relatively short-lived feelings (e.g., anger over what someone just said or did), the second referring to a relatively enduring evaluative or sentimental disposition (e.g., a lasting fondness for the Thanksgiving holidays). This is an important distinction (compare Stets 2003). A recurring, recognizable kind of social situation is likely be a well-functioning social setting if it is infused with positive affect, but this fairly well-established conclusion from network exchange research tells us little about the part played by emotions as the latter are understood by most social psychologists who study emotional processes. Let us now turn to emotions as short-lived reactions to immediate events, which include such states as anger, joy, sadness, and fear.

SOCIAL EMOTIONS Some emotional states clearly are rooted in physiological activity within the autonomic nervous system, that is, within the limbic system, the adrenal glands, and so forth. When we are feeling joy or anger, for instance, our hearts are pounding, our blood vessels are constricted in some pattern, and our hands

CH08-P374095.indd 135

5/26/2008 2:45:20 PM

136

Social Structure and Emotion

are perspiring more than usual. According to Paul Ekman, “The most common way in which [such] emotions occur is when we sense, rightly or wrongly, that something that seriously affects our welfare, for better or worse, is happening or about to happen” (Ekman, 2003:19). As an emotion begins, it affects a person within a few milliseconds. Almost instantaneously, her or his facial expression and behavior may change. Some writers take pains to distinguish between the “core emotions,” firmly rooted in DNA and genetic inheritance, and the more culturally variable emotions, where cultural learning and cognitive information processing impinge more heavily upon the experiencing of those emotions. There is less-than-perfect agreement on what those core emotions are (compare Ekman and Friesen 1975, with Shaver et al. 1987), but there is wide agreement that cognitive information processing often plays a part in people’s emotional experience. Cultural norms about appropriate emotional expression, called display rules (Ekman 1972), clearly require some cognitive information processing if they are to be applied in a social setting. Likewise, emotion work (Hochschild 1983), which refers to a conscious effort to make feelings correspond to the culturally approved requirements of the occasion, entails an interplay between physiological activity and cognition. People may suppress, exaggerate, or alter their feelings to meet the requirements of the occasion, as when they amplify their joy at a wedding or their sadness at a funeral.

SOCIAL STRUCTURE AND EMOTIONAL EXPERIENCE Consider the diagram in Figure 8.1. On the left of the diagram, we have variables stemming from social interaction in an exchange network, such as the structure of reciprocity and each actor’s behavior. On the right, we have variables pertaining to integrative bonds, such as the level of trust, the sense of fairness, and so forth. In the middle, we have the products of emotional arousal and cognitive information processing. Any kind of human–social interaction, including that in exchange networks, is likely to produce at least some cognitive information processing, as people seek to

Social interaction in an exchange network

FIGURE 8.1

CH08-P374095.indd 136

Cognitive information processing Integrative bonds General emotional arousal Social structure, emotional experience, and integrative bonds.

5/26/2008 2:45:20 PM

Commentary

137

make sense of their situation. Also, it is likely to produce at least some level of emotional arousal, as people are likely to sense that what is happening affects their welfare, for better or worse. Stanley Schachter’s (1964) classic article on the interaction of the cognitive and the physiological determinants of emotional states develops this assertion about the cognitive and emotional interplay quite convincingly. Besides happenings in the social setting, emotional arousal itself also triggers some cognitive information processing, as the person seeks to identify, interpret, and label her or his state of arousal, using cues from the precipitating situation. This cognitive activity, in turn, feeds back upon the emotional arousal, giving it direction as well as its motivational significance. The person who is fearful may run; but the person who is angry may attack. The label matters. Earlier generations of psychologists used the expression, “fight or flight,” to signify that even among lower animals emotional arousal requires cues from the environment before it is translated into action. More generally, emotional arousal and cognitions together create the emotional variables of interest. The interaction between emotional arousal and cognitive processes does not stop there, however, and further aspects of that interaction may also be important in social exchange. Besides the labeling aspects, the level of emotional arousal influences the efficacy of that cognitive information processing more generally, including that which is directly pertinent to the social exchange in question. Translated into a modern terminology, the so-called Yerkes–Dodson Law (Yerkes and Dodson 1908) says that the relationship between emotional arousal and performance is curvilinear. Too little arousal inhibits performance; too much arousal also inhibits performance. Somewhere in between is the optimal level. Intuitively, the person who is not sufficiently aroused is likely to be mentally dull and error-prone, while the person who is excessively pumped up also is likely to be error-prone, her or his mental precision having been compromised. This law (or principle) further states that the optimal level depends upon the complexity of the thinking that the person in question is called upon to do. With greater problemcomplexity, a lower level of emotional arousal becomes the optimal one. Now if cognitive efficacy becomes seriously suboptimal, what then? How could that affect social exchange processes? One plausible conjecture is that a person’s quest to organize and understand her or his situation would become more infused with cognitive biases and errors in attribution. There is much evidence that people rely on their cognitive schemas (essentially their personal theories about things) to conserve their limited supplies of time and energy (Fiske and Taylor 1991), but their use of such schemas as shortcuts carries a risk of erroneous inference. Rather than engaging in accurate, effective, rational thought, people become more inclined to fall back on mental shortcuts. People have this tendency even under the best of circumstances, but this is especially likely to happen when their efficacy is suboptimal – that is, when it would require the most time and energy to do a more careful rational analysis. One theoretical task, then, if this suggestion proves valid, would be to identify the severity and direction of such biases in cognitive information processing and attribution.

CH08-P374095.indd 137

5/26/2008 2:45:20 PM

138

Social Structure and Emotion

Most of the classic work on emotions focuses on the nature and sources of emotions. In contrast, that on network exchange and emotion treats emotional experience both as an independent and as a dependent variable, operating in an encompassing system of action. This conception poses some challenges. Whether the perspective portrayed in Figure 8.1 is a fully adequate paradigm for the investigations to come remains to be seen, but it suggests where emotional and cognitive processes best fit into the network exchange framework, and it also suggests the likely importance of the interplay of emotional and cognitive processes. All the parts of this (or some similar) perspective will require further explication as theoretical and empirical investigations continue.

CONCLUSION Network exchange theorists and researchers assume the possibility of cumulative theory development, in which a formulation increases in scope, rigor, precision, and empirical adequacy as work on the theory continues (Wagner and Berger 1985). Given this faith in theory growth, it seems a foregone conclusion that theorizing and research on network exchange processes and emotion will accelerate in the coming years, inspired in significant part by the papers that follow. Cognitive and emotional processes will be incorporated into the general NET framework. Perhaps the best strategy for insuring a cumulative development of our knowledge about network exchange phenomena, including their cognitive and emotional components, is to develop falsifiable hypotheses about limited parts of the general model, test those limited hypotheses via careful research, and build incrementally upon the results of successive tests. Trying to do too much all at once, tempting as that might be, probably would not build the kind of solid foundation necessary for cumulative knowledge. Looking ahead, we eventually will need better measurements of emotion, because otherwise our statistical controls involving emotion are not going to be as convincing as we would like them to be. Even without major methodological improvements, however, if an area of theorizing and research has a sense of where it is headed, individual research studies that take us part way have greater meaning than do research not guided by a meaningful framework. The chapters that follow, written by some of the most productive contributors to network exchange theory, provide sound examples for others who will be participating in what promises to be an exciting journey toward an increasingly adequate understanding of social structure and emotion.

REFERENCES Adams, J. Stacey. 1963. “Toward an Understanding of Inequity.” Journal of Abnormal and Social Psychology 67:422–436. Blau, Peter M. 1964. Exchange and Power in Social Life. New York: Wiley.

CH08-P374095.indd 138

5/26/2008 2:45:20 PM

Commentary

139

Cook, Karen S. and Richard M. Emerson. 1978. “Power, Equity, and Commitment in Exchange Networks.” American Sociological Review 43:721–739. Cook, Karen S., Richard M. Emerson, Mary R. Gillmore, and Toshio Yamagishi. 1983. “The Distribution of Power in Exchange Networks: Theory and Experimental Results.” American Journal of Sociology 89:275–305. Emerson, Richard M. 1962. “Power-Dependence Relations.” American Sociological Review 27:31–40. Ekman, Paul. 1972. “Universals and Cultural Differences in Facial Expressions of Emotion.” In: J. K. Cole (ed.), Nebraska Symposium on Motivation, 1971. Lincoln Nebraska: University Press. Ekman, Paul. 2003. Emotions Revealed. New York: Henry Holt and Company. Ekman, Paul and Wally V. Friesen. 1975. Unmasking the Face: A Guide to Recognizing Emotions from Facial Clues. Upper Saddle River, NJ: Prentice-Hall. Fiske, Susan T. and Shelley E. Taylor. 1991. Social Cognition, 2nd edition. New York: McGraw-Hill. Hochschild, Arlie R. 1983. The Managed Heart: Commercialization of Human Feeling. Berkeley: University of California Press. Homans, George C. 1958. “Social Behavior as Exchange.” American Journal of Sociology 62:597–606. Homans, George C. 1961. Social Behavior: Its Elementary Forms. New York: Harcourt, Brace & World. Lawler, Edward J. and Jeongkoo Yoon. 1993. “Power and the Emergence of Commitment Behavior in Negotiated Exchange.” American Sociological Review 58:465–481. Lawler, Edward J. and Jeongkoo Yoon. 1996. “Commitment in Exchange Relations: Test of a Theory of Relational Cohesion.” American Sociological Review 61:89–108. Lawler, Edward J. and Jeongkoo Yoon. 1998. “Network Structure and Emotion in Exchange Relations.” American Sociological Review 63:871–894. Markovsky, Barry, David Willer, and Travis Patton. 1988. “Power Relations in Exchange Networks.” American Sociological Review 53:220–236. Molm, Linda D. 2003. “Theoretical Comparisons of Forms of Exchange.” Sociological Theory 21:1–17. Molm, Linda D. and Karen S. Cook. 1995. “Social Exchange and Exchange Networks.” In: Karen S. Cook, Gary Alan Fine, and James S. House (eds), Sociological Perspectives on Social Psychology, pp. 209–235. Boston: Allyn and Bacon. Molm, Linda D., Gretchen Peterson, and Nobuyuki Takahashi. 1999. “Power in Negotiated and Reciprocal Exchange.” American Sociological Review 64:876–890. Schachter, Stanley. 1964. “The Interaction of Cognitive and Physiological Determinants of Emotional State.” In: L. Berkowitz (ed.), Advances in Experimental Social Psychology, Vol. 1. New York: Academic Press. Shaver, Phillip R., Judith C. Schwartz, Donald Kirson, and Cary O’Connor. 1987. “Emotion Knowledge: Further Exploration of a Prototype Approach.” Journal of Personality and Social Psychology 52:1061–1086. Stets, Jan E. 2003. “Emotions and Sentiments.” In: J. DeLamater (ed.), Handbook of Social Psychology. New York: Kluwer. Thibaut, John W. and Harold H. Kelley. 1959. The Social Psychology of Groups. New York: Wiley. Wagner, David G. and Joseph Berger. 1985. “Do Sociological Theories Grow?” American Journal of Sociology 90:697–728. Willer, David. 1986. “Vulnerability and the Location of Power Positions.” American Journal of Sociology 92:441–444. Willer, David (ed.). 1999. Network Exchange Theory. New York: Praeger. Willer, David and John V. Skvoretz. 1991. “Power in Exchange Networks: Setting and Structural Variations.” Social Psychology Quarterly 54:224–238. Yerkes, Robert M. and John D. Dodson. 1908. “The Relation of Strength of Stimulus to Rapidity of Habit Formation.” Journal of Comparative Neurology and Psychology 18:459–482.

CH08-P374095.indd 139

5/26/2008 2:45:20 PM

9 The Effect of Perceived Trustworthiness on Affect in Negotiated and Reciprocal Exchange Alexandra Gerbasi and Karen S. Cook Stanford University, Stanford, CA, USA

ABSTRACT This chapter investigates the link between perceived trustworthiness and affect toward one’s exchange partner(s) in negotiated and reciprocal exchange, types of exchange that vary in the nature of the uncertainty involved in completing an exchange. We argue that the more cooperative an exchange partner is in either type of exchange, the higher the perceived trustworthiness of that exchange partner and the more positive the affect toward the partner. In addition, we argue that cooperative behavior has greater impact on perceived trustworthiness when uncertainty is high since such behavior more clearly signals trustworthiness. One hundred and fifty subjects participated in five experimental conditions in which they interacted with simulated partners who were either more or less cooperative. We varied the level of (simulated) partner cooperation and type of exchange. Our results indicate a clear effect of uncertainty on perceived trustworthiness, which is related to positive affect, especially when the available partners are highly cooperative. Low rates of Social Structure and Emotion

CH09-P374095.indd 141

141

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 2:48:16 PM

142

Social Structure and Emotion

cooperation have the expected negative effect on perceived trustworthiness and affect. Key Words: Perceived Trustworthiness, Affect, Negotiated Exchange, Reciprocal Exchange

INTRODUCTION In the past several decades social exchange theorists in sociology have primarily studied the structural determinants of exchange, in part due to the rapid growth of social network methods and approaches to social relations throughout the 1970s and 1980s. Richard Emerson (1972) was the first social exchange theorist to build a bridge between exchange theory and network conceptions of social relations and the structures created by their linkages. In this way he created a direct connection between micro-level exchange processes and macro social structures. While Blau (1964) also tried to bridge the micro- and macrolevels of social exchange, he did so with the notion of nested groups and relations of association (sometimes fairly complex) rather than with the conception of social networks linking actors through exchange. More recently, exchange theorists have turned to the analysis of the nonstructural aspects of exchange focusing on processes of exchange and attendant emotional responses. This focus is reflective of the concerns of some of the earliest exchange theorists and, especially, the anthropologists who wrote about the ways in which informal exchange created relations of social solidarity (LeviStrauss 1969; Mauss 1950 [1990]). Even Homans, the first sociologist to write about social behavior as exchange, included emotions in his theoretical treatise (Homans 1961 [1974]). He discussed the anger or guilt that might accompany an exchange between two parties that was felt to be unfair either by a party who received less than she viewed as “fair” or more than she viewed as “fair.” But, few other emotional responses were discussed in the early work on exchange, except the emotions and related sentiments that attended justice evaluations by either party to the exchange (or by third parties in indirect exchange). In an effort to extend recent work on affect and exchange, we investigate the effect of cooperation and perceived trustworthiness on affect towards one’s exchange partner in two different types of exchange, negotiated and reciprocal exchange. Before commenting on the recent work on emotions and exchange, however, let us briefly define some key terms including affect and emotion, which we distinguish from sentiment (a more enduring feeling toward another). We adopt the definition of emotion provided by Thoits (1990): An emotion includes four interconnected components: (1) situational cues, (2) physiological changes, (3) expressive gestures, and (4) an emotion label that names the specific configuration of components (see Stets 2003: 309–310). A further distinction is sometimes made in the literature on emotions between primary and secondary

CH09-P374095.indd 142

5/26/2008 2:48:16 PM

Effect of Perceived Trustworthiness

143

emotions. The five basic (or primary) emotions are: fear, anger, sadness, disgust, and happiness, often viewed as common across humans in all settings (see Ekman and Friesen 1975; Stets 2003).1 Secondary emotions are viewed as more culturally variable and typically as derivatives of the primary emotions. Affect is defined as any evaluative orientation towards an object, highlighting the more cognitive aspect of the concept.2 Within exchange theory recent efforts to understand the role of emotions in exchange represent an important return to fundamental issues of social solidarity and cohesion, but until recently little data has been collected on affect or emotion in exchange settings. In the past two decades research has been decidedly more structural focusing on the role of network position and characteristics of the network that might affect the distribution of power across those network positions. In many cases networks are actually the primary focus of analysis, not the social relations of which they are composed.

EXCHANGE AND EMOTION Among the more recent efforts to examine the non-structural elements of exchange is the work of Lawler and Yoon (1993) on relational cohesion as well as Molm’s (2003) research on variations in behavioral responses to different types of exchange. Lawler and his colleagues developed what they call the “affect theory of social exchange” to explain the effects of emotional responses to exchanges on exchange relationships (e.g., attendant affect) and exchange outcomes including the degree of commitment and solidarity. The goal of this work has been to develop a more comprehensive theory of exchange that includes emotions as a key element in the formation and continuation of relations of exchange. Affect and emotion in Lawler’s work are viewed as emergent properties of exchange that mediate the effects of structure on behavioral outcomes including commitment and cohesion (For more detail, see Chapter 3 by Thye, Lawler, and Yoon). We begin with a brief discussion of the research on uncertainty and commitment that preceded Lawler’s research program on relational cohesion and affect in social exchange. In our view uncertainty is a key determinant not only of commitment in exchange relations, but also of the perceived trustworthiness of one’s exchange partner(s) and emotional responses to exchange, including affect toward one’s partner(s).

1 Kemper (1987) identifies only four: anger, fear, depression, and satisfaction, which are viewed as innate (general or primary) emotions. 2 Gordon (1982, p. 563) writes that sentiment involves “combinations of bodily sensations, gestures and cultural meanings that we learn in enduring social relationships.” Affect control theorists also view sentiments as providing cultural meanings relevant to how the self, others, and social actions in situations are to be interpreted.

CH09-P374095.indd 143

5/26/2008 2:48:16 PM

144

Social Structure and Emotion

UNCERTAINTY AND COMMITMENT IN EXCHANGE RELATIONS Initially, the reduction of uncertainty was viewed as the primary determinant of commitment formation in exchange relations (Cook and Emerson 1978, 1984). Facing uncertain environments, actors involved in exchange are more likely to form committed exchange relations (Cook and Emerson 1978; Kollock 1994) or networks of trusted exchange partners. A significant effect of the emergence of commitment is that it reduces the extent to which actors seek exchange with alternative partners and thus reduces power inequalities within the exchange relation and the network in which it is embedded (Rice 2002). This work, however, does not include reference to the role of affect or emotions in commitment formation, as did Lawler’s subsequent research (e.g., Lawler and Yoon 1996). Cook and Emerson (1984) explicitly explored the effects of uncertainty on relational commitment in exchange networks. They conceived of uncertainty in terms of the subjective probability of concluding a satisfactory exchange with any partner. Low subjective probability of a successful exchange creates high levels of uncertainty. They found that commitment varied directly with uncertainty, increasing when uncertainty was high. As the likelihood of concluding a transaction decreased, an actor was more likely to exchange with one partner exclusively, ignoring possible alternatives. They also found that commitment formation reduced this uncertainty. Cook and Emerson (1984, 13) argue that commitment increases the frequency of exchange and thus over time improves the benefits for those within the relationship. Dyadic commitments may also result in changes in the network structure in which these relations are embedded. Kollock (1994) subsequently investigated commitment formation under conditions of low and high uncertainty. He conceptualized uncertainty in terms of the unknown quality of the goods being exchanged, thus focusing on a different source of uncertainty than Cook and Emerson (1978, 1984) did. Kollock (1994) argued that committed relationships were more likely to form under conditions of high uncertainty about quality to reduce risk and to assure profit. Kollock viewed commitment as a strategy for reducing uncertainty in exchange situations, testing the argument posed originally by Cook and Emerson (1978, 1984). In addition, Kollock investigated the role of trust. He found that perceived trustworthiness of a partner was directly related to increased rates of commitment (under uncertainty) and reduced rates of malfeasance, but again there was no explicit investigation of the affective responses to commitment in this research program. More recent work on trust (Cook 2005) in social exchange relations treats trust as an emergent property of social relations in certain types of exchange situations and trust can be related to both affect and commitment. Yamagishi, Cook, and Watabe (1998) reported that trust emerges in exchange relations under conditions of high uncertainty when actors begin to form commitments to exclusive exchange relations in an attempt to avoid the possibility of exploitation by unknown actors who enter the exchange opportunity structure.

CH09-P374095.indd 144

5/26/2008 2:48:16 PM

Effect of Perceived Trustworthiness

145

Given low uncertainty, actors are more likely to continue to play the market and avoid forming commitments to specific partners to maximize their continued access to highly valued resources. Uncertainty in these experiments refers to the likelihood of being exploited by a new partner in a network of exchange opportunities that changes over time. Uncertainty and vulnerability to exploitation are often defined as two key elements in situations in which trust becomes paramount (Heimer 2001). Lawler views uncertainty as one independent source of commitment, but treats the primary determinant of commitment as the frequency of positive exchange between two parties (especially those of equal power). With the exception of Lawler’s work on commitment and cohesion, and Molm’s research on affective and behavioral commitment in different types of exchange, little of the experimental research in this tradition has focused much attention on the emotional components of exchange relations, though clearly trust and distrust are often emotionally charged. Molm, Takahashi, and Peterson (2000) initiated a research program to examine the effects of type of exchange (reciprocal or negotiated) on various outcomes of exchange, including affective commitment and trust. In their analysis of variations in types of exchange, Molm and her collaborators argue that reciprocal exchange is inherently more uncertain than negotiated exchange. Because exploitation is always possible, actors in reciprocal exchange risk giving benefits unilaterally while receiving little or nothing in return. Moreover, Molm, Takahashi, and Peterson (2000) argue that affective commitment is more likely to form in reciprocal exchange than in negotiated exchange. Because of the inherent uncertainty, actors are likely to attribute a partner’s positive behaviors to personal traits and intentions, which results in the emergence of stronger positive feelings in reciprocal exchange than in negotiated exchange. Molm, Takahashi, and Peterson (2000) also argue that trust should be higher in reciprocal exchange when compared with negotiated exchange, precisely because the fear of exploitation should be stronger in reciprocal exchange. In negotiated exchange (binding or non-binding) people typically have more control over the final terms of exchange because they can opt out if the terms of exchange are unacceptable. Molm’s research facilitates the investigation of more complex exchange situations. In addition, Molm has begun to specify the nature of the precise mechanisms that relate structural determinants of power with the actual use of power by those in various positions of power, something previous theorists had not yet accomplished. Norms of fairness or justice and attitudes toward risk play a central role in this analysis. Conceptions of fairness constrain the use of power under some conditions; especially the use of coercive power, and risk aversion makes some actors unwilling to use the structural power at their disposal for fear of loss (see also Kahneman). In more recent work, Molm, Takahashi and Peterson (2003) have begun to investigate more fully the relative importance of fairness, risk aversion, perceptions of conflict and the strategic use of power in different forms of exchange (i.e., negotiated or reciprocal exchange (for more detail see Chapter 10 by Molm).

CH09-P374095.indd 145

5/26/2008 2:48:16 PM

146

Social Structure and Emotion

One way in which social structure and emotion are thus linked in recent developments within exchange theory, especially in the work of Lawler and to some extent Molm, is through the investigation of the differential effects of the various forms of exchange on emotional responses – such as affective commitment, expressions of affect and perceptions of trust or trustworthiness. (Another line of work compares types of exchange and power use, see Molm, Peterson, Takahashi 1999; Molm, Takahashi, Peterson 2000.) Power differences among actors are generally not the primary focus of the work on emotions, though effects of power differentials are hypothesized. Equal power relations, it is argued, for example, are more likely to lead to positive exchange and positive affect. Emotion, however, is not often dealt with explicitly in the research on power.

FORMS OF EXCHANGE, UNCERTAINTY, PERCEIVED TRUSTWORTHINESS AND AFFECT Molm has developed the argument that different forms of exchange involve different types of uncertainty and risk. This is a fundamental feature of her work on the effects of different forms of exchange (negotiated, reciprocal, generalized, or indirect reciprocity) on the use of power, affect toward an exchange partner, commitment between exchange partners, and assessments of whether one can trust one’s partner. These different levels of uncertainty and types of risk are hypothesized to alter the exchange dynamics and thus the nature of the relationships that form (Molm, Takahashi and Peterson 2000). In general there are two types of exchange relationships, direct and indirect exchanges. Direct exchange is a relationship in which each actor’s outcome is directly dependent on the other actor’s behavior (see Cook and Emerson 1978; Molm and Cook 1995). Indirect exchange is an exchange relationship in which each actor’s outcome is dependent not on the person they gave to but either on the collective entity, or on another member of the network. We have chosen to focus only on direct exchange in this study, since it is the most commonly discussed form of exchange. There are three types of direct exchange that we are concerned within this chapter: binding negotiated, non-binding negotiated exchange, and reciprocal exchange. In negotiated exchange, actors engage in a joint decision process, such as explicit bargaining, in which they seek agreement on the terms of exchange. Most economic exchanges and some social exchanges fit this category. There are two subtypes of negotiated exchange: binding and non-binding. In binding negotiated exchange, once an agreement is reached, it is typically immediately enacted or consummated (Cook and Emerson 1978; Cook et al. 1983; Kollock 1994; Lawler and Yoon 1993, 1996). In non-binding negotiated exchange, once an agreement is reached, some further action must be taken for the actor to

CH09-P374095.indd 146

5/26/2008 2:48:16 PM

Effect of Perceived Trustworthiness

147

follow through on her agreement to complete the exchange (Heckathorn 1985; Rice 2002). In contrast, in reciprocal exchange, actors’ contributions to the exchange are separately performed and not explicitly negotiated. That is, in reciprocal exchange, an actor initiates exchanges individually, by performing a beneficial act for another actor, without knowing whether, when or to what extent the other actor will reciprocate this beneficial act (Molm, Peterson and Takahashi 1999; Molm, Takahashi and Peterson 2000). Many social exchanges take this form. Uncertainty in exchange situations occurs when there is fear on the part of each actor that she may be cheated by engaging in an exchange, i.e., the fear of exploitation or opportunism exists.3 Opportunism (the term preferred by Oliver Williamson, among others) occurs when an actor reneges on a promise, regardless of whether the promise is implicit or explicit. The primary source of uncertainty in binding negotiated exchange is typically the uncertainty that surrounds finding an exchange partner and sometimes reaching an agreement.4 But, once an agreement is made, the transaction is completed, allowing no chance for opportunism on the part of the exchange partner. Thus binding negotiated exchange has the least amount of uncertainty. Non-binding negotiated exchange involves considerably more uncertainty. The uncertainty about finding an exchange partner and reaching an agreement remains. But, in addition, there is a greater potential for opportunism since actors can renege on their agreements or simply fail to follow through on their agreements. Rice (2002) compared binding and nonbinding negotiated exchange and showed that actors engaged in non-binding negotiated exchange were more likely to engage in repeat positive exchange than actors in binding negotiated exchange. This behavioral commitment between specific exchange partners served to reduce the uncertainty. As Molm argues reciprocal exchange entails the most uncertainty, since there is little information, at least initially to predict when and if a partner will return a “benefit.” Exchange is initiated without knowledge that there will ever be a return. Several studies support the argument that a certain degree of uncertainty or risk is necessary for perceptions of trustworthiness to emerge (Gambetta 1988; Gerbasi 2007; Malhotra and Murnighan 2002). We define perceived trustworthiness as a belief that one’s exchange partner will not behave opportunistically. Without uncertainty, actors are unlikely to attribute the cause of the partner’s behavior to dispositional factors hence they have little basis for an assessment of perceived trustworthiness. Hence, under certainty no causal attributions are necessary. Uncertainty in the exchange situation, however, does create the need for causal attributions that provide the basis for assessments of perceived 3 These situations include when the quality of the goods exchanged are unknown, when there is an opportunity to exploit, when the interaction partner has an incentive to act in a way that imposes costs (or harm) to the actor, and when the actor does not have enough information to predict whether the partner will act in a way that exploits her. 4 Another possible source of uncertainty is the quality of the goods exchanged. See Kollock (1994) for further discussion.

CH09-P374095.indd 147

5/26/2008 2:48:16 PM

148

Social Structure and Emotion

trustworthiness. Thus as uncertainty increases the perceived trustworthiness of one’s exchange partner should increase if the partner cooperates and is not opportunistic. Hypothesis 1:

Given positive exchange, as uncertainty (measured by the type of exchange) increases so should the levels of perceived trustworthiness of one’s exchange partner(s).

PARTNER BEHAVIOR In addition to the uncertainty induced by the type of exchange, clearly the behavior of the partner also determines the amount of uncertainty a participant faces, as noted above. Experience with a particular partner increases or decreases uncertainty over time. If a partner consistently engages in positive exchange (meaning that the partner does not behave opportunistically), uncertainty about interacting with this partner will be reduced. Similarly if this partner consistently defects (i.e., acts opportunistically), uncertainty will increase. Perceived uncertainty of the partner’s behavior should be inversely associated with perceived trustworthiness; focal actors are likely to perceive the partners who cooperate consistently (entailing low uncertainty and low risk) as more trustworthy. Conversely, those partners who fail to cooperate and behave opportunistically (i.e., generating high uncertainty and potential risk) are likely to be perceived as less trustworthy. Hypothesis 2:

When opportunism emerges, as uncertainty increases, levels of perceived trustworthiness of one’s exchange partner(s) should decrease.

In addition to being associated with perceived trustworthiness, the behavior of the other has a direct effect on perceived affect. Work by Lawler and his colleagues (2000) predicts that repeated positive exchange in binding negotiated and productive exchange situations can lead to positive affect. Thus partner behavior clearly has an effect on affect toward that partner. Positive exchange leads to positive affect. The majority of the work by Lawler and his colleagues (Lawler and Yoon 1993, 1996, 1998) does not address types of exchange in which there is the possibility for opportunism hence they hypothesize that repeated exchange leads to positive affect. But when the possibility for opportunism is introduced, the situation becomes more complicated for the actors involved and affect should be greatly influenced by the knowledge of this potential for opportunism (or exploitation). When the possibility of opportunism exists, but is not taken, which may occur in non-binding negotiated exchange, high positive levels of affect are likely to emerge. On the other hand, when the possibility of opportunism exists and exploitation occurs, then negative affect is likely to emerge. Hypothesis 3a: Hypothesis 3b:

CH09-P374095.indd 148

Given positive exchange behavior, as uncertainty increases positive affect toward one’s exchange partner(s) should also increase. When opportunism occurs, positive affect toward one’s exchange partner(s) should decrease as uncertainty increases.

5/26/2008 2:48:16 PM

Effect of Perceived Trustworthiness

149

PERCEIVED TRUSTWORTHINESS Through repeated exchange actors “learn” about one another. By testing relationships with multiple partners, actors also begin to perceive “stable” traits of their potential partners. For example, the behavior of partners can lead to perceptions of their trustworthiness relative to other potential exchange partners. We argue that perceived trustworthiness precedes affect. Clearly low levels of trustworthiness should be associated with negative affect. When a partner is frequently non-cooperative, or acts opportunistically, the focal actor will tend to have negative affect toward that partner. On the other hand, when the partner is cooperative, the focal actor will tend to have positive affect toward that partner. Hypothesis 4:

Perceived trustworthiness of one’s exchange partner(s) is positively associated with positive affect toward one’s exchange partner(s).

ATTRIBUTION AND PERCEIVED INTENTIONALITY Uncertainty is directly related to the type of exchange, which in turn conditions the causal attribution of the exchange partner’s behavior. The causal attribution of the partner’s behavior subsequently conditions the formation of perceptions of a partner’s trustworthiness which influences affect. After the initial completion (or not) of an exchange, we argue that the actor attempts to understand her partner’s behavior through an attribution process, by which the actor uses information to arrive at causal explanations for events (Fiske and Taylor 1991). An attribution process is initiated when an actor needs to predict the future and control events. We are primarily concerned here with dispositional attributions in which the actor points to something internal to her partner to explain her behavior.5 We assume that dispositional attributions provide a basis for assessments of partner characteristics such as trustworthiness and promise-keeping. When actors make a dispositional attribution regarding the cause of their partner’s behavior they typically assume the behavior is intentional, rather than caused by the situation (a.k.a. a situational causal attribution6). Work by Gerbasi (2007) supports the findings of Hegtvedt, Thompson, Cook (1993) and Newman (1981) that actors are likely to make both situational and dispositional attributions in exchange situations, suggesting that these attributions are not mutually exclusive. Gerbasi (2007) also finds that exchange participants generally make stronger situational attributions than dispositional attributions in exchange situations, 5 An example of a statement indicative of a dispositional attribution is: “My partner defected or failed to complete our exchange because she is irresponsible.” 6 An example of a statement indicative of a situational attribution is: “My partner completed our exchange because she was bound by our agreement.”

CH09-P374095.indd 149

5/26/2008 2:48:17 PM

150

Social Structure and Emotion

although the situational attributions do not have a strong effect on relational outcomes. Given our focus on perceived trustworthiness and affect, however, we focus on dispositional causal attributions of the perceived intention of the behavior of one’s exchange partner. We agree with Lawler and his colleagues that positive exchange can bring about positive affect. Unlike prior work on emotions in exchange interactions, we take note of the prevalence of negative emotions and affect that also tend to emerge under various conditions. The conditions that we argue tend to produce negative emotions and affect are those types of exchange in which opportunism is possible and one partner has engaged in exploitation. These feelings, we argue, derive from dispositional causal attributions made concerning the intentions revealed by the behavior of others, especially when their behavior is non-cooperative or opportunistic. These dispositional attributions help actors make sense of their partner’s behavior and in turn interpret their own emotional responses and affect. We argue that structural uncertainty and the behavior of exchange partners directly affect perceived trustworthiness. Specifically, given positive exchange structural uncertainty is positively related to perceived trustworthiness. Behavior is associated with perceived trustworthiness, meaning that the more cooperative a partner is, the higher the perceived trustworthiness of that partner should be. Perceived trustworthiness directly influences perceived affect toward this exchange partner is a positive way (i.e., the higher the perceived trustworthiness of a partner, the more positive the affect toward that partner). Partner behavior also has a direct (non-mediated) effect on affect toward that partner (see Figure 9.1 for a graphical depiction). We provide evidence concerning the role of uncertainty (which varies by type of exchange) and partner behavior (i.e., the extent of cooperative exchange) on assessments of trustworthiness of one’s exchange partner(s) and the degree of positive or negative affect toward the partner(s). Relational outcomes (such as affect) are primarily mediated by assessments of the trustworthiness of one’s exchange partners which is determined by the level of uncertainty in the exchange situation and the extent of cooperative (or non-cooperative) exchange behavior of the partner. We include some evidence on the effect of the dispositional causal attributions actors make regarding the source of the positive or

Structural uncertainty

Perceived trustworthiness

Affect

Partner behavior

FIGURE 9.1

CH09-P374095.indd 150

Proposed theoretical model.

5/26/2008 2:48:17 PM

Effect of Perceived Trustworthiness

151

negative exchange behavior, though this is not the primary focus of our chapter. We hope to begin to identify the primary drivers of emotional responses in exchange situations, and to investigate the role of negative behavior on emotion and other relational outcomes in the future. We present preliminary results from a set of ongoing studies on this topic. In conclusion, we discuss the ways in which key variables may be interrelated in fairly complex ways as determinants of affective and emotional responses to exchange processes in different types of exchange. In particular, like Molm, we draw a contrast between negotiated (including non-binding negotiated) and reciprocal exchanges. The findings we report are from an experiment (Gerbasi 2007) on perceived trustworthiness and attributions in various types of exchange (in which the behavior of the other participants is controlled to vary the extent of cooperation). Here we focus on trustworthiness and affect.

METHODS We explored these assertions in negatively connected exchange networks. Exchange relations are negatively connected when exchange in one relation decreases the frequency or amount of exchange in another relation.7 In each network there was one participant and three simulated actors. Hence, each actor could interact with each potential partner but only complete one exchange on each trial in the experiment (over multiple trials). There were no power differences in these networks and each person believed that she could potentially exchange with any other partner in the four-person network. EXPERIMENTAL DESIGN AND BASIC PROCEDURES

We used a 3 ⫻ 2 modified factorial design. The two factors that are crossed in the design are the form of exchange (three types: binding negotiated exchange, non-binding negotiated exchange and reciprocal exchange) with the level of cooperation of the simulated exchange partners (two levels: high or low cooperation) to examine the effects of high and low cooperation on the attribution process and perceived trustworthiness in different types of exchange settings. This should result in six conditions. Binding negotiated exchange, however, has not traditionally been operationalized in such a way that there is opportunity for 7 Emerson (1972, p. 70) defines connected networks as “two exchange relationships A:B and A:C as connected at A if the frequency or magnitude of a transaction in one relationship is a function of transaction of the other relation.” Thus a negatively connected network is one in which an increase in the frequency or amount of exchange in one relation implies an inverse function, or a decrease in the frequency of amount of exchange in another relation(s). On the other hand a positively connected network is one in which an increase in the frequency or amount of exchange in one relationship implies a positive function, or an increase in the frequency or amount of exchange in the other relation(s).

CH09-P374095.indd 151

5/26/2008 2:48:17 PM

152

Social Structure and Emotion

defection or failed exchange hence we do not include a low cooperation, binding negotiated exchange condition. The result is five experimental conditions. Thirty subjects were randomly assigned to each of the resulting five conditions for a total of 150 undergraduate participants in the study.8 Subjects engaged in either binding negotiated, non-binding negotiated or reciprocal exchanges to earn money. The exchange relations were negatively connected such that in negotiated exchanges participants could negotiate with all potential partners, but could only complete one transaction on each exchange opportunity. In reciprocal exchange a participant could only initiate exchange with one other partner, but could receive benefits from multiple partners (Molm, Peterson, Takahashi 2003). All interaction was mediated through computers placed in isolated rooms. After reading detailed instructions and engaging in practice trials, the participants engaged in a series of exchanges. At the end of each opportunity for exchange they were informed of the source and the amount of any points they had gained, and their own total earnings were also displayed on their computer screens. There were two phases to the experiment. In each phase the participants engaged in numerous rounds of exchange and then answered questionnaires at the end of each phase. EXPERIMENTAL MANIPULATIONS

Forms of Exchange: Three different settings were created to manipulate the type of exchange: binding negotiated, non-binding negotiated, and reciprocal. The settings were similar to those used by Molm and her colleagues in previous studies (Molm, Peterson, Takahashi 1999; Molm, Takahashi, Peterson 2000). These exchange conditions were designed to be comparable except for their defining differences in the actual structure of the exchange (i.e., negotiated versus reciprocal). In binding negotiated exchange, during each exchange opportunity, the participant can make several offers to multiple partners but in the end only one exchange can occur on each trial. The participant is also given a pool of “points” they can use to negotiate with her potential partners. The participant is unaware of her partner’s alternative exchange opportunities. The “participants” sent offers to others in the network, stating their initial offers. If after a randomized amount of time (ranging from 10 to 20 seconds) the participant has not sent any offers to her potential partners, the simulated actors sent pre-determined offers. Once a request has been made, the other could accept the offer, reject the offer or make a counter offer. The simulated actors were programmed to accept profitable deals. If after 2 minutes of negotiation no agreement was reached, actors 8 Although gender is not a variable in this study, we used a 50/50 gender distribution randomized across conditions. Other basic demographic information (i.e., age and ethnicity) were also collected. There were no effects for gender, age or ethnicity so these variables are excluded from further analysis.

CH09-P374095.indd 152

5/26/2008 2:48:17 PM

Effect of Perceived Trustworthiness

153

retained their original “points” and a new round commenced. On the second exchange opportunity, if no agreement was reached the simulated actors repeated the programmed offer for the previous exchange. If an agreement was reached, the simulated actor who completed the agreement on the first trial repeated that offer, while another simulated actor offered a more profitable option to the participant (a randomly assigned value between one to five points more than the participant’s previous exchange outcome). These patterns continued until all of the exchange opportunities ended. Non-binding negotiated exchange followed the same procedures as binding negotiated exchange, except that it added a step. If an agreement was reached, the participant faced the additional decision of keeping their agreement or defecting (i.e., not following through with the exchange at the agreed terms). If both partners (the participant and the simulated actor) followed through on the agreement, the participants received the profit as negotiated. The participants were told if they reneged on the agreement they would keep all of their initial “points” plus any “profit” coming to them if their partner followed through. However, if the participant followed through and their partner defected, the participant would lose the “profit.” If both partners reneged, they would each keep only their initial “points.” In reciprocal exchange, participants were told they could choose to give “benefits” to other “participants” in the network. It was possible for a participant to give from 0 to 12 points to any other “participant” in the network on each trial. It was thus possible to receive benefits from multiple actors, but participants could only give benefits to one other “participant.” The exchanges were organized sequentially, such that the simulated actors gave benefits on the first exchange opportunity to induce reciprocal exchange. If the participant returned the benefit, the simulated actor who gave the benefit on the first trial would repeat that gift at the same value, while another simulated actor would give a more profitable benefit to the participant (based on a random increment ranging from 1 to 6 points). In the two forms of negotiated exchange, the participants engaged in 30 exchange opportunities and in the reciprocal exchange conditions the participants engaged in 60 exchange opportunities.9 All five conditions took roughly the same amount of time (one hour) and subjects were not told the number of exchange opportunities that would take place to avoid end game effects. In addition, the monetary value of points was adjusted so that the participants in all conditions would earn approximately the same amount of money. Level of Cooperation: There were two types of exchange environments created by the simulated actors, a high cooperation environment and a low cooperation environment. The environment was manipulated to test the effects of 9 The number of exchange opportunities was based on the number of interactions that could be completed in 1 hour. In addition this operationalization is consistent with that used by Molm and her colleagues (1999, 2000) who argue that relational phenomena take longer to emerge in reciprocal exchange due to the sequential nature of the interactions.

CH09-P374095.indd 153

5/26/2008 2:48:17 PM

154

Social Structure and Emotion

different cooperation levels on perceived trustworthiness, causal attributions, and affect. Each participant was randomly assigned to one of these conditions. In the high cooperation environment the participant interacted with simulated actors who cooperated 100% of the time.10 In binding negotiated exchange, this translates into the simulated partners accepting any deal that either breaks even or gains them profit. In non-binding negotiated exchange, the simulated others, accept any deal that either breaks even or gains them profit. They also follow through on each one of their agreements (i.e., they do not renege on agreements). In reciprocal exchange, the simulated others returned the benefit they receive on the following trial 100% of the time. Alternatively, in the low cooperation environment the participant interacted with simulated actors who cooperated slightly more than 50% of the time.11 A random number generator created a probability for defection on each trial. If the random number was greater than 0.56 the simulated actor would defect and if it was lower than 0.56 the simulated actor would cooperate. As mentioned above, there is not a low cooperation binding negotiated exchange condition, because negotiated exchange is typically operationalized in such a way that there is no opportunity for defection or failed exchange. In the non-binding negotiated exchange condition, the simulated partners agreed to exchanges that would either break even or gain profit, but only 56% of the time would they follow through on those agreements. The remaining 44% of the time they would defect on these agreements. In the reciprocal exchange condition, the simulated partners would return a benefit they had received on the subsequent trial 56% of the time and would not return that benefit 44% of the time. Measures Affect: To obtain measures of affect, we used Lawler and Molm’s scales measuring affect. Affect was measured on a 10 point scale (1 negative, 10 positive) based on the following questions: Describe your feelings towards your partners (1 negative/10 positive): These measures include: bad/good; awful/nice; unhelpful/ helpful; uncooperative/cooperative; displeased/pleased; unhappy/happy; not satisfied/satisfied; not contented/contented; boring/exciting; not interesting/interesting; unenthusiastic/enthusiastic; unmotivating/motivating; uncommitted/committed. The affect score used here was the average of all of these measures (Cronbach alpha ⫽ 0.95). Perceptions of Trustworthiness: We measured trustworthiness using several attitudinal scales (with 1 being the most negative and 10 being the most positive). Participants were asked to rate their partners as untrustworthy/trustworthy, 10

The rate of cooperation in the high cooperation condition is at 100% so that it would be clearly differentiated from the low cooperation condition. 11 The simulated actors in the low cooperation conditions cooperate 56% of the time. This rate of cooperation is based on prior data collection with four actual participants occupying each of the roles.

CH09-P374095.indd 154

5/26/2008 2:48:17 PM

Effect of Perceived Trustworthiness

155

unreliable/reliable, undependable/dependable, uncredible/credible. Scores were the average of all of these measures (Cronbach alpha ⫽ 0.93). Attributions: Participants were asked to rate the perceived source of their partner’s behavior. The participant had to determine if the attribution was dispositional or situational. To gauge this attribution, the participant was asked to rate how much of his/her partner’s behavior is due to something internal to his/her partner (responding to a 1 to 10 scale, 1 being “none” and 10 being “all”). This measure was collected at the end of each phase of exchange.

RESULTS PERCEIVED TRUSTWORTHINESS

In Hypothesis 1 we argued that uncertainty was positively related to perceived trustworthiness. Specifically we argued that given cooperative behavior, as uncertainty increases (as measured by the type of exchange), so should the perceived trustworthiness of one’s exchange partner(s). This would predict that actors in the reciprocal exchange condition would display the highest levels of perceived trustworthiness as they face the highest level of uncertainty, followed by those in non-binding negotiated exchange and finally those in binding negotiated exchange who face the lowest level of uncertainty. We present a figure depicting the mean levels of perceived trustworthiness for the three high cooperation conditions (see Figure 9.2). Analysis of variance shows that the levels of perceived trustworthiness vary significantly between conditions, F (2, 87) ⫽ 6.52, p ⬍ 0.005. The levels of perceived trustworthiness are in the following order, reciprocal exchange has the highest level of perceived trustworthiness of one’s exchange partners, followed by non-binding negotiated exchange and then binding negotiated exchange. The mean differences between the reciprocal exchange and binding negotiated exchange conditions are significant, as is the difference between non-binding negotiated exchange and binding negotiated exchange. The difference between reciprocal exchange and non-binding negotiated exchange is in the predicted direction, but does not reach significance (see Table 9.1). In Hypothesis 2 we compare the effect of partner behavior on perceived trustworthiness. We suggest that high levels of cooperation will lead to high levels of perceived trustworthiness and low levels of cooperation will lead to low levels of perceived trustworthiness. Analysis of variance shows that the levels of perceived trustworthiness vary significantly between the high and low cooperation conditions, F (4, 145) ⫽ 10.622, p ⬍ 0.001, providing initial support for Hypothesis 2. Examining Figure 9.3, it is clear that in comparison to the high cooperation conditions, the corresponding low cooperation conditions produced lower levels of perceived trustworthiness. Interestingly, post-hoc analysis of the data suggests an interaction between type of exchange and behavior (see Table 9.2). There is a significant difference in perceived trustworthiness between the non-binding negotiated exchange low cooperation condition and the reciprocal exchange low

CH09-P374095.indd 155

5/26/2008 2:48:17 PM

156

Social Structure and Emotion

7

6

5

4

3 Reciprocal high cooperation

FIGURE 9.2

Non-binding negotiated high cooperation

Binding negotiated

Average level of trustworthiness in the high cooperation conditions.

Post-hoc analysis of mean difference scores for perceived trustworthiness between types of exchange (Bonferroni comparisons)

TABLE 9.1

(I) Exchange type

(J) Exchange type

Binding negotiated exchange Non-binding negotiated exchange Non-binding negotiated exchange

Dispositional attribution mean difference (I ⫺ J) Total ⫺1.65* (0.6)

Reciprocal exchange

⫺2.05** (0.6)

Reciprocal exchange

⫺0.41 (0.6)

N ⫽ 90. Standard errors in parentheses. p ⬍ 0.05 p ⬍ 0.005 (two-tailed tests).

*

**

cooperation condition. The non-binding negotiated exchange low cooperation condition has a much lower mean level of perceived trustworthiness than the low cooperation reciprocal exchange condition. These results suggest that it is something about the rules of exchange that heighten the salience of opportunistic behavior. In non-binding negotiated exchange opportunism creates significantly

CH09-P374095.indd 156

5/26/2008 2:48:18 PM

157

Effect of Perceived Trustworthiness Average level of trustworthiness 8 7 6 5 4 3 2 1 0 Reciprocal high cooperation

FIGURE 9.3

Reciprocal low cooperation

Binding negotiated

Non-binding Non-binding negotiated negotiated high cooperation low cooperation

Average levels of perceived trustworthiness in all five conditions (N ⫽ 150).

Post-hoc analysis of mean difference scores for perceived trustworthiness between types of exchange (Bonferroni comparisons)

TABLE 9.2

(I) Exchange type

(J) Exchange type

Binding negotiated exchange

Non-binding negotiated exchange high cooperation Non-binding negotiated exchange low cooperation Reciprocal exchange high cooperation Reciprocal exchange low cooperation Non-binding negotiated exchange low cooperation

Non-binding negotiated exchange high cooperation

Non-binding negotiated exchange low cooperation Reciprocal exchange high cooperation

Dispositional attribution mean difference (I ⫺ J) Total ⫺1.65* (0.59) 1.3 (0.59) ⫺2.05** (0.59) ⫺1.24 (0.59) 2.95** (0.59)

Reciprocal exchange high cooperation Reciprocal exchange low cooperation Reciprocal exchange high cooperation

⫺.41 (0.59) .41 (0.59) ⫺3.36** (0.59)

Reciprocal exchange low cooperation

⫺2.54** (0.59)

Reciprocal exchange low cooperation

0.82 (0.59)

N ⫽ 150. Standard errors in parentheses. * p ⬍ 0.05 ** p ⬍ 0.005.

CH09-P374095.indd 157

5/26/2008 2:48:18 PM

158

Social Structure and Emotion

lower levels of perceived trustworthiness, but in reciprocal exchange, the same level of opportunism does not. AFFECT

In Hypothesis 3a we argued that there is a positive relationship between uncertainty and affect. Figure 9.4 presents evidence that supports this argument. In addition, analysis of variance shows that the levels of perceived affect vary significantly between conditions F (4,145) ⫽ 10.03, p ⬍ 0.001. Further support for this hypothesis is revealed in post-hoc analysis. A similar pattern to the results for perceived trustworthiness is seen when the different types of exchange and partner behavior are compared (see Table 9.3). The conditions in which there are higher levels of uncertainty and high levels of cooperation (and hence higher perceived trustworthiness) tend to have higher levels of reported positive affect. Hypothesis 3b also received some support. We argued when opportunism exists, positive affect toward one’s exchange partner(s) should decrease as uncertainty increases. The reciprocal exchange high cooperation condition, for example, has the highest level of reported affect toward one’s partner. This was expected given the high level of perceived trustworthiness in this condition (a finding also supported by Molm’s recent work 2000, 2003). The non-binding negotiated

Average level of positive affect 8 7 6 5 4 3 2 1 0 Reciprocal high cooperation

FIGURE 9.4

CH09-P374095.indd 158

Reciprocal low cooperation

Binding negotiated

Non-binding negotiated high cooperation

Non-binding negotiated low cooperation

Average levels of positive affect in all five conditions (N ⫽ 150).

5/26/2008 2:48:18 PM

159

Effect of Perceived Trustworthiness

Post-hoc analysis of mean difference scores for affect between types of exchange (Bonferroni comparisons)

TABLE 9.3

Dispositional attribution mean difference (I ⫺ J) Total

(I) Exchange type

(J) Exchange type

Binding negotiated exchange

Non-binding negotiated exchange high cooperation

⫺ 1.24 (0.51)

Non-binding negotiated exchange low cooperation

1.18 (0.51)

Non-binding negotiated exchange high cooperation

Reciprocal exchange high cooperation

⫺ 1.53* (0.51)

Reciprocal exchange low cooperation

⫺ 1.27 (0.51)

Non-binding negotiated exchange low cooperation

2.42** (0.51)

Reciprocal exchange high cooperation

⫺ 0.28 (0.51)

Reciprocal exchange low cooperation

⫺ 0.01 (0.51)

Non-binding negotiated exchange low cooperation

Reciprocal exchange high cooperation

⫺ 2.71** (0.51)

Reciprocal exchange low cooperation

⫺ 2.45** (0.51)

Reciprocal exchange high cooperation

Reciprocal exchange low cooperation

0.36 (0.51)

N ⫽ 150. Standard errors in parentheses. p ⬍ 0.05 p ⬍ 0.005.

*

**

exchange condition with high levels of partner cooperation has the next highest level of reported affect. These findings support the hypothesized effect of levels of uncertainty (and the extent of cooperative partner behavior) on affect. The one relatively surprising result is the high level of affect in the reciprocal exchange low cooperation condition. The level of affect in this condition is only significantly different from the level of affect in the low cooperation, non-binding negotiated exchange condition. This suggests that opportunism has a different effect depending on the type of exchange (as was the case concerning the effect of opportunism on perceived trustworthiness in reciprocal exchange). This specific finding may be the result of an interesting structural feature of reciprocal exchange. In Malinowski’s (1922) account of the Kula ring, participants often excused temporal delays of gift giving by their partners because they were aware of the situational constraints on their partners. Partners may have many gifts they need to return or they do not have the resources to return all gifts in kind in Malinowski’s account, thus creating constraints on a partner’s ability to reciprocate. In fact, in reciprocal exchange there is often a temporal lag in the return of gifts.

CH09-P374095.indd 159

5/26/2008 2:48:18 PM

160

Social Structure and Emotion

This finding may also partially be a result of a constrained social process that emerged in the experimental situation. In this experimental setting, the actor was allowed to give only one gift at a time, but could receive multiple gifts (i.e., a gift from more than one person at a time) on any exchange opportunity. Most actors seemed to feel a necessity to reciprocate all gifts they received, which resulted in a temporal lag in reciprocity12. The behavior of the actors did not differ in this respect between the two reciprocal exchange conditions. They did not always reciprocate immediately in either condition. Since the focal actor wasn’t immediately reciprocating gifts, due to situational constraints, the behavior of the pre-programmed partners in the low cooperation condition was not viewed as explicit defection, but rather as the partners’ fulfilling of their obligations to others resulting from the nature of the exchange situation. Qualitative data collected at the end of the experiment shows that 70% of the participants who engaged in reciprocal exchange viewed any delays in reciprocity as a result of their partners fulfilling obligations to other partners (i.e., “I didn’t mind when partner X took awhile to return my coins, I had to return gifts to partners Y and Z too.”). In short, complex turn taking emerged in this exchange condition. Gerbasi (2005) suggests that actors in reciprocal exchange create accounts for their partner’s behavior, thereby reducing their uncertainty and the need for retaliation (in the case of defection) or the immediate return of gifts (in the case of positive exchange). The high levels of affect and perceived trustworthiness in the low cooperation reciprocal exchange condition may also be explained in part by the attributions made by the actors in reciprocal exchange. The account making done by actors in reciprocal exchange suggests they make situational causal attributions regarding their partner’s behavior. That is, they think the primary cause of their partner’s behavior is external rather than internal. In this case, the account they create regarding other obligations explains the behavior of their partner. Many of the existing arguments regarding attribution in reciprocal exchange suggest that the actors will make dispositional causal attributions concerning the cause of their partner’s behavior. Examining the types of attributions made in reciprocal exchange, there is very little difference between the high cooperation and low cooperation conditions in this study. In both conditions actors tend to make situational causal attributions regarding the cause of their partner’s behavior consistent with Gerbasi’s (2007) argument. There is some evidence that they also make internal attributions regarding their partner’s behavior (i.e., dispositional attributions), but there is a stronger tendency for them to make an external attribution (to situational constraints). Figure 9.5 includes the rates of dispositional and situational causal attributions. It is clear from this table that participants in both of the reciprocal 12

This resulted from getting gifts from Actors B and C at time 1, such that the actor would reciprocate Actor B at time 2 and Actor C at time 3, and at time 4 she would reciprocate the gifts she was given at time 2.

CH09-P374095.indd 160

5/26/2008 2:48:19 PM

161

Effect of Perceived Trustworthiness 8

7

6

5

4

3

2

1

0 Reciprocal Reciprocal exchange exchange high cooperation low cooperation

Binding negotiated exchange high cooperation

Dispositional attribution

FIGURE 9.5

Non-binding negotiated exchange high cooperation

Non-binding negotiated exchange low cooperation

Situational attribution

Average levels of attributions in all five conditions.

exchange conditions have a strong tendency to make situational attributions, even taking into account the behavior of the others (see Figure 9.5 for a graphical display of these findings). Our results for binding negotiated exchange replicate findings obtained by Lawler and Yoon (1993) in which positive exchange produces positive affect among exchange partners, even with low levels of uncertainty. The low cooperation non-binding negotiated exchange condition has the lowest level of affect (see Table 9.3), as predicted in Hypothesis 3b. Given the relatively high level of defection and the explicit nature of the partner’s reneging on their promises, we would be quite surprised if the participants in this condition did not exhibit the lowest levels of affect. Qualitative data collected at the end of the experiment suggest that participants in the low cooperation conditions, particularly in the non-binding negotiated exchange condition, had extremely negative emotional responses to the exchange experience. They frequently displayed anger toward their exchange partners. They used words such as “hate, frustration, anger/angry, betrayal.” Approximately 80% of the participants in the non-binding negotiated exchange low cooperation condition used such phrases. The rate was approximately 65%

CH09-P374095.indd 161

5/26/2008 2:48:19 PM

162

Social Structure and Emotion

in the reciprocal exchange low cooperation condition. On the other hand, the high cooperation conditions seemed to induce positive feelings of satisfaction. Words such as “nice, cooperative, friendly, helpful” were consistently used in these conditions. In the non-binding negotiated exchange high cooperation condition, the findings suggest that guilt occasionally arises and it correlates highly with behavior. Participants who reneged on promises to their partners tended to express guilt, or say they felt bad about defecting on a partner. These participants also made excuses for their behavior such as, “I was trying to earn as many points as possible” and “I didn’t mean to hurt my partner.”

THE RELATIONSHIP BETWEEN PERCEIVED TRUSTWORTHINESS AND AFFECT In Hypothesis 4 we argued that there is a positive relationship between perceived trustworthiness and affect. To test this hypothesis we estimated linear regression equations to examine the effects of level of cooperation (partner behavior) and perceived trustworthiness on affect (see Table 9.4). The results support Hypothesis 4, perceptions of trustworthiness are a positive and significant predictor of positive affect, controlling for the type of exchange and the behavior of the other. The uncertainty created by the type of exchange clearly has an effect on the emergence of relational phenomena in exchange situations including affect toward one’s exchange partners and related emotions. These results are consistent with the recent findings from studies by Lawler, Molm, and their collaborators. Besides investigating emotional responses as they vary by type of exchange, an interesting question is what happens when the type of exchange changes

TABLE 9.4

Effects of the type of exchange and perceived trustworthiness on affect (N ⫽ 150)

Non-binding negotiated exchange1

Model 1

Model 2

0.7

0.46

Reciprocal exchange

2.07**

1.76*

Partner behavior

1.34**

0.95*

Constant

3.84

3.14

R2

0.15

0.2

Perceived trustworthiness

0.23**

Notes: Data Gerbasi (2006). 1 Comparison category is binding negotiated exchange. * p ⬍ 0.01 **

p ⬍ 0.005.

CH09-P374095.indd 162

5/26/2008 2:48:19 PM

Effect of Perceived Trustworthiness

163

during the course of an interaction. This is one direction for future research on types of exchange and relational outcomes. It moves us beyond more static comparisons between types of exchange to an investigation of the dynamics of exchange relations (Cook, Cheshire and Gerbasi, 2006).

CONCLUDING COMMENTS In this chapter we have begun to investigate the affective and emotional dynamics of social exchange. Different modes of exchange, as Molm and Lawler have noted in their research, create quite different interaction processes. Direct two-party negotiation differs along a number of dimensions from reciprocal exchange in which two parties may end up in exchange over time, but only through reciprocal acts of giving and receiving (i.e., no negotiation occurs). In addition, even a simple distinction between binding and non-binding negotiated exchange leads to different predictions about the relational, affective and emotional components of the social process we call exchange. In binding negotiated exchange, the two parties can rest assured that the terms of trade will be honored by both parties. The terms will be “enforced” in the exchange setting by the institutional framework in which the exchange is embedded (i.e., in this case by the experimental authority). In non-binding negotiated exchange it is much less certain that the parties involved will abide by the terms of trade, since “defection” and opportunism can occur (and does in our experimental world). Here the parties to the exchange are more vulnerable and they must come to their own conclusions about the trustworthiness of their potential exchange partners. In such a world we have argued perceived trustworthiness (along with the degree of cooperation of the partner) is the main determinant of the affective relation between the two parties. In addition, we hypothesize that it is in this condition positive emotions are more likely to emerge (though we would need better evidence to back up these claims). Our study was limited in the nature of the data we could bring to bear on such hypotheses since emotional responses were volunteered by our participants at the end of the study and were not systematically collected on a survey or in a more formal interview. The clear link between perceived trustworthiness in the different exchange settings and affect toward one’s partners makes us believe that were we to collect more systematic data on emotions, these two variables would be significant determinants of the extent to which positive as compared with negative emotions emerge in these exchange settings. We anticipate the most negative emotions to appear in the condition in which subjects are allowed to “defect” on their partners or to renege on the promised terms of trade. The data we obtained in the low cooperation non-binding negotiated exchange condition provide support for this claim. An interesting exception, however, is revealed in the findings with respect to low cooperation (or “defection”) under reciprocal exchange. Here the

CH09-P374095.indd 163

5/26/2008 2:48:19 PM

164

Social Structure and Emotion

participants had a tendency to attribute non-exchange to “other obligations” the partner might have had to fulfill (i.e., they made a situational rather than a dispositional causal attribution). By making it less a reflection on the trustworthiness of their partner (i.e., a dispositional attribution), they were able to maintain relatively high levels of cooperation even though in general their partners were less cooperative. In addition, the higher rate of non-cooperation in this condition did not lower affect as much as it did in the non-binding negotiated exchange condition in which promises were broken in many subjects’ minds. Future research should explore more fully the nature of the emotional dynamics of these different conditions of exchange.

REFERENCES Blau, Peter. [1964] 1986. Exchange and Power. New York: Wiley. Cook, Karen S. 2005. “Networks, Norms and Trust: The Social Psychology of Social Capital.” Social Psychology Quarterly 68:4–14. Cook, Karen S., Coye Cheshire, and Alexandra Gerbasi. 2006. “Exchange Network Transitions: Uncertainty, Risk and Shifts in Mode of Exchange.” Presented at American Sociological Association Meetings, Montreal, Canada. Cook, Karen S. and Richard M. Emerson. 1978. “Power, Equity and Commitment in Exchange Networks.” American Sociological Review 43:721–739. Cook, Karen S. and Richard M. Emerson. 1984. “Exchange Networks and the Analysis of Complex Organizations.” Research in the Sociology of Organizations 3:1–30. Cook, Karen S., Richard Emerson, Mary R. Gillmore, and Toshio Yamagishi. 1983. “The Distribution of Power in Exchange Networks: Theory and Experimental Results.” American Journal of Sociology 89:275–305. Ekman, Paul and W. V. Friesen. 1975. Unmasking the Face. A Guide to Recognizing Emotions from Facial Cues. Englewood Cliffs, NJ: Prentice-Hall. Emerson, Richard M. 1972. “Exchange Theory, Part II: Exchange Relations and Networks.” In Joseph Berger, Morris Zelditch and B. Anderson (eds.). Sociological Theories in Progress, pp. 58–87. Boston, MA: Houghton Mifflin. Fiske, Susan T. and Shelley E. Taylor. 1991. Social Cognition. New York: McGraw-Hill. Gambetta, Diego. 1988. Trust: Making and Breaking Cooperative Relations. New York: Blackwell. Gerbasi, Alexandra. 2005. “Cooperation and Defection in Reciprocal Exchange and the Effects on Affect, Cohesion and Commitment.” Group Processes Conference, Philadelphia, PA. Gerbasi, Alexandra. 2007. Attribution and Commitment in Different Types of Exchange. Stanford University. Gordon, Steven L. 1982. “The Sociology of Sentiment.” In: Morris Rosenberg and Ralph H. Turner (eds), Social Psychology: Sociological Perspectives, pp. 562–592. New York: Basic Books. Heckathorn, Douglas D. 1985. “Power and Trust in Social Exchange.” In: Edward J. Lawler (ed.), Advances in Group Processes, pp. 143–167. Greenwich, CT: JAI Press. Hegtvedt, Karen A., Elaine A. Thompson, and Karen S. Cook. 1993. “Power and Equity: What Counts in Attributions for Exchange Outcomes?” Social Psychology Quarterly 56:100–119. Heimer, Carol. 2001. “Solving the Problem of Trust.” In: Karen Cook (ed.), Trust in Society, pp. 40–88. New York: Russell Sage. Homans, G. C. 1974. Social Behavior and Its Elementary Forms. New York: Harcourt, Brace and World. Kemper, Theodore D. 1987. A Social Interactional Theory of Emotions. New York: John Wiley & Sons Inc.

CH09-P374095.indd 164

5/26/2008 2:48:19 PM

Effect of Perceived Trustworthiness

165

Kollock, Peter. 1994. “The Emergence of Exchange Structures: An Experimental Study of Uncertainty, Commitment, and Trust.” American Journal of Sociology 100:313–345. Lawler, Edward J. and Jeongkoo Yoon. 1993. “Power and the Emergence of Commitment Behavior in Negotiated Exchange.” American Sociological Review 58:465–481. Lawler, Edward J. and Jeongkoo Yoon. 1996. “Commitment in Exchange Relations: A Test of a Theory of Relational Cohesion.” American Sociological Review 61:89–108. Lawler, Edward J. and Jeongkoo Yoon. 1998. “Network Structure and Emotion in Exchange Relations.” American Sociological Review 63:871–894. Lawler, Edward J., Jeongkoo Yoon, and Shane R. Thye. 2000. “Emotion and Group Cohesion in Productive Exchange.” American Journal of Sociology 106:616–657. Levi-Strauss, Claude. 1969. The Elementary Structures of Kinship, revised edition. Boston: Beacon. Malinowski, Bronislaw. 1922. Argonauts of the Western Pacific. New York: E.P. Dutton. Malhotra, Deepak and J. Keith Murnighan. 2002. “The Effects of Contracts on Interpersonal Trust.” Administrative Science Quarterly, 47:534–559. Mauss, Marcel. 1950 [1990]. The Gift: The Form and Reason for Exchange in Archaic Societies. New York: W.W. Norton. Molm, Linda D. and Karen S. Cook. 1995. “Social Exchange and Exchange Networks.” In Karen S. Cook, Gary Alan Fine and James S. House (eds.), Sociological Perspectives on Social Psychology. pp. 209–235. Boston, MA: Allyn and Bacon. Molm, Linda D., Gretchen Peterson, and Nobuyuki Takahashi. 1999. “Power in Negotiated and Reciprocal Exchange.” American Sociological Review 64:876–890. Molm, Linda D., Gretchen Peterson, and Nobuyuki Takahashi. 2003. “In the Eye of the Beholder: Procedural Justice in Social Exchange.” American Sociological Review 68:128–152. Molm, Linda, N. Takahashi, and Gretchen Peterson. 2000. “Risk and Trust in Social Exchange: An Experimental Test of a Classical Proposition.” American Journal of Sociology 105:1396–1427. Newman, H. 1981. “Communication with Ongoing Intimate Relationships.” Personality and Social Psychology Bulletin 48:59–80. Rice, Eric R. 2002. “The Effect of Social Uncertainty In Networks of Social Exchange.” Unpublished Ph.D. dissertation. Stanford University, Stanford, CA. Stets, Jan. 2003. “Emotions and Sentiments.” In: John DeLamater (ed.), Handbook of Social Psychology, pp. 309–335. New York: Kluwer Academic/Plenum. Thoits, Peggy. 1990. “Emotional Deviance: Research Agendas.” In: Theodore D. Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 180–203. Albany, NY: SUNY Press. Yamagishi, Toshio, Karen S. Cook, and M. Watabe. 1998. “Uncertainty, Trust and Commitment Formation in the United States and Japan.” American Journal of Sociology 104:165–194.

CH09-P374095.indd 165

5/26/2008 2:48:19 PM

10 The Structure of Reciprocity and Integrative Bonds: The Role of Emotions Linda D. Molm University of Arizona, Tucson, AZ, USA

ABSTRACT A theoretical program of research comparing different forms of exchange – direct negotiated, direct reciprocal, and indirect generalized – has found striking differences in their effects on integrative outcomes, including trust, affective regard, relational solidarity, and interactional fairness. Reciprocal exchange produces stronger integrative bonds than negotiated exchange, while indirect (generalized) exchange produces stronger integrative bonds than either form of direct exchange. Reciprocity theory, the theory generated from this program of research, explains these effects by arguing that two dimensions of the structure of reciprocity in exchange – whether benefits are reciprocated directly or indirectly and whether reciprocity occurs through bilateral or reciprocal unilateral flows of benefit – influence the development of integrative bonds through three intervening causal mechanisms: the risk of nonreciprocity, the expressive value of acts of reciprocity, and the relative salience of the cooperative or competitive “faces” of exchange. While integrative outcomes incorporate both affective and cognitive elements, the role of emotions in this process, as distinct from affect, has not previously been considered. This chapter identifies emotions that are likely to be involved in this process, discusses the potential role of emotions in the causal mechanisms that link the structure of reciprocity to integrative Social Structure and Emotion

CH10-P374095.indd 167

167

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 2:52:13 PM

168

Social Structure and Emotion

outcomes, and asks what the formal incorporation of emotions in reciprocity theory might contribute to our theoretical understanding of these processes. Key Words: Social Exchange, Reciprocity Theory, Cohesion

Reciprocity is a defining feature of social exchange, the one that gives exchange its name: “Benefits obtained through social process are contingent upon benefits provided in exchange” (Emerson 1981:32). The importance of reciprocity in social life has been recognized by numerous scholars; Hobhouse (1906:12) called reciprocity “the vital principle of society,” Becker (1956:1) referred to our species as “homo reciprocus,” and Simmel (1950:387) noted that social equilibrium and cohesion could not exist without “the reciprocity of service and return service.” While all forms of exchange are characterized by some type of reciprocity, the structure of reciprocity varies. For the past decade, my collaborators and I have conducted a program of experimental research comparing three forms of exchange – direct negotiated exchange, direct reciprocal exchange, and indirect generalized exchange – that vary on two key dimensions: (1) whether benefits are reciprocated directly or indirectly and (2) whether reciprocity occurs through bilateral or reciprocal unilateral flows of benefits.1 These differences in the structure of reciprocity have important implications for the integrative bonds that develop in exchange relations and networks. It is these integrative aspects of exchange on which the research program has focused, four of which I discuss here: trust (the belief that the exchange partner will not exploit the actor), affective regard (positive feelings for, and evaluations of, exchange partners), relational solidarity (perception of the relationship as a single social unit, with actors united in purpose and interests), and interactional fairness (perceptions of fair interpersonal treatment from the partner). The theory that has developed from this program of research, which I will call the reciprocity theory of exchange, argues that the structure of reciprocity affects integrative bonds through three causal mechanisms: the structural risk of nonreciprocity, the expressive value of voluntary acts of reciprocity, and the relative salience of the cooperative or competitive “faces” of exchange (Molm 2003a). While integrative outcomes incorporate both affective and cognitive components, the theoretical program has not explicitly considered the role of emotions in this process. That is the objective of this chapter. Numerous scholars have conceptualized emotions as a key mechanism producing social solidarity and group cohesion. Both the classical functionalists (Bales 1950) and conflict theorists (Coser 1956, 1967) strongly emphasized the role of emotions. Among contemporary theorists, Ridgeway and Johnson (1990)

1 For reports of research from this program, see Molm (2003a,b); Molm, Collett, and Schaefer (2006, 2007); Molm, Peterson, and Takahashi (1999, 2001); Molm, Schaefer, and Collett (2007); Molm, Takahashi, and Peterson (2000, 2003); and Schaefer, Collett, and Molm (2004).

CH10-P374095.indd 168

5/26/2008 2:52:13 PM

Structure of Reciprocity and Integrative Bonds

169

have continued the functionalist concern with the role of emotions in maintaining group cohesion, Collins (1975, 1990, 1993) has integrated Coser’s ideas with Durkheim’s interaction rituals, and Lawler (2001) has developed an affect theory of social exchange that includes emotions as a key factor linking structural features of exchange relationships with cohesion and solidarity. Much of the work on emotions in social exchange relations begins with the structure of power (Lawler and Yoon 1993, 1996).2 I focus, instead, on the structure of reciprocity, and later consider how power may modify the processes involved. The following sections summarize the conceptual background of the research program, the causal model that describes the logic of reciprocity theory, and the research methods and empirical findings from the dozen or so experiments that have been conducted to date. I then turn to a consideration of the role of emotions in the theory. My aim is not to develop a revised theory that integrates emotions in any formal way, but rather to discuss how emotions might be involved in the processes that my colleagues and I have studied, what kinds of theoretical issues are raised by this possibility, and some of the implications for the structure and process of social exchange.

THE THEORETICAL RESEARCH PROGRAM FORMS OF EXCHANGE AND THE STRUCTURE OF RECIPROCITY

All forms of social exchange occur within structures of mutual dependence; that is, structures in which actors are mutually, or reciprocally, dependent on one another for valued outcomes. The mutual or reciprocal dependence can be either direct (A provides value to B and B to A) or indirect (the recipient of the benefit does not return benefit directly to the giver, but to another actor in the social circle). Direct exchanges can be further distinguished by whether they are negotiated or reciprocal (Emerson 1981; Molm 1994). Figure 10.1 illustrates these distinctions in small networks of three or four actors. The research program began by focusing on the two forms of direct exchange. In direct negotiated exchange, two actors jointly negotiate the terms of an agreement that benefits both parties, either equally or unequally (Figure 10.1a). Both sides of the exchange are agreed upon at the same time, and the benefits for both exchange partners are easily identified as paired contributions that form a discrete transaction. In most negotiated exchanges studied by exchange researchers, these agreements are also strictly binding; that is, they automatically produce the benefits agreed upon (Cook et al. 1983; Markovsky, Willer, and Patton 1988).

2 For more discussion of the relation between power and emotions, see Chapter 3 by Thye, Lawler, and Yoon in this book.

CH10-P374095.indd 169

5/26/2008 2:52:13 PM

170

Social Structure and Emotion

(a)

(b)

(c) Exchange networks illustrating three forms of exchange: (a) direct negotiated, (b) direct reciprocal, and (c) indirect chain-generalized.

FIGURE 10.1

In direct reciprocal exchange, actors instead perform individual acts that benefit another, such as giving assistance or advice, without negotiation and without knowing whether or when the other will reciprocate (Figure 10.1b). Relations of reciprocal exchange evolve gradually, as beneficial acts prompt reciprocal benefits, in a series of sequentially contingent, individual acts. Reciprocal exchanges were assumed by most classical exchange theorists (Blau 1964; Homans 1974) but have received less attention from contemporary researchers, with the exception of my own work (Molm 1997; Molm and Wiggins 1979). More recently, the research program was extended to include generalized forms of exchange with indirect reciprocity (Molm, Collett, and Schaefer 2007). The purest form of indirect, generalized exchange, and the one that I have studied, is the chain-generalized form that the classical anthropologists first documented (Lévi-Strauss 1969; Malinowski 1922). In chain-generalized exchange, benefits flow in one direction in a circle of giving that eventually returns benefit to the giver (Figure 10.1c). In a 3-actor chain, for example, A gives to B, B gives to C, and C gives to A; thus, A’s giving to B is reciprocated indirectly, from B through C to A, rather than directly from B to A. Classic examples of chaingeneralized exchange are the Kula ring (Malinowski 1922; Ziegler 1990) and matrilateral cross-cousin marriage (Bearman 1997; Lévi-Strauss 1969). The two structural dimensions of reciprocity – whether benefits are reciprocated directly or indirectly, and whether benefits can flow unilaterally or only bilaterally – cut across these three forms of exchange. The first dimension, direct or indirect reciprocity, corresponds to the basic distinction I have already described between direct (or restricted) forms of exchange, either negotiated or reciprocal, and indirect (or generalized) forms of exchange. Whether reciprocity is direct or indirect implies two additional, and related, structural differences: whether exchange is dyadic or collective and whether actors are dependent on the actions of a single other actor or

CH10-P374095.indd 170

5/26/2008 2:52:13 PM

171

Structure of Reciprocity and Integrative Bonds

TABLE 10.1

The Structure of Reciprocity in Three Forms of Exchange The structure of reciprocity

Form of exchange

Flow of Benefits

Reciprocity

Negotiated exchange

Bilateral

Direct

Reciprocal exchange

Unilateral

Direct

Generalized exchange

Unilateral

Indirect

multiple actors. Direct exchanges are dyadic and each actor is dependent solely on the other; indirect exchanges are collective and each actor is dependent on all other actors in the generalized network for the benefits that are ultimately received from one particular actor. The second dimension, whether benefits can flow unilaterally or only bilaterally, cuts across the first dimension. Although reciprocity is direct in both negotiated and reciprocal exchange, benefits can flow unilaterally in both reciprocal and generalized exchange. In both, each actor’s outcomes are contingent solely on another’s individual actions, and actors can initiate exchanges that are not reciprocated (and vice versa). This means that the timing of reciprocity can be delayed in both reciprocal and generalized exchange (although the greater number of actors in a chain of indirect reciprocity implies greater delay, on average, than in a dyadic relation of direct reciprocity), and in both, any inequality that occurs emerges only over time. In contrast, when exchanges are negotiated, each actor’s outcomes depend on the joint actions of self and other, and the flow of benefits is always bilateral. Each transaction produces an agreement that provides either equal or unequal outcomes for actors, and each transaction is complete in itself; i.e., actors are not embedded in the “chains of mutual obligation and debt” that are characteristic of both direct reciprocal and indirect generalized exchange (Bearman 1997:1413). Theoretically, then, each of these three forms of exchange consists of a different configuration of the two structural dimensions, as shown in Table 10.1. In negotiated exchange, benefits always flow bilaterally and reciprocity is direct; in reciprocal exchange, benefits flow unilaterally but can be reciprocated directly; in generalized exchange, benefits flow unilaterally and reciprocity is indirect. These differences in the structure of reciprocity affect the development of integrative bonds among exchange partners. INTEGRATIVE OUTCOMES OF SOCIAL EXCHANGE

Integrative outcomes refer to behaviors, cognitions, or feelings that result from the structure and process of social exchange and that increase bonds of attachment – both objective and subjective – between actors. The tension between integration and differentiation is an important feature of social exchange

CH10-P374095.indd 171

5/26/2008 2:52:13 PM

172

Social Structure and Emotion

(Blau 1964). The classical exchange theorists devoted fairly equal attention to the integrating and differentiating aspects of exchange and the relations between them (Blau 1964; Homans 1974; Thibaut and Kelley 1959), but until recently, contemporary theorists have focused primarily on processes of differentiation: power, inequality, and conflict. Integrative outcomes capture those aspects of exchange that bring people together and provide the foundation for long-term relationships. In the broader sociological literature, these outcomes are often considered important components of the “social capital” that actors accumulate through their relationships to one another and that provide the basis for acquiring many other kinds of benefits, both economic and social (Coleman 1990; Portes 1998). In this chapter I focus on four key variables that have been studied extensively in this research program and that tap core elements of social integration in exchange relations: trust, affective regard, relational solidarity, and interactional fairness. I define trust as an expectation or belief that an exchange partner will behave benignly toward the actor, based on the attribution of positive traits, dispositions, or intentions to the partner. The latter element of trust distinguishes it from assurance (Yamagishi and Yamagishi 1994), which refers to expectations of benign behavior from an exchange partner based on knowledge of an incentive structure (such as contracts or warranties) that encourages such behavior rather than exploitation. Affective regard refers to an actor’s overall positive or negative feelings toward an exchange partner, combined with assessments of the partner on dimensions with evaluative connotations, such as good or bad. While trust involves attributions of predictability as well as motivation and intent, affective regard is purely evaluative. For both trust and affective regard, the target of evaluation is another actor: the exchange partner in direct exchange relations and an actor’s potential benefactor in indirect exchange relations. In contrast, the target of evaluation for the third integrative outcome, relational solidarity, is the social unit: the dyadic exchange relationship in direct exchange networks and the generalized exchange network in indirect exchange networks. Relational solidarity refers to actors’ perceptions of their relationship or generalized exchange system as a united, harmonious partnership versus a divided, conflictual adversarial relation. The fourth integrative outcome, interactional justice or fairness, refers to the fairness of the interpersonal treatment that an actor receives from an exchange partner (Bies 2001; Cropanzano et al. 2001; Molm, Collett, and Schaefer 2006). The term was originally introduced by scholars of organizational justice to distinguish this aspect of procedural justice – the fairness of interpersonal treatment – from the fairness of formal decision procedures. While organizational scholars tend to focus on aspects of the communication process (such as politeness, honesty, and respect) in their conceptualizations of interpersonal justice (Cohen-Charash and Spector 2001; see also Chapter 11 by Clay-Warner in this book), my colleagues and I have adopted the term solely to denote that the target of the fairness evaluations is the

CH10-P374095.indd 172

5/26/2008 2:52:13 PM

Structure of Reciprocity and Integrative Bonds

173

exchange partner or benefactor, rather than the outcome (as in distributive justice) or the procedures (as in procedural justice). Thus, like trust and affective regard, the target of interactional fairness is the exchange partner in direct exchange relations and the actor’s potential benefactor in indirect exchange relations. Justice judgments have typically been associated with the differentiating aspects of exchange, such as power and inequality, rather than with the development of integrative bonds. But there are several reasons for including fairness in this discussion. First, the perception of fair dealing may well be a prerequisite for the development of integrative bonds. Second, the structure of reciprocity affects perceptions of fairness even when actors are equal in power and their exchanges produce equal benefits. Third, and perhaps most relevant to this book, our research on interactional fairness has important and interesting implications for the role of emotions in exchange relations. THE RESEARCH SETTING

The early work in the program compared the two direct forms of exchange, negotiated and reciprocal, on the three integrative outcomes in which the partner is the target of evaluation: trust, affective regard, and interactional fairness. Later work added the fourth integrative outcome, relational solidarity, and the third form of exchange – chain-generalized exchange with indirect reciprocity. All of the research was conducted in laboratory experiments in which undergraduate students earned money by making exchanges with partners to whom they were connected, via computer, in small networks of three or four actors (see Figure 10.1). In most of the experiments, all actors in the networks were real subjects, but in some – most notably the experiments on fairness – one real actor interacted with two computer-simulated partners. With simulated partners, the objective levels of equality or inequality of benefits could be held constant across different forms of exchange. Some of the experiments testing particular causal mechanisms also used simulated partners, again for their value in physically controlling exchange partners’ behaviors. Three standardized experimental settings were created to allow comparison of the forms of exchange under conditions that held constant all potential differences other than their defining characteristics. Thus, the network structures, the resources exchanged, the potential value of the exchange, and the length of interaction were always equivalent in these comparisons. In the direct negotiated setting, actors negotiated the division of a fixed amount of benefit on each exchange opportunity, which consisted of up to five rounds of negotiation. On each round, all actors in the network simultaneously made offers to each of their potential partners. After the first round, actors could accept another’s offer, repeat their last offer, or make a counteroffer. Negotiations continued until all potential agreements were made or the five rounds were up. As soon as an agreement was reached, both actors received the amounts they had agreed upon.

CH10-P374095.indd 173

5/26/2008 2:52:13 PM

174

Social Structure and Emotion

In the direct reciprocal setting, each actor in the network could give a fixed number of points (equal to one-half the total points that actors in the negotiated setting could divide) to one of his or her exchange partners on each exchange opportunity. They made these choices simultaneously and independently, without knowing whether the other would reciprocate, and then were informed of their exchange partners’ actions. In the chain-generalized setting, each actor could give a fixed number of points to one actor in the network and receive a fixed number of points from a different actor (the number of points given and received was equivalent to that in the reciprocal exchange setting). Because each actor was linked to one potential benefactor and one potential recipient in the closed chain (see Figure 10.1c), actors could choose only whether or not to give points to their potential recipient on each exchange opportunity. As in the reciprocal setting, they made these choices simultaneously and independently, without knowing what choices other actors made, and then were informed of their potential benefactor’s actions. Most experiments compared two or more network structures. The two basic structures that we studied most often were the 3- and 4-actor closed networks illustrated in Figure 10.1. As shown, each actor was connected to two other actors in the network. To ensure that actors obtained benefits from only one exchange (direct or indirect) on each opportunity, relations in the direct forms of exchange were negatively connected: in the negotiated exchanges, making an agreement with one partner precluded an agreement with another partner on that opportunity; in the reciprocal exchanges, initiating exchange with one partner precluded initiating exchange with another on that opportunity. The earlier work comparing the two forms of direct exchange studied power-imbalanced networks, created by varying whether actors’ alternative exchange relations were high or low value. The later work comparing all three forms of exchange studied power-balanced networks, in which all actors had alternatives of equal value. The distinctions between the 3- and 4-actor networks – one of relative power advantage in the earlier research, and one of size in the later research – are not important for most findings and will not be discussed here. The distinction between power-balanced or -imbalanced networks is important, however, both theoretically and empirically. The integrative outcomes were measured by 7-point semantic differential scales administered at the end of the exchange period. Trust was initially measured by a single item and later by 3- or 4-item scales. The single item asked subjects how much they trusted the partner or benefactor (very little/very much); other items asked subjects for their impression of the direct exchange partner or generalized exchange benefactor as untrustworthy/trustworthy, unreliable/reliable, and undependable/dependable. Affective regard was measured by multi-item scales with four or five items. The scale always included one item asking subjects to describe their general feelings toward the partner or benefactor as positive or negative; other items asked subjects for their impression of the

CH10-P374095.indd 174

5/26/2008 2:52:14 PM

Structure of Reciprocity and Integrative Bonds

175

partner or benefactor as awful/nice, bad/good, uncooperative/cooperative, and unhelpful/helpful. Relational solidarity was measured by a 3- or 4-item scale, based on subjects’ descriptions of their relationship with a dyadic exchange partner (direct exchange) or with all of the other participants (indirect exchange) as divided/united, adversaries/partners, and one or more of the following: conflictual/harmonious, self-oriented/team-oriented, and coming apart/coming together. Interactional fairness was measured by a 3-item scale asking subjects to evaluate their partner’s or benefactor’s behavior toward them as unfair/fair, unjust/just, and inequitable/equitable. Cronbach’s alpha for all of these scales is high, ranging from .89 to .96. FORMS OF EXCHANGE AND INTEGRATIVE BONDS

The early research in the program established strong differences in integrative bonds between reciprocal and negotiated exchange, holding constant network structure and partner’s behavior. Experiments comparing the two direct forms of exchange in power-imbalanced networks consistently show that subjects in reciprocal exchanges trust their partners more, express more positive affective regard for them, and perceive their partners as fairer than subjects in negotiated exchanges (Molm, Takahashi, and Peterson 2000, 2003). Values for negotiated exchange typically hover around or just below the neutral point on the semantic differential scales, while the values for reciprocal exchange are typically about 1 point higher on the 7-point scale for trust and affective regard and 1.5 points higher for interactional fairness. Subjects in reciprocal exchanges perceive their partners as fairer than those in negotiated exchanges – and by about the same amount – regardless of whether the exchanges are objectively equal, moderately unequal, or highly unequal. The experiments on interactional fairness have also found marked differences in behavioral responses to objective inequalities in exchange benefits. Subjects engaged in negotiated exchange strongly resist making unequal exchanges that disadvantage them, even when this action is very costly in absolute terms. The frequency of exchange with an unequal partner is only half as great for subjects in negotiated exchanges as for subjects in reciprocal exchanges (Molm, Collett, and Schaefer 2006; Molm, Takahashi, and Peterson 2003). More recent experiments, extending comparisons of the two direct forms of exchange to include indirect, chain-generalized exchange, and extending integrative outcomes to include relational solidarity, have found that generalized exchange produces significantly greater trust, affective regard, and relational solidarity than either form of direct exchange, with differences between direct and indirect exchange averaging about 1 point on the 7-point scale (Molm, Collett, and Schaefer 2007). Replications of the original comparisons of negotiated and reciprocal exchange also show that the differences observed for trust and affective regard extend to relational solidarity; that is, the effects hold regardless of whether the target of evaluation is the partner or the relationship.

CH10-P374095.indd 175

5/26/2008 2:52:14 PM

176

Social Structure and Emotion

THE CAUSAL MODEL

The second phase of the program was devoted to constructing and testing a theoretical model that explains these differences. This phase tested the independent effects of three causal mechanisms proposed as intervening processes that link the structure of reciprocity with integrative bonds: the risk of nonreciprocity, the expressive value conveyed by the act of reciprocity itself, and the relative salience of the cooperative or conflictual aspects of exchange. Predictions in the first phase of the research were based on one or more of these mechanisms, but research in that phase was not designed to distinguish among them – the two forms of exchange differed on all three. The second phase of the research tested the independent effects of each of these mechanisms by varying them within forms of exchange. The first two mechanisms – the risk of nonreciprocity and the expressive value conveyed by the act of reciprocity itself – entail direct structural effects of reciprocity on social solidarity; the third mechanism – the salience of conflict – affects integrative bonds through the intervening cognitions of the actors. Figure 10.2 shows the causal model that resulted from this phase of building and testing theory. Overall, the model proposes that forms of exchange with indirect (rather than direct) reciprocity, and with unilateral (rather than bilateral) flow of benefits, have positive effects on the formation of integrative bonds, producing stronger trust, affective regard, relational solidarity, and interactional fairness. These two structural dimensions of reciprocity increase integrative bonds by increasing the risk of nonreciprocity, increasing the expressive value of the act of reciprocity, and decreasing the relative salience of conflict in the relationship. Risk of Nonreciprocity The risk of nonreciprocity refers to the structural or situational potential for one actor to incur a net loss in exchange by giving benefits to an exchange partner and receiving little or nothing in return. Theoretically, all forms of exchange involve risk because the mutual dependence of actors on one another makes them vulnerable to another’s actions. The amount and kind of risk vary, however, with the structure of reciprocity. The risk of nonreciprocity is present only when exchange benefits flow unilaterally and actors make individual decisions about whether or not to reciprocate another’s giving. When actors jointly negotiate bilateral agreements with binding terms, the risk of nonreciprocity is eliminated. Risk is also greater when reciprocity is indirect rather than direct, because actors are dependent on the actions of multiple others, with risk increasing in proportion to the size of the chain-generalized network. Consequently, generalized exchange is riskier than reciprocal exchange, and reciprocal exchange is riskier than negotiated exchange. Risk has long been linked to the development of trust in relations (Gambetta 1988; Kelley and Thibaut 1978; Kollock 1994; Molm, Takahashi, and Peterson

CH10-P374095.indd 176

5/26/2008 2:52:14 PM

Causal mechanisms

⫹

Risk of nonreciprocity

⫹

Expressive value

Indirect versus direct reciprocity Unilateral versus bilateral flow of benefits

Integrative bonds ⫹ ⫹ ⫺

⫺

⫹ ⫺

Salience of conflict

FIGURE 10.2

Trust

Affective regard Relational solidarity Interactional fairness

Structure of Reciprocity and Integrative Bonds

CH10-P374095.indd 177

Structure of reciprocity

Causal model for the reciprocity theory of social exchange.

177

5/26/2008 2:52:14 PM

178

Social Structure and Emotion

2000). Acts of trust, and attributions of trustworthiness, can only be made in situations in which the partner has both the incentive and the opportunity to exploit the actor, but instead behaves benignly. If, under such risky conditions, the partner behaves in a trustworthy manner, the actor’s trust in the partner should increase. Untrustworthy behavior, on the other hand, should lead to distrust. Thus, risk is a necessary condition for building trust, but not a sufficient one; both risk and behavioral evidence of trustworthiness are required. Without risk, neither trust nor distrust is likely to develop – trust is unnecessary and another’s trustworthiness is impossible to judge. The experimental research in this program supports this logic, showing that the differences in risk between the three forms of exchange affect the development of actors’ trust in their exchange partners. As we would expect, trust is greater, on average, in reciprocal exchange than in negotiated exchange, and the relative trustworthiness of the partner’s behavior is more strongly correlated with an actor’s trust in the partner in reciprocal than in negotiated exchange (Molm, Takahashi, and Peterson 2000). Generalized exchange produces stronger trust than either form of direct exchange, as well as stronger correlations between another’s behavior and an actor’s trust in the other (Molm, Collett, and Schaefer 2007). As the model in Figure 10.2 shows, however, the effects of risk, per se, appear to be specific to trust. This specification comes from experiments that manipulated risk within a single form of exchange, by comparing negotiated exchanges with or without binding agreements (Schaefer, Collett, and Molm 2004). When agreements were nonbinding, actors made individual decisions about whether to honor the agreement, and exploitation was possible. Under these conditions, risk – when combined with high levels of trustworthy behavior – increased trust to levels comparable to those in reciprocal exchanges, but did not increase affective regard or relational solidarity.3 Expressive Value of Reciprocity The second causal mechanism, the expressive value of reciprocity, refers to the symbolic or communicative value attached to the act of reciprocity itself, over and above the instrumental benefits produced by the partner’s reciprocity (Ekeh 1974; Hass and Deseran 1981; Kollock and O’Brien 1992; Lévi-Strauss 1969; Macneil 1986; Offer 1997). While all forms of exchange involve reciprocity, defined as benefits given in return for benefits received, the behavioral act of reciprocity, per se, carries expressive value only when (a) its occurrence is uncertain, with no terms or deadlines stipulated, and (b) it is a voluntary choice of the reciprocator. In other words, the questions of whether, when, and to what extent another’s beneficial action will be reciprocated must be left to the discretion of the recipient (Kranton 1996; Larson 1992; Offer 1997). Under these conditions, the act of reciprocity not only gives benefit in return for benefit received; 3 See Chapter 9 by Gerbasi and Cook in this book for additional research on nonbinding negotiated exchanges and relations among risk, trustworthiness, and affect.

CH10-P374095.indd 178

5/26/2008 2:52:14 PM

Structure of Reciprocity and Integrative Bonds

179

it also communicates regard for the partner and the relationship, acknowledging and conveying appreciation for benefits received and demonstrating a willingness and a desire to continue the relationship (Offer 1997). The structure of reciprocity directly affects the conditions necessary for the act of reciprocity to carry expressive value. Bilateral flows of benefit, in either negotiated exchanges with binding agreements or fixed-price trades, make reciprocity a taken-for-granted feature of exchange. Even when agreements are nonbinding, however, each actor’s decision to honor the agreement meets only one of the requirements just described. Terms have been stipulated and agreed upon; only the decision to carry them out remains. Not keeping an agreement is likely to convey negative expressive value, but keeping an agreement simply meets expectations – it builds trust, but conveys little in the way of positive regard. The unilateral flow of benefits in both reciprocal and generalized exchange, in contrast, meets both requirements: terms are uncertain and reciprocity is voluntary. Actors initiate both reciprocal and generalized exchange without knowing whether, when, or to what extent their giving will be reciprocated in the future. These questions are left to the discretion of the potential reciprocator, and delays in reciprocity are common, often because an opportunity for reciprocation may not immediately occur. Acts of indirect reciprocity carry additional expressive value. In the direct reciprocity of reciprocal exchange, the relationship unfolds as a series of individually performed, sequentially contingent acts, in which obligations are repeatedly created and repaid. But in the indirect reciprocity of generalized exchange, the giver is not repaying a direct obligation. Participation in a generalized exchange network benefits another to whom the giver owes no debt and, at the same time, contributes to a collective enterprise that benefits the network as a whole. Experiments in the research program show that the expressive value of reciprocity, like risk, builds trust, but unlike risk, it also contributes to the development of affective bonds – both affective regard for the partner and feelings of relational solidarity (Molm, Schaefer, and Collett 2007). In addition, it increases perceptions of interactional fairness. These dimensions of relational integration arise from feelings of being valued, respected, and treated well by the other that both unilateral acts of reciprocity and indirect reciprocity promote. Salience of Conflict In all relations of social exchange, actors’ interests partially correspond (both actors are better off with exchange than without it) and partially conflict (each actor benefits in inverse proportion to what he or she gives the other). This duality is inherent in “mixed-motive” structures like exchange. In any particular exchange relation, however, one of these two “faces” of exchange – the cooperative face or the competitive face – may be more salient to the actors than the other. The greater the relative salience of the conflictual, competitive aspects of exchange, the weaker the integrative bonds that develop, as shown by the negative sign of the relation in Figure 10.2. The opposite is true, of course, for the

CH10-P374095.indd 179

5/26/2008 2:52:14 PM

180

Social Structure and Emotion

cooperative aspects of exchange; the greater their salience, the stronger the integrative bonds. But it is the salience of conflict that has the most interesting implications for emotions and on which I focus here. The structure of reciprocity affects the relative salience to actors of the cooperative or competitive aspects of their exchange relation. Bilateral flows of benefit heighten awareness of the competitive, conflictual elements of exchange, while unilateral flows mute their salience. This seems counterintuitive, because negotiated exchanges – in which benefits flow bilaterally – are structurally more cooperative than reciprocal or generalized exchanges, in which benefits flow unilaterally (Molm 1994). Lawler’s (2001) recent affect theory of exchange, which argues that solidarity should be stronger in forms of exchange with greater jointness of task, is based on this difference. Reciprocity theory predicts the opposite relation, arguing that the very features that make negotiated exchange more structurally cooperative than reciprocal exchange – the joint decision-making that creates the bilateral flow of benefits – also bring into sharp relief the inherent conflict of interests between actors in exchange relations. When benefits flow bilaterally, outcomes are more easily compared (the “strict accounting” that Ekeh [1974] associated with restricted exchange is possible); the relation of one actor’s gain to another’s loss is more obvious and direct (making the costs of exchange more transparent); and actions that produce inequality (offering less, bargaining harder) are more explicitly and intentionally directed at the other party. In contrast, the unilateral acts of giving that comprise the structure of reciprocal and generalized exchange mute the inherent conflict in actors’ interests, by making it harder to compare outcomes and by making both the costs of exchange and inequalities in benefits less directly tied to the other actor. The indirect reciprocity of generalized exchange further reduces the salience of conflict, by removing any direct relation between benefactor and recipient. Experiments support the prediction of a strong negative relation between both dimensions of the structure of reciprocity and actors’ beliefs that their relation is conflictual and their partner’s motives are competitive (Molm, Collett, and Schaefer 2006, 2007). Even when power is balanced and exchanges are equal, actors perceive greater conflict when benefits flow bilaterally rather than unilaterally, and when reciprocity is direct rather than indirect (Molm, Collett, and Schaefer 2007). And the greater the salience of conflict, the weaker the development of integrative bonds (Figure 10.2). In the social psychological literature, conflict is most directly linked to distributive and procedural justice (Deutsch 2000; Hegtvedt and Killian 1999; Thompson and Lowensten 1992). Our experimental comparisons of negotiated and reciprocal exchange, however, show that the salience of conflict affects integrative bonds more generally; increased salience of conflict also reduces trust, affective regard, and relational solidarity. Interactions with Structural Power How does power, perhaps the most studied variable in social exchange, affect these relations? Unequal power accentuates the effects of the structure

CH10-P374095.indd 180

5/26/2008 2:52:14 PM

Structure of Reciprocity and Integrative Bonds

181

of reciprocity by heightening the salience of conflict (particularly in negotiated exchange), making reciprocal exchange riskier, and increasing the expressive value of the partner’s reciprocity (Molm, Collett, and Schaefer 2007; Molm, Takahashi, and Peterson 2000). All of these effects are greater for powerdisadvantaged than for power-advantaged actors. Power differences are primarily relevant for direct forms of exchange, in which actors can be equal or unequal in power to one another. In chain-generalized exchange networks, the unidirectional and collective nature of the exchange system makes power differences unlikely: all actors in the chain occupy equivalent structural positions (Bearman 1997; Ekeh 1974). Consequently, our comparisons of all three forms of exchange have been conducted under conditions of equal power, while comparisons of the two direct forms of exchange have been conducted under conditions of either high- or low-power imbalance. As we would expect, under equal structural power, the differences between negotiated and reciprocal exchange are muted; when differences in exchange frequency are controlled for, reciprocal exchange still produces greater relational solidarity than negotiated exchange, but only marginally greater affective regard and no stronger trust (Molm 2006).

BRINGING EMOTIONS INTO RECIPROCITY THEORY This, then, is the outline of the theory to date: forms of exchange in which benefits flow through unilateral acts of reciprocal giving, and in which reciprocity is indirect rather than direct, produce stronger integrative bonds, to both exchange partners and social units, on multiple dimensions. They do so through the joint contributions of three distinct but related causal mechanisms. Unequal power accentuates the effects of the causal mechanisms; equal power mutes them. I turn now to the task of considering the role that emotions might play in this process. I first consider the distinctions between emotions and the affective and cognitive components of integrative outcomes, and then discuss the potential role of emotions in the causal mechanisms that link the structure of reciprocity to integrative outcomes. Finally, I ask what is perhaps the most important question: Do we need emotions to explain the processes in reciprocity theory? Do emotions contribute substantially to our theoretical understanding of the relation between the structure of reciprocity and integrative bonds, or to our explanation of empirical findings? AFFECT, COGNITION, AND EMOTIONS

The integrative outcomes that are the focus of reciprocity theory have both cognitive and affective components; that is, they involve both beliefs about the exchange partner and the relationship and evaluative (positive or negative)

CH10-P374095.indd 181

5/26/2008 2:52:14 PM

182

Social Structure and Emotion

orientations toward those objects – the partner and the relationship. Some scholars have labeled some of these outcomes as emotions; Sprecher (1986), for example, lists trust, liking, and respect as emotions, and Collins (1990:14) calls trust a “long-term emotional disposition.” Kelley (1984) makes no distinction between emotion and affect. As recent reviews of the literature on emotions and affect illustrate, the boundaries between affective orientations and emotions are not always clear, and there continue to be many debates and unanswered questions about the definition of emotions (Lawler and Thye 1999; Smith-Lovin 1995; Stets 2003). I distinguish between emotions and affect, and conceptualize emotions as more fleeting feelings than the integrative bonds that are the focus of reciprocity theory. Both involve cognitive and affective elements, but emotions also include an autonomic or neurological component that may not be entirely under the control of the actor. For my purposes here, two dimensions of emotion are important: (1) their valence (positive or negative) and (2) their intensity (mild to strong). To simplify the discussion, I focus on what Kemper (1987), Turner (2002), and others have labeled primary or basic emotions.4 These include anger, fear, sadness (or depression), and pleasure (or satisfaction or happiness). Pleasure and anger are the two that I will propose are most important for linking the structure of reciprocity to the development of integrative bonds. Although both pleasure and anger are typically classified as primary emotions, some scholars have distinguished between the two on a different basis. Kelley (1984) and Weiner, Russell, and Lerman (1979) suggest that pleasure is a direct emotional reaction to a particular positive event – similar to behavioral accounts of emotions as linked to reinforcement – but that anger is a more derivative emotion that depends on whether an actor attributes a negative event to self or other, with anger occurring when the other is blamed for the negative event. I agree with this distinction – that pleasure is a more direct reaction and anger more derivative – but suggest that anger can be provoked without causal or attributional analysis when negative events occur in a relational context of heightened conflict in which cognitive beliefs about the competing interests of self and other already predispose an actor to blame the other for negative outcomes. THE ROLE OF EMOTIONS IN THE CAUSAL PROCESS

In considering how emotions might be implicated in the causal processes of reciprocity theory, let me begin by sketching the general outlines of the approach I will take. This approach is perhaps most consistent with Kemper’s (1991) structural theory of emotions, although it differs in some key respects. 4 Secondary emotions, which are more socially constructed, may also be involved. One might argue, for example, that feelings of gratitude – a secondary emotion – are sometimes produced when another actor initiates exchange, particularly generalized exchange.

CH10-P374095.indd 182

5/26/2008 2:52:14 PM

Structure of Reciprocity and Integrative Bonds

183

First, like Kemper (and also Lawler 2001), I view emotions as emerging from the structure of relationships and behavioral interaction, not from normative prescriptions, although the structural dimensions on which I focus obviously differ. Kemper proposes that changes in power and status are the primary triggers of emotions; I consider a different set of structural dimensions, those associated with the structure of reciprocity. Kemper’s dimensions of power and status, as he defines these terms, are based on the hierarchy and valence of relationships – dimensions that are roughly equivalent to integration (status) and differentiation (power) in relationships. The reciprocity dimensions on which I focus are based instead on the interdependencies between actors’ exchange behaviors and exchange outcomes. Second, a key component of Kemper’s (1991) theory is actors’ attributional analyses of the source or cause of emotions produced by changes in power or status; the nature of the emotions predicted depends, in most cases, on whether self or other (or fate) is the causal agent. Other theories that take a structural or relational approach to emotions (Lawler 2001; Ridgeway and Johnson 1990) also propose that attributional processes play a key role in determining the nature of emotions, behavioral responses to emotions, or the role of emotions in producing relational solidarity. In contrast, the cognitions that are central to emotions in reciprocity theory are not causal attributions, but rather beliefs about the conflictual or harmonious nature of the exchange relation and the competitive or cooperative motives of the exchange partner. Third, like Kemper, I conceptualize negative emotions as playing a stronger motivational role than positive emotions, as actors seek to change the conditions that cause the negative emotion. Stets (2003) notes that the majority of primary emotions are negative, suggesting that humans may be more attentive to negative than to positive cues. In accord with that view, I will propose that anger, a negative emotion, has stronger effects than pleasure, a positive emotion. Positive emotions tend to reinforce and maintain the status quo, deepening integrative bonds and maintaining established behavioral patterns, while negative emotions can lead to changes in both affect and behavior. While behavior, per se, is not an outcome included in the causal model, the frequency of exchange with a partner may be regarded as both a behavioral indicator of integrative bonds and a prerequisite for their development and/or continuation. Finally, while structural conditions are the fundamental causes of emotions, the three intervening causal mechanisms in the theory are the likely “triggers” for both cognitive and emotional responses. The effects of these mechanisms may depend on exchange outcomes (i.e., the benefits obtained from exchange), but outcomes alone do not determine emotional responses. The first causal mechanism, the risk of nonreciprocity, primarily triggers cognitive processing. Risk provides the situational context for making attributions about the partner’s behavior; specifically, whether the partner is trustworthy or untrustworthy. The relation between risk and trust, I would argue, is mediated

CH10-P374095.indd 183

5/26/2008 2:52:14 PM

184

Social Structure and Emotion

almost entirely by cognition – not emotion.5 Trust has emotional and affective components, but those elements come from the other causal mechanisms. As our experiments have shown, trust can develop without increased affective regard, and – according to subjects’ self-reports of trust – without exchange necessarily producing positive outcomes for the actor. An exchange partner who carries through on his or her stated intentions is trusted more, even when those intentions are contrary to the actor’s interests. That is, a partner who states that he or she intends to initiate reciprocal exchange with someone other than the actor, and then does exactly that, is nevertheless “trusted” by the actor. That kind of behavior is not positive, but neither is it exploitive: it allows the actor to make informed choices that are more beneficial to the actor than some alternatives might be. To produce emotional reactions that affect integrative bonds, the risk of nonreciprocity must be accompanied by the conditions that make the act of reciprocity – in a context of risk – an act of expressive value. That is, there must be an absence of terms stipulating whether, when, or to what extent reciprocity will occur, and reciprocity must be a voluntary act of the reciprocator of benefit. Under these conditions, reciprocity not only provides instrumental value, but expressive value that communicates both the actor’s regard for the exchange partner and the actor’s willingness to invest in the continuation of the exchange relationship. Receiving another’s expression of regard and relationship investment should prompt relatively mild, but certainly positive, emotions of pleasure and satisfaction. There is little need for cognitive processing of the causes of these pleasurable emotions, and they should contribute both to affective regard for the partner (or benefactor) – in reciprocation of the partner’s behavioral expression of regard – and to feelings of relational solidarity. They also add an affective and emotional character to trust and promote the perception of fair dealing by the partner or benefactor. It is the last causal mechanism, the salience of conflict, that has the strongest and most interesting implications for emotions. To the extent that the competitive, conflictual aspects of exchange become salient, actors are more likely to perceive inequalities as unjust and exchange partners as unfair. The relation between conflict and fairness has traditionally been assumed to be mediated by attributional processes (Cohen 1982; Utne and Kidd 1980), with heightened conflict increasing tendencies toward self-serving attributions. That is, the more actors perceive their relation as conflictual, the more likely they are to hold the other responsible for unsatisfactory outcomes and to perceive the other’s behavior as intentional (Hegtvedt and Killian 1999; Thompson and Lowenstein 1992). Attributing responsibility and intent to the other then heightens the severity of felt injustices and increases the likelihood that disadvantaged actors will protest or take action to restore equity. 5

One conference participant suggested that “fear” might be an emotional reaction to risk. I agree that fear, anxiety, and related emotions might all be associated with a context of risk and uncertainty. Fear is unlikely to mediate a positive relation between risk and trust, however.

CH10-P374095.indd 184

5/26/2008 2:52:14 PM

Structure of Reciprocity and Integrative Bonds

185

Experiments in which my collaborators and I manipulated the salience of conflict in direct forms of exchange, however, do not support mediation by attribution (Molm, Collett, and Schaefer 2006). The salience of conflict mediates and explains the relation between the form of exchange and perceptions of both interactional fairness and distributive justice, but conflict affects fairness judgments directly, rather than indirectly through individuals’ tendencies toward self-serving attributions. Actors who are disadvantaged on exchange correctly perceive the distribution of benefits as unequal in their partner’s favor in both negotiated and reciprocal forms of exchange. But perception of inequality in exchange is transformed into a judgment of the partner’s behavioral unfairness only when the salience of conflict is high, as it is in negotiated exchange. Furthermore, high salience of conflict lowers perceptions of fairness even when exchanges are equal. Rather than triggering attributional processes, we propose that conflict produces a direct emotional reaction to the process of exchange, by creating a relational context in which the partner’s behavior is perceived in a negative light, inequalities are more likely to be perceived as unfair, and even equal exchanges are perceived as less than fair (Molm, Collett, and Schaefer 2006). Homans’s (1974) classic analysis of distributive justice proposed that inequitable distributions of benefit make people angry. Indeed they do – but only in relations in which actors perceive the other as an adversary with competing interests rather than as a partner whose interests correspond to their own. These strong negative emotions also motivate behavioral change, as Kemper (1978) and others predict. Subjects in our negotiated exchange conditions, who experienced their relation as highly conflictual and competitive, frequently refused to engage in unequal exchanges, even though the monetary cost of doing so was very high (Molm, Collett, and Schaefer 2006; Molm, Takahashi, and Peterson 2003). Rather than agreeing to an unequal exchange in which their partner earned somewhat more than they did, they opted to make equal exchanges with an alternative partner who controlled resources of much lower value, worth only a fifth as much as they could have received from unequal exchanges with the other partner. In a sense, they were restoring justice in the only way available to them. But they were also behaviorally expressing strongly charged emotions of anger toward an unfair exchange partner. In open-ended responses that our subjects wrote at the conclusion of the experiment, many of those in the negotiated exchange conditions explained their refusal to exchange with the unequal partner, X, as an action taken “to spite X,” to “punish X,” to “make sure that X gets nothing.” They also described X as not only unfair and uncooperative, but “stingy,” “cheap,” and “a complete idiot.” Subjects who had engaged in reciprocal exchange responded very differently to inequality: although our data show that they clearly perceived their objective disadvantage (the inequality in benefits given and received), they evaluated their partners neutrally rather than perceiving them as unfair, and they continued to exchange with them the majority of the time. These subjects used the term “fairness” in their open-ended responses only to explain why they sometimes gave to their low-value (but equal) partner rather

CH10-P374095.indd 185

5/26/2008 2:52:14 PM

186

Social Structure and Emotion

than to their high-value (but unequal) partner: they did so out of “appreciation and fairness.” This account of the relation between inequality and perceptions of fairness is at odds with the strong cognitive flavor of the distributive justice tradition. Although emotions have always been a component of justice theories, far greater emphasis has been placed on reasoned information processing (Jasso 1980; Markovsky 1985). That is, actors obtain outcomes, they compare those outcomes with the outcomes of relevant others or with some comparison standard, they weigh inputs or other factors to determine what is “just,” they evaluate their own outcomes as either exceeding or falling short of this “just reward,” and – if their outcomes do fall short – they then engage in attributional processes to determine the cause of the unjust reward (self, other, fate, etc.). Only then, after determining that their outcomes are unjust and the other is to blame for that injustice, are actors expected to experience the emotion of anger. Scher and Heise (1993) have challenged this sequence of events, arguing that emotional reactions, rather than cognitive processes, lead parties to judgments about fairness. Our research supports that view and further implies that the perception of relational conflict, created by forms of exchange in which reciprocity is structured as direct bilateral flows of benefit, provides a relational context in which unequal exchanges provoke strong, negative emotions, directed at the other party. When the other party is perceived as an adversary whose interests conflict with one’s own, there is no need to engage in attributional processes to assign blame – emotions, not cognitions, play that role. Although the salience of conflict has been discussed primarily in relation to perceptions of fairness, our research shows that its effects spread to all of the other integrative outcomes as well. When actors feel that they are treated unfairly by their exchange partner, integrative bonds are weakened; as Hegtvedt (1990) has found, the perception of fair treatment is a highly salient issue for actors. Actors’ trust in their partner, their affective regard for their partner, and their perception of relational solidarity all drop below the neutral point, into the negative range of evaluation. These effects are consistent with theories that propose that emotional states influence social judgments by triggering cognitions that are similar in valence to the emotions (Bower 1981, 1991; Schwarz 1990; Schwarz and Clore 1983). Thus, anger toward an exchange partner can lead not only to judgments of the partner’s unfairness, but to negative judgments about the partner and the relationship on a number of dimensions. As Lawler and Thye (1999) suggest, these negative emotions and negative cognitions, fueled by repeated, unsatisfactory social exchanges, can build on one another and ultimately bias decision-making. Positive emotions, in contrast, lead to positive evaluations – even in contexts of inequality – and “relaxed” accounting systems that promote cooperation, trust, and solidarity. These parallel processes, both involving emotional states that bias how individuals perceive their exchange partners and relationships, may explain the very different responses of actors in different reciprocity structures to objectively similar, unequal

CH10-P374095.indd 186

5/26/2008 2:52:14 PM

Structure of Reciprocity and Integrative Bonds

187

exchanges: the negative emotions produced by the heightened conflict of bilateral, negotiated agreements produce equally negative cognitive judgments and strong behavioral reactions; the positive emotions produced by reciprocal acts of unilateral giving produce positive judgments and continued (albeit somewhat reduced) exchange – even though the exchanges provide unequal benefits. DO WE NEED EMOTIONS?

As we have seen, emotions can logically be entered as intervening variables in the causal model; this revised model is shown in Figure 10.3. We might still ask, however, how much explanatory power they add to the model. Do emotions contribute substantially to our theoretical understanding of the relation between the structure of reciprocity and integrative bonds? Do emotions explain some findings, some aspects of exchange process or exchange outcomes, that other variables in the model cannot explain? Do any of the causal mechanisms operate solely or primarily through emotions, so that our theoretical account is inaccurate or incomplete without emotions? Let me recap what I have proposed: First, I suggested that emotions play no significant role in mediating the relation between the risk of nonreciprocity and trust. Second, I suggested that mild, positive emotions of pleasure or satisfaction are involved in the positive relation between the expressive regard communicated by voluntary acts of reciprocity and the formation of integrative bonds. Third, I proposed that strong, negative emotions of anger link the salience of conflict to perceptions of interactional unfairness and weakened integrative bonds. In this revised model, emotions are not implicated in the first causal link (between the risk of nonreciprocity and trust). Emotions do contribute to our understanding of the second causal link (between expressive regard and integrative bonds), but are unlikely to be essential to the causal process. The experience of positive emotions may strengthen feelings of affective regard and relational solidarity, but it is likely that the communication of regard would produce reciprocal feelings of regard and perceptions of relational solidarity – the product of the mutual (or generalized) exchange of regard – even without an emotional experience, purely as a result of cognitive inferences. Emotions serve to reinforce the positive experiences with exchange, thus deepening the bond that is developing, but whether they are essential for producing the bond seems more questionable. This position is contrary, however, to both classic and contemporary arguments describing positive emotions as the “glue” that maintains solidarity (Collins 1990; Coser 1956), and it should be tested. It is to our understanding of the third causal link, the negative relation between the salience of conflict and integrative bonds, that an analysis of emotions can potentially contribute the most. Ironically, this causal process has previously been theorized almost exclusively in cognitive, information-processing terms, at least when applied to one integrative outcome – perceived fairness. But when actors experience inequality and disadvantage in a relational context

CH10-P374095.indd 187

5/26/2008 2:52:14 PM

188

CH10-P374095.indd 188

Structure of reciprocity ⫹

Causal mechanisms

Emotions

Risk of nonreciprocity

[None]

Expressive value

⫺ Salience of conflict

FIGURE 10.3

Trust

⫹

⫹

⫺ Pleasure

⫹

Affective regard

⫺

Relational solidarity

Anger

Revised causal model, incorporating the potential role of emotions.

Interactional fairness

5/26/2008 2:52:14 PM

Social Structure and Emotion

Unilateral versus bilateral flow of benefits

⫹ ⫹

Indirect versus direct reciprocity ⫹

Integrative bonds

Structure of Reciprocity and Integrative Bonds

189

of heightened conflict, they react far more strongly than we would predict from cognition alone. Actors who exchange under these conditions are angry. They evaluate their partners and their relationships in highly negative terms, and they behave in ways that are not in their self-interest, at least as that interest is typically conceptualized in the framework of exchange theory. Without bringing in emotions, the effect of emotions on cognitive evaluations of partners and relationships, and the motivational force of negative emotions on behavior, our understanding of these effects may well be incomplete.

CONCLUSIONS The structure of reciprocity comprises an important set of dimensions that affect the process and outcomes of exchange relations. The strong and consistent relations that have been empirically observed between dimensions of reciprocity and the integrative bonds that exchange partners develop may well be mediated, in part, by emotions that serve to strengthen or weaken social solidarity. Here, I have considered the potential role that emotions might play in this causal process, proposing that two emotions – mild feelings of pleasure or satisfaction and strong feelings of anger – may be implicated in some of the causal mechanisms previously supported for the theory. This discussion of emotions is only a preliminary step. If emotions are to be integrated in reciprocity theory, then the next step must be empirical tests of the intervening role of emotions proposed here. That could easily be done with additional experiments in which emotions are measured at one or more stages during the exchange process and their indirect role in the causal process is tested statistically. If the effects of structural dimensions of reciprocity affect integrative bonds only indirectly, through emotions, then the theory should clearly be revised to reflect this specification. If emotions are, instead, outcomes of reciprocity, equivalent in theoretical status to the more enduring affective states that have characterized our study of integrative bonds, then their inclusion in the theory is more optional. In either case, such work would help to clarify long-standing questions about the importance of emotion in linking reciprocity in social life with dimensions of social solidarity.

ACKNOWLEDGMENTS This research was supported by two grants from the National Science Foundation (#SBR-9514911 and #SES-0217287). The author gratefully acknowledges their support and the work of the collaborators on this project: Jessica Collett, Gretchen Peterson, David Schaefer, and Nobuyuki Takahashi. The author also thanks the editors of this book and participants at the Social Structure and Emotions Conference for helpful comments on an earlier draft of this chapter.

CH10-P374095.indd 189

5/26/2008 2:52:14 PM

190

Social Structure and Emotion

REFERENCES Bales, Robert F. 1950. Interaction Process Analysis: A Method for the Study of Small Groups. Cambridge, MA: Addison-Wesley. Bearman, Peter. 1997. “Generalized Exchange.” American Journal of Sociology 102:1383–1415. Becker, Howard. 1956. Man in Reciprocity. New York: Praeger. Bies, Robert S. 2001. “Interactional (In)Justice: The Sacred and the Profane.” In: Jerald Greenberg and Russell Cropanzano (eds), Advances in Organizational Justice, pp. 89–118. Stanford, CA: Stanford University Press. Blau, Peter. 1964. Exchange and Power in Social Life. New York: Wiley. Bower, Gordon H. 1981. “Mood and Memory.” American Psychologist 36:129–148. Bower, Gordon H. 1991. “Mood Congruity of Social Judgments.” In: Joseph P. Forgas (ed.), Emotion and Social Judgments, pp. 31–53. Oxford: Pergamon. Cohen, Ronald L. 1982. “Perceiving Justice: An Attributional Perspective.” In: Jerald Greenberg and Ronald L. Cohen (eds), Equity and Justice in Social Behavior, pp. 119–160. New York: Academic. Cohen-Charash, Yochi and Paul E. Spector. 2001. “The Role of Justice in Organizations: A MetaAnalysis.” Organizational Behavior and Human Decision Processes 86:278–321. Coleman, James S. 1990. Foundations of Social Theory. Cambridge: Harvard University Press. Collins, Randall. 1975. Conflict Sociology. New York: Academic. Collins, Randall. 1990. “Stratification, Emotional Energy, and the Transient Emotions.” In: Theodore D. Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 27–57. New York: State University of New York Press. Collins, Randall. 1993. “Emotional Energy as the Common Denominator of Rational Action.” Rationality and Society 5:203–230. Cook, Karen S., Richard M. Emerson, Mary R. Gillmore, and Toshio Yamagishi. 1983. “The Distribution of Power in Exchange Networks: Theory and Experimental Results.” American Journal of Sociology 89:275–305. Coser, Lewis. 1956. The Functions of Social Conflict. New York: Free Press. Coser, Lewis. 1967. Continuities in the Study of Conflict. New York: Free Press. Cropanzano, Russell, Zinta S. Byrne, D. Ramona Bobocel, and Deborah E. Rupp. 2001. “Moral Virtues, Fairness Heuristics, Social Entities, and Other Denizens of Organizational Justice.” Journal of Vocational Behavior 58:164–209. Deutsch, Morton. 2000. “Justice and Conflict.” In: Morton Deutsch and Peter T. Coleman (eds), The Handbook of Conflict Resolution: Theory and Practice, pp. 41–64. San Francisco: Jossey-Bass. Ekeh, Peter P. 1974. Social Exchange Theory: The Two Traditions. Cambridge, MA: Harvard University Press. Emerson, Richard M. 1981. “Social Exchange Theory.” In: Morris Rosenberg and Ralph H. Turner (eds), Social Psychology: Sociological Perpsectives, pp. 30–65. New York: Basic Books. Gambetta, Diego. 1988. “Can We Trust Trust?” In: Diego Gambetta (ed.), Trust: Making and Breaking Cooperative Relations, pp. 213–237. New York: Basil Blackwell. Hass, David F. and Forrest A. Deseran. 1981. “Trust and Symbolic Exchange.” Social Psychology Quarterly 44:3–13. Hegtvedt, Karen A. 1990. “The Effects of Relationship Structure on Emotional Responses to Inequity.” Social Psychology Quarterly 53:214–228. Hegtvedt, Karen A. and Caitlin Killian. 1999. “Fairness and Emotions: Reactions to the Process and Outcomes of Negotiations.” Social Forces 78:269–302. Hobhouse, Leonard T. 1906. Morals in Evolution: A Study in Comparative Ethics. London: Chapman and Hall. Homans, George C. 1974. Social Behavior: Its Elementary Forms. Revised edition. NY: Harcourt, Brace & World. Jasso, Guillermina. 1980. “A New Theory of Distributive Justice.” American Sociological Review 45:3–32.

CH10-P374095.indd 190

5/26/2008 2:52:15 PM

Structure of Reciprocity and Integrative Bonds

191

Kelley, Harold H. 1984. “Affect in Interpersonal Relations.” In: Phillip Shaver (ed.), Review of Personality and Social Psychology, Vol. 5, pp. 89–115. Newbury Park, CA: Sage. Kelley, Harold H. and John W. Thibaut. 1978. Interpersonal Relations: A Theory of Interdependence. NY: Wiley. Kemper, Theodore D. 1978. A Social Interactional Theory of Emotions. New York: Wiley. Kemper, Theodore D. 1987. “How Many Emotions Are There? Wedding the Social and the Autonomic Components.” American Journal of Sociology 93:263–289. Kemper, Theodore D. 1991. “Predicting Emotions from Social Relations.” Social Psychology Quarterly 54:330–342. Kollock, Peter. 1994. “The Emergence of Exchange Structures: An Experimental Study of Uncertainty, Commitment, and Trust.” American Journal of Sociology 100:313–345. Kollock, Peter and Jodi O’Brien. 1992. “The Social Construction of Exchange.” In: Edward J. Lawler, Barry Markovsky, Cecilia Ridgeway, and Henry A. Walker (eds), Advances in Group Processes, Vol. 9, pp. 89–112. Greenwich, CT: JAI Press. Kranton, Rachel E. 1996. “Reciprocal Exchange: A Self-Sustaining System.” The American Economic Review 86:830–851. Larson, Andrea. 1992. “Network Dyads in Entrepreneurial Settings: A Study of the Governance of Exchange Relationships.” Administrative Science Quarterly 37:76–104. Lawler, Edward J. 2001. “An Affect Theory of Social Exchange.” American Journal of Sociology 107:321–352. Lawler, Edward J. and Shane R. Thye. 1999. “Bringing Emotions into Social Exchange Theory.” Annual Review of Sociology 25:217–244. Lawler, Edward J. and Jeongkoo Yoon. 1993. “Power and the Emergence of Commitment Behavior in Negotiated Exchange.” American Sociological Review 58:465–481. Lawler, Edward J. and Jeongkoo Yoon. 1996. “Commitment in Exchange Relations: Test of a Theory of Relational Cohesion.” American Sociological Review 61:89–108. Lévi-Strauss, Claude. 1969. The Elementary Structures of Kinship, Revised edition. Boston: Beacon. Macneil, Ian R. 1986. “Exchange Revisited: Individual Utility and Social Solidarity.” Ethics 96:567–593. Malinowski, Bronislaw. 1922. Argonauts of the Western Pacific. New York: E. P. Dutton. Markovsky, Barry. 1985. “Toward a Multilevel Distributive Justice Theory.” American Sociological Review 50:822–839. Markovsky, Barry, David Willer, and Travis Patton. 1988. “Power Relations in Exchange Networks.” American Sociological Review 53:220–236. Molm, Linda D. 1994. “Dependence and Risk: Transforming the Structure of Social Exchange.” Social Psychology Quarterly 57:163–176. Molm, Linda D. 1997. Coercive Power in Social Exchange. Cambridge University Press. Molm, Linda D. 2003a. “Theoretical Comparisons of Forms of Exchange.” Sociological Theory 21:1–17. Molm, Linda D. 2003b. “Power, Trust, and Fairness: Comparisons of Negotiated and Reciprocal Exchange.” In: Shane R. Thye and John Skvoretz (eds), Advances in Group Processes: Power and Status Organizing Processes, Vol. 20, pp. 31–65. Oxford, UK: Elsevier Press. Molm, Linda D., Jessica L. Collett, and David R. Schaefer. 2006. “Conflict and Fairness in Social Exchange.” Social Forces 84:2331–2352. Molm, Linda D., Gretchen Peterson, and Nobuyuki Takahashi. 1999. “Power in Negotiated and Reciprocal Exchange.” American Sociological Review 64:876–889. Molm, Linda D., Gretchen Peterson, and Nobuyuki Takahashi. 2001. “The Value of Exchange.” Social Forces 80:159–185. Molm, Linda D., David R. Schaefer, and Jessica L. Collett. 2007. “The Value of Reciprocity.” Social Psychology Quarterly 70:199–217. Molm, Linda D., Nobuyuki Takahashi, and Gretchen Peterson. 2000. “Risk and Trust in Social Exchange: An Experimental Test of a Classical Proposition.” American Journal of Sociology 105:1396–1427.

CH10-P374095.indd 191

5/26/2008 2:52:15 PM

192

Social Structure and Emotion

Molm, Linda D., Nobuyuki Takahashi, and Gretchen Peterson. 2003. “In the Eye of the Beholder: Procedural Justice in Social Exchange.” American Sociological Review 68:128–152. Molm, Linda D. and James A. Wiggins. 1979. “A Behavioral Analysis of the Dynamics of Social Exchange in the Dyad.” Social Forces 57:1157–1179. Offer, Avner. 1997. “Between the Gift and the Market: The Economy of Regard.” Economic History Review 3:450–476. Portes, Alejandro. 1998. “Social Capital: Its Origins and Applications in Modern Sociology.” Annual Review of Sociology 22:1–24. Ridgeway, Cecilia and Cathryn Johnson. 1990. “What Is the Relationship Between Socioemotional Behavior and Status in Task Groups?” American Journal of Sociology 95:1189–1212. Schaefer, David A., Jessica L. Collett, and Linda D. Molm. 2004. “Risk, Trust, and Solidarity: Explaining Effects of the Form of Exchange.” American Sociological Association Meetings, San Francisco, CA. Scher, Steven J. and David R. Heise. 1993. “Affect and the Perception of Injustice.” In: Edward J. Lawler, Barry Markovsky, Karen Heimer, and Jodi O’Brien (eds), Advances in Group Processes, Vol. 10, pp. 223–252. CT: JAI. Schwarz, Rorbert. 1990. “Feelings as Information: Informational and Motivational Functions of Affective States.” In: R. M. Sorrentino and E. T. Higgins (eds), Handbook of Motivation and Cognition: Foundations of Social Behavior, pp. 265–290. New York: Guilford. Schwarz, Rorbert. and Gerald L. Clore. 1983. “Mood, Misattribution, and Judgments of Well-Being: Informative and Directive Functions of Affective States.” Journal of Personality and Social Psychology 45:513–523. Simmel, Georg. 1950. The Sociology of Georg Simmel, translated and edited by Kurt Wolff. Glencoe, IL: Free Press. Smith-Lovin, Lynn. 1995. “The Sociology of Affect and Emotion.” In: Karen S. Cook, Gary Alan Fine, and James S. House (eds), Sociological Perspectives on Social Psychology, pp. 118–148. Boston: Allyn and Bacon. Sprecher, Susan. 1986. “The Relationship Between Inequity and Emotions in Close Relationships.” Social Psychology Quarterly 49:309–321. Stets, Jan E. 2003. “Emotions and Sentiments.” In: John Delamater (ed.), Handbook of Social Psychology, pp. 309–335. New York: Kluwer Academic/Plenum. Thibaut, John W. and Harold H. Kelley. 1959. The Social Psychology of Groups. New York: Wiley. Thompson, Leigh and George Lowenstein. 1992. “Egocentric Interpretations of Fairness and Interpersonal Conflict.” Organizational Behavior and Human Decision Processes 51:176–197. Turner, Jonathan H. 2002. Face-to-Face: Towards a Sociological Theory of Interpersonal Behavior. Stanford: Stanford University Press. Utne, Mary Kristine and Robert F. Kidd. 1980. “Equity and Attributions.” In: Gerald Mikula (ed.), Justice and Social Interaction, pp. 63–93. New York: Springer-Verlag. Weiner, Bernard, Dan Russell, and David Lerman. 1979. “The Cognition–Emotion Process in Achievement-related Contexts.” Journal of Personality and Social Psychology 37:1211–1220. Yamagishi, Toshio and Midori Yamagishi. 1994. “Trust and Commitment in the United States and Japan.” Motivation and Emotion 18:129–166. Ziegler, Rolf. 1990. “The Kula: Social Order, Barter, and Ceremonial Exchange.” In: Michael Hechter, Karl-Dieter Opp, and Reinhard Wippler (eds), Social Institutions: Their Emergence, Maintenance, and Effects, pp. 141–170. New York: Aldine de Gruyter.

CH10-P374095.indd 192

5/26/2008 2:52:15 PM

PA R T

IV Justice and Moral Emotions

CH11-P374095.indd 193

5/26/2008 2:53:37 PM

11 Commentary Jody Clay-Warner University of Georgia, Athens, GA, USA

Emotions have always been central to the study of justice. Recently, however, researchers have begun to ask questions that move the study of justice and emotion beyond the traditional equity model and its focus on outcome distributions. Some of these questions are a result of a move toward examining variations in discrete emotions, while other of these questions are derived from newer models of justice that challenge some of the basic assumptions of the traditional equity model. Here I trace the development of the study of justice and emotions and discuss the contributions made by the two chapters in this section.

EMOTIONS AND THE EQUITY MODEL The most widely utilized formulation of micro-level justice is the equity model. In this model, inequity is said to exist when an actor’s inputs are not equivalent to his or her outcomes, or when the ratio of the actor’s inputs to outcomes is not equivalent to those of a comparison with other or exchange partner (Adams 1965; Blau [1964](1986); Homans 1974; see also Berger, Cohen, and Zelditch 1972). Though the equity model has been presented in slightly different ways by different scholars, a shared assumption is that inequity results in emotional distress. Empirical research largely supports the model’s predictions with respect to under-reward, with research finding that under-reward is associated with strong, negative emotional reactions. In a classic study of injustice and naturally occurring emotions, Mikula (1986) asked subjects to describe a situation in which they had been treated unfairly and then to recall their feelings. Anger emerged as the primary emotional reaction (see also Clayton 1992). In a more recent Social Structure and Emotion

CH11-P374095.indd 195

195

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 2:53:38 PM

196

Social Structure and Emotion

study of naturally occurring emotions, Mikula, Scherer, and Athenstaedt (1998) confirmed the link between anger and injustice in a large multi-national sample of university students. Using vignette methodology, Hegtvedt (1990) found that under-reward produced resentment. Other work in this area has focused on inequity in interpersonal relationships. It is generally found that those who perceive themselves to be under-rewarded in a relationship report higher levels of depression and anger than those who are equitably rewarded (e.g. Kluijer, Buunk, Ybema, and Wobbes 2002; Longmore and Demaris 1997; Sprecher 1986, 1992). The equity model’s predictions regarding emotions and over-reward, however, have not found support. Homans (1974) predicted that inequity of any kind would result in emotional distress and that in the case of over-reward the distress would take the form of guilt. He acknowledged, however, that guilt would most likely emerge when the over-reward was at the expense of another and that low levels of over-reward may not evoke strong emotional reactions. Empirical research has failed to establish a consistent link between over-reward and guilt or other negative emotions. Austin and Walster (1974) found that over-rewarded subjects reported slightly more distress than equitably treated subjects but less distress than under-rewarded subjects. Gray-Little and Teddlie (1978), however, found no differences between these groups. In Hegtvedt’s (1990) vignette study, over-rewarded actors did report slightly higher levels of guilt than did equitably rewarded and under-rewarded actors. The intensity of the guilt, though, was very low, leading Hegtvedt to suggest that arguments about emotional reactions to inequity most centrally apply to situations of under-reward. Over-reward has been found to be related to guilt and other negative emotions in personal relationships, however (Sprecher 1992, 1986). There is also evidence to suggest that over-reward may actually lead to positive affect, which is contrary to the equity model. In a recent study of consumer behavior, Oliver, Shor, and Tidd (2004) reported that subjects who received an unexpected coupon that lowered their purchase price felt more satisfied about their transaction than did those in the control group. Weiss, Suckow, and Cropanzano (1999) reported similar results from their laboratory study, finding that the most important predictor of happiness was outcome favorability and that there was not a significant difference between the reported levels of happiness of subjects in the fair versus the over-rewarded condition.

PROCEDURAL JUSTICE AND EMOTIONS While the focus of the equity model is on outcome fairness (distributive justice), scholars working in this tradition have also utilized the basic assumptions of the equity model to explain why individuals value fair procedures. Thibaut and Walker (1975) argued that people desired fair procedures because fair procedures were likely to lead to fair distributions. Thus, procedures had instrumental value. Thibaut and Walker’s research focused primarily upon preferences for particular procedural rules, as well as attitudinal and behavioral responses to fair

CH11-P374095.indd 196

5/26/2008 2:53:38 PM

197

Commentary

versus unfair procedures. Emotions were not an explicit focus of their model. There was, however, an implicit link between procedural justice and emotions. Specifically, fair procedures were expected to increase positive outcomes which, in turn, bring about positive emotions. Emotions have recently become more central to the study of procedural justice as theories have moved away from instrumental models and toward relational models. Lind and Tyler’s (1988) and Tyler and Lind’s (1992) group-value model asserts that people desire fair procedures not only because fair procedures lead to fair outcomes but, more importantly, because procedurally fair treatment by in-group authorities is an indication of high status within one’s group. As a result, people value fair procedures in and of themselves and not simply for their instrumental effects on distributions. Building upon Leventhal, Karuza, and Fry (1980), Tyler and Lind (1992) define fair procedures as those that reflect standing, neutrality, and trustworthiness (see also Tyler 1989). Standing is demonstrated though respectful treatment, while neutrality refers to bias suppression and honesty. Authorities communicate trust by considering the opinions of those in the group and by showing concern for their well-being. Empirical research has confirmed these predictions, with researchers finding not only that people interpret procedurally fair treatment as an indication that they are well regarded by other group members, but that these effects exist above and beyond the effects of distributive justice (e.g. Tyler, Degoey, and Smith 1996). According to the group-value model, fair procedures evoke positive emotions because fair procedures indicate favorable standing in the group. Alternatively, perceiving that one is a low status member of the group would be associated with negative emotions. To test this prediction, Tyler (1994) measured emotional responses to procedural and distributive justice at work among a random sample of Chicago residents. He found that procedural justice was an important predictor of emotional reactions to these workplace events and that the effects were independent of the effects of distributive justice. Other researchers have confirmed these findings. In a survey of workplace behaviors, Fox, Spector and Miles (2001) found that procedural justice was negatively correlated with a 17-item scale of negative affect, while Hegtvedt and Killian (1999) found in their experimental study of negotiated exchange that participants’ ratings of procedural justice were linked to their reported levels of depression and negativity (see also Barclay, Skarlicki, and Pugh 2005; Krehbiel and Cropanzano 2000). In their meta-analysis, Cohen-Charash and Spector (2001) also report that procedural injustice is related to negative affect.

CURRENT DIRECTIONS Though researchers have tested the basic propositions regarding emotions offered by the dominant theories of justice, we know far less about emotional reactions to justice/injustice than we do about attitudinal and behavioral responses (Cropanzano et al., 2000). As a result, the link between justice and

CH11-P374095.indd 197

5/26/2008 2:53:38 PM

198

Social Structure and Emotion

emotions is under-theorized, and many questions remain unanswered (Weiss, Suckow, and Cropanzano 1999). In the last few years, however, justice scholars have increasingly turned their attention to the study of emotions, creating a promising surge of new research. These researchers have focused our attention on a number of new or under-studied areas, including the impact of contextual factors on emotional reactions to injustice, moral emotions, the importance of examining discrete emotions, and third-party reactions to injustice. Chapters 12 and 13 reflect many of the new directions in the study of justice and emotions. In discussing these chapters by Hegtvedt et al. and Stets et al., I focus upon these new directions and highlight additional lines of research suggested by their research. In Chapter 12, Hegtvedt, Johnson, and Ganem explore emotional reactions to observed distributive injustice, thereby moving justice research beyond its traditional focus on the self. Using a vignette methodology, they examine how contextual factors affect observer’s emotional responses to an unfair outcome experienced by an in-group member. They are particularly interested in the ways in which contextual factors alter the observer’s felt and expressed emotions, with a focus on procedural justice, endorsement, and strength of interpersonal ties. They find that procedural justice does increase both felt and expressed satisfaction concerning an unjust distribution. The results for negative emotions are less straightforward, with one of the most interesting findings being an interaction between endorsement and procedural justice: it is when the decision-maker has support that procedural justice produces the greatest degree of resentment and frustration. Their results highlight the importance of examining multiple contextual factors when studying emotional reactions to injustice. Doing so more accurately simulates real-world environments in which outcome decisions take place and also advances our understanding of the intertwined social processes that produce reactions to injustice. Hegtvedt et al.’s research on third-party reactions to injustice is unique. Not only has empirical research on injustice focused almost exclusively on injustice to self, but research that has considered third-party reactions to injustice has not explored the ways in which such reactions affect emotions. This is an interesting omission in the literature since the group-value model of justice asserts that people desire not only to be respected members of their group, but that they also want to feel proud of their group. A sense of “group pride” is derived from knowing that one is a member of a group that enacts fair procedures. ClayWarner (2001) expanded upon the concept of group pride to explain why high status authorities are unlikely to perceive procedurally unjust treatment of ingroup members. Clay-Warner (2001), however, did not examine emotional reactions to perceived injustice. Similarly, other research has examined how overall perceptions fairness in the group is affected by varying levels of injustice toward self and others but, again, this research has not considered emotional reactions to such injustice (e.g. Kray and Lind, 2002; Lind, Kray, and Thompson, 1998). Thus, Hegtvedt et al. take an important step in recognizing that not only are

CH11-P374095.indd 198

5/26/2008 2:53:38 PM

Commentary

199

people attuned to the procedural injustice experienced by other group member but that under certain conditions observing another’s injustice can affect one’s own emotions. Stets et al. (Chapter 13) also offer an empirical chapter that moves the field beyond its current parameters. These authors explore moral emotions through the lens of identity theory. Though moral emotions tend to be complex, at their heart is a sense of injustice. Moral emotions occur in response to violation of moral codes – codes to which all members of society are equally bound. When individuals violate such codes, they receive benefits that non-violators do not, and often the benefits result in losses for others. Inequity is thus the basis of the shame, guilt, anger, and empathy that occur in response to such violation of the moral code. Stets et al. are interested in the ways in which one’s moral identity and status affect moral emotion. They predict that when an individual with a high moral identity behaves in non-normative ways, moral emotions will result (e.g., guilt over failure to notify a clerk of being undercharged). While their results confirm that the moral identity does influence moral emotions, they find that the effects are not simply a result of a discrepancy between the identity and behavior. Instead, there are main effects for the moral identity on certain moral emotions, while normative action serves to mediate the relationship for other moral emotions. Skitka, Bauman, and Mullen (forthcoming) argue that the recent surge in research on justice and morality signals a shift in emphasis from homo socialis (humans as social beings who have an innate need to belong) to homo moralitis (humans as moral beings who care about basic issues of right and wrong). They argue that this shift has brought justice researchers a new set of questions that challenge our assumptions about human nature. In focusing upon moral emotions, Stets et al.’s study brings homo moralitis to the fore. Instead of arguing that morality is an innate concern, however, Stets and colleagues view morality in identity terms – people are more or less invested in maintaining “moral” selves. Thus, concern with morality is variable across individuals and becomes part of the overall control system that regulates emotion and behavior. Stets and colleagues also argue that status affects moral emotions, which suggests that social structure and identity work together to maintain the normative order. This linking of the micro- and macro-level is an important step in the study of moral emotions, where most research has focused exclusively on the individual. Together, these chapters increase our understanding of justice and emotions and also suggest new questions. Are concerns for the justice of others reflective of morality or self-interest? When does the expression of moral emotions serve to maintain the status quo versus challenge the status quo? And, when and how do people manage the moral emotions of other group members? Answering such questions brings sociology closer to understanding why people continue to accept and legitimate injustice. As Hegtvedt et al. and Stets et al. remind us, people feel uncomfortable when exposed to unjust situations. These feelings of distress sometimes motivate people to rectify injustice. Yet, oftentimes, distress – or

CH11-P374095.indd 199

5/26/2008 2:53:38 PM

200

Social Structure and Emotion

even moral outrage – does not evoke behaviors designed to change the situations. Emotions, of course, are complex, quickly changing and are, themselves, affected by social structural factors, such as power and status. Thus, continued study is necessary to determine the role of emotions in system change.

REFERENCES Adams, J. Stacy. 1965. “Inequity in Social Exchange.” Advances in Experimental Social Psychology 2:267–299. Austin, W. and Elaine Walster. 1974. “Reactions to Confirmation and Disconfirmation of Equity and Inequity.” Journal of Personality and Social Psychology 30:208–216. Barclay, Laurie J., Daniel P. Skarlicki, and S. Douglas Pugh. 2005. “Exploring the Role of Emotions in Injustice Perceptions and Retaliation.” Journal of Applied Psychology 90:629–643. Blau, Peter. [1964] 1986. Exchange and Power. New York: Wiley. Berger, Joseph, Bernard P. Cohen, and Morris Zelditch, Jr.. 1972. “Status Characteristics and Social Interaction.” American Sociological Review 37:241–255. Clayton, Susan. 1992. “The Experience of Injustice: Some Characteristics and Correlates.” Social Justice Research 5:71–92. Clay-Warner, Jody. 2001. “Perceiving Procedural Injustice: The Effects of Group Membership and Status.” Social Psychology Quarterly 64:224–238. Cohen-Charash, Yochi and Paul E. Spector. 2001. “The Role of Justice in Organizations: A MetaAnalysis.” Organizational Behavior and Human Decision Processes 86:278–321. Cropanzano, Russell, Howard M. Weiss, and Kathleen J. Suckow. 2000. “Doing Justice to Workplace Emotion.” In: Neal M. Ashkanasy, Charmine E. Härtel, and Wilfred J. Zerbe (eds.), Emotions in the Workplace: Research, Theory, and Practice, pp. 49–62. Westport, CT: Quorum. Fox, Suzy, Paul E. Spector, and Don Miles. 2001. “Counterproductive Work Behavior (CWB) in Response to Job Stressors and Organizational Justice: Some Mediator and Moderator Tests for Autonomy and Emotions.” Journal of Vocational Behavior 59:291–309. Gray-Little, Bernadette and Charles B. Teddlie. 1978. “Racial Differences in Children’s Responses to Inequity.” Journal of Applied Social Psychology 8:107–116. Hegtvedt, Karen A. 1990. “The Effects of Relationship Structure on Emotional Responses to Inequity.” Social Psychology Quarterly 53:214–228. Hegtvedt, Karen A. and Caitlin Killian. 1999. “Fairness and Emotions: Reactions to the Process and Outcomes of Negotiation.” Social Forces 78:269–303. Homans, George C. 1974. Social Behavior: Its Elementary Forms. Hartcourt Brace Jovanovich. Kluijer, Roeline G., Bram P. Buunk, Jan Y. Ybema, and Theo Wobbes. 2002. “The Relations between Perceived Inequity, Marital Satisfaction and Emotions among Couples Facing Cancer.” The British Journal of Social Psychology 41:39–56. Kray, Laura E. and E. Alan Lind. 2002. “The Injustices of Others: Social Reports and the Integration of Others’ Experiences in Organizational Justice Judgments.” Organizational and Human Decision Processes 89:906–924. Krehbiel, Patricia J. and Russell Cropanzano. 2000. “Procedural Justice, Outcome Favorability, and Emotion.” Social Justice Research 13:339–360. Leventhal, Gerald S., J. Karuza, Jr, and W. R. Fry. 1980. “Beyond Fairness: A Theory of Allocation Preferences.” In: Gerold Mikula (ed.), Justice and Social Interaction, pp. 167–218. New York: Springer-Verlag. Lind, E. Alan, Laura Kray, and Leigh Thompson. 1998. “The Social Construction of Injustice: Fairness Judgments in Response to Own and Others Unfair Treatment by Authorities.” Organizational Behavior and Human Decision Processes 75:1–22.

CH11-P374095.indd 200

5/26/2008 2:53:38 PM

Commentary

201

Lind, E. Alan and Tom R. Tyler. 1988. The Social Psychology of Procedural Justice. New York: Plenum. Longmore, Monica A. and Alfred Demaris. 1997. “Perceived Inequity and Depression in Intimate Relationships: The Moderating Effect of Self-esteem.” Social Psychology Quarterly 60:172–184. Mikula, Gerold, Klaus R. Scherer, and Ursula Athenstaedt. 1998. “The Role of Injustice in the Elicitation of Differential Emotional Reactions.” Personality and Social Psychology Bulletin 7:769–783. Mikula, Gerold. 1986. “The Experience of Injustice: Toward a Better Understanding of Its Phenomenology.” In: H.W. Bierhoff, R.L. Cohen, and J. Greenberg (eds.), Justice in Social Relations, pp. 103–124. New York: Plenum. Morgan, Rick and David R. Heise. 1988. “Structure of Emotions.” Social Psychology Quarterly 51:19–31. Oliver, Richard L., Mikhael Shor, and Simon T. Tidd. 2004. “Induced Over-Benefiting and UnderBenefiting on the Web: Inequity Effects on Feelings and Motivations with Implications for Consumption Behavior.” Motivation and Emotion 28:85–106. Skitka, Linda J., Christopher J. Bauman, and Elizabeth Mullen. (forthcoming). “Morality and justice: An Expanded Theoretical Perspective and Empirical Review.” Advances in Group Processes, 25. Sprecher, Susan. 1986. “The Relation between Inequity and Emotions in Close Relationships.” Social Psychology Quarterly 49:309–321. Sprecher, Susan. 1992. “How Men and Women Expect to Feel and Behave in Response to Inequity in Close Relations.” Social Psychology Quarterly 55:57–69. Thibaut, John and Laurens Walker. 1975. Procedural Justice: A Psychological Analysis. Hillsdale, NJ: Erlbaum. Tyler, Tom R. 1989. “The Psychology of Procedural Justice: A Test of the Group-Value Model.” Journal of Personality and Social Psychology 57:830–838. Tyler, Tom R. 1994. “Psychological Models of the Justice Motive: Antecedents of Distributive and Procedural Justice.” Journal of Personality and Social Psychology 5:850–863. Tyler, Tom R., Peter Degoey, and Heather J. Smith. 1996. “Understanding Why the Justice of Group Procedures Matter.” Journal of Personality and Social Psychology 70:913–930. Tyler, Tom R. and E. Alan Lind. 1992. “A Relational Model of Authority in Groups.” Advances in Experimental Social Psychology 25:115–191. Weiss, Howard W., Kathleen Suckow, and Russell Cropanzano. 1999. “Effects of Justice Conditions on Discrete Emotions.” Journal of Applied Psychology 84:786–794.

CH11-P374095.indd 201

5/26/2008 2:53:38 PM

12 It’s Not Just What You Feel: Expressing Emotional Responses to the Injustice of Others Karen A. Hegtvedt1, Cathryn Johnson1 and Natasha M. Ganem2 1 Emory University, Atlanta, GA, USA Savannah College of Art and Design, Savannah, GA, USA

2

ABSTRACT Although many studies address emotional reactions to injustice, few analyze how observers respond emotionally to the injustices suffered by others. Here we examine the influence of procedural justice, interpersonal ties, and legitimacy of the group decision-maker on observers’ emotional experiences and expressions. Hypotheses suggest that procedurally fair treatment attenuates negative emotional reactions and may enhance satisfaction. Interpersonal ties and legitimacy are likely to impact emotional responses when procedures are unfair. Student study volunteers (N  173) read a vignette describing a distributively unjust situation in a volunteer organization and answered a series of questions. Analyses confirm most hypotheses pertaining to procedural justice and, to some extent, those regarding legitimacy. Observers’ emotional experiences and expressions are most intense when a legitimate decision-maker uses unfair procedures. Procedural justice effects are strongest under conditions of high legitimacy. Discussion focuses on the implications of emotional responses for ensuring justice for others in both volunteer and paid organizational situations. Key Words: Justice, Emotion, Legitimacy

Social Structure and Emotion

CH12-P374095.indd 203

203

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 2:54:56 PM

204

Social Structure and Emotion

INTRODUCTION Whites fight for the civil rights of minority Americans. Men support women’s endeavors for equality in the workplace. Heterosexuals defend provisions for “domestic partners’” benefits in organizations. What is common to all of these examples is that people who are advantaged by the social system act on behalf of individuals who are disadvantaged by the system to redress inequality. Or, more simply, the advantaged seek justice for and with (disadvantaged) others. The study of justice in a wide range of fields – psychology, social psychology, sociology, political science, and so forth (see Scherer 1992) – has largely focused on antecedents and consequences of injustice that befalls oneself and results in disadvantage or “under-reward.” Such studies examine social factors affecting perceptions of distributive, procedural, and interactional injustice and accompanying emotional, cognitive, or behavioral responses (see e.g., Colquitt et al. 2001; Hegtvedt and Cook 2001; Tyler et al. 1997). Recent research, however, has begun to draw attention to how observers or “third parties” respond to the injustice suffered by another person (e.g., Kray and Lind 2002; Lind, Kray, and Thompson 1998; Lupfer et al. 2000; Skarlicki, Ellard, and Kelln 1998; van den Bos and Lind 2001). To extend this novel area of research, this study focuses explicitly on observers’ emotional experiences and expressions. Many studies show that when people directly experience disadvantageous outcome injustice they feel anger and resentment (e.g., Hegtvedt 1990; Hegtvedt and Killian 1999; Sprecher 1992). Negative feelings of injustice, however, may not be expressed. Johnson, Ford, and Kaufman (2000) show that individuals are likely to suppress the display of negative feelings in response to an unfair pay raise when they are in a highly dependent position or when their superior is strongly supported by other co-workers. In other words, situational factors attenuate the display of felt emotions by those who directly experience injustice. No research to date has examined what observers feel in response to another’s injustice. The emotional experiences of observers may enhance motivation to rectify an unfair situation. Whether situational factors attenuate observers’ emotional expressions also remains to be investigated. In general, emotion management (e.g., Hochschild 1983; Erickson and Ritter 2001) suggests that individuals follow rules about how they should feel in a situation, but other expectations associated with a situation may modify how they feel or the extent to which they express their true feelings. For example, following from Johnson et al. (2000), subordinates may feel angry about failing to receive a deserved pay raise, but they hide that anger around their boss for fear of losing their jobs. Building upon Johnson et al., here we examine how the social context affects both the experience and expression of emotions in response to the distributive injustice suffered by another person. Specifically, we examine how procedural justice, interpersonal ties, and the legitimacy of the decision-maker affect how observers respond emotionally to

CH12-P374095.indd 204

5/26/2008 2:54:56 PM

Expressing Emotional Responses to Injustice of Others

205

a colleague’s injustice.1 These factors may shape emotional experiences, which may or may not be expressed. Although much research indicates that procedural justice (i.e., fairness of the decision-making process) reduces the perceived severity of distributive injustice (see Tyler et al. 1997), few studies focus on how it may also reduce the negativity of emotional responses. In contrast, interpersonal ties between group members may intensify the emotional response, especially the experience of it. Also, the legitimacy of the decision-maker draws attention to the role of other group members in a manner different from that of interpersonal ties. If peers of the observer legitimate an authority by their support, perceptions accompanying the expression of negative emotions grow more complicated. These three contextual factors, thus, have differential implications for observers’ experience and expression of emotional responses to another group member’s unfair outcome. Concentration on observers’ emotions is a unique contribution of this study. By focusing on both experienced and expressed emotions, we may tap into processes that perpetuate inequality. In any distribution situation, if individuals other than the recipients themselves provide no visible indication of their distress over injustice within the situation, authorities have no reason to alter their decisions or allocations. In other words, emotion management disallows assessment of not only individual sentiment, but collective sentiment as well. The contrast between experienced and expressed emotions in conjunction with the group level contextual factors provides a basis for a clearer understanding of the reciprocal influences of individuals embedded in a social system. Such an understanding augments current theorizing about justice processes and addresses Simon and Nath’s (2004) call to consider the social context of emotions. Below we first review elementary propositions regarding emotional response to distributive injustice. Then we introduce tenets of approaches to emotional management. We integrate both justice and emotions theorizing in our specification of hypotheses about the effects of procedural justice, interpersonal ties, and legitimacy on observers’ emotional responses to injustice. We use data from a vignette study to test our hypotheses.

1

As most models of justice suggest (e.g., Tyler et al. 1997), perceptions of injustice may mediate between situational conditions and emotional, cognitive, and behavioral responses to injustice. In another paper using data from this vignette study (Morgan et al. 2005) we examine the direct effect of procedural justice, interpersonal ties, and legitimacy of the authority on justice perceptions, and their indirect effects – through justice perceptions – on anticipated behavioral responses. Given that analysis, we simplify our analysis here by focusing only on the effects of the context on experienced and expressed emotional responses. Such a strategy is akin to previous research (e.g., Hegtvedt 1990; Johnson et al. 2000) and is consistent with those who argue that the emotional response may precede the cognitive evaluation of fairness (Scher and Heise 1993). Our data, however, do not allow us to adjudicate whether the emotional response or the cognitive assessment comes first.

CH12-P374095.indd 205

5/26/2008 2:54:56 PM

206

Social Structure and Emotion

EMOTIONAL RESPONSES TO DISTRIBUTIVE INJUSTICE A key proposition of classic models of distributive justice is that a person’s experience of injustice creates distress (Adams 1965; Homans 1974; Walster, Walster, and Berscheid 1978). Feelings of distress emanate from unmet expectations about levels of actual outcomes compared to expected outcomes based on past experience or generalized standards or from inequality ratios of outcomes to inputs between individuals. Homans (1974) specifically argues that disadvantageous injustice stimulates feelings of anger whereas advantageous injustice may result in feelings of guilt, if advantage is at another person’s expense. The experience of justice, in contrast, produces feelings of satisfaction. While less explicitly discussed, procedural justice researchers also presume that negative emotions results from procedural injustice (see Tyler et al. 1997). Such distress, in turn, presumably stimulates other cognitive or behavioral reactions to the injustice, much like balance or consistency theories would predict (Heider 1958). A number of studies confirm the predictions of the classic models of distributive justice regarding the experience of feelings. Early work on emotional responses using global measures of distress and occasionally employing physiological measures demonstrates higher levels of distress among those suffering injustice (see Hegtvedt and Markovsky 1995). Later work focuses on comparisons of the distinct emotions experienced by those whose outcomes represent disadvantaged or advantaged injustice or fair outcomes. Research on both personal (e.g., Sprecher 1992) and impersonal (e.g., Hegtvedt 1990) relationships shows far less satisfaction and greater anger among those suffering disadvantageous injustice than those who are fairly rewarded. Results are more equivocal regarding feelings of guilt among those receiving higher than expected outcomes. As Homans (1974) noted, unless the over-benefit is at another’s expense in an interdependent relationship (Sprecher 1992), guilt levels rarely vary systematically across reward levels or vary at such low levels to be relatively meaningless (Hegtvedt 1990). Of the studies directly pertaining to observers’ perceptions of the fairness of outcomes or procedures for other group members, none focus on emotions. Two other studies, however, indirectly bear upon observers’ emotions. Hegtvedt and Killian (1999) examine emotional responses to perceptions of procedural and distributive justice in an experimental bargaining situation. Their results confirm the typical pattern of negative feelings in response to procedural and distributive injustices pertaining to one’s own situation and also show that such negative responses are not as evident with regard to injustices to others. Subjects who judge the outcome to their partners as unfair, however, report higher levels of guilt. Such a finding is consistent with Homans’ theorizing about when individuals are likely to feel guilty in response to over-reward. In an allocation study, Peters, van den Bos, and Bobocel (2004) show that observers are likely to believe that others are more satisfied with over-reward than they themselves are.

CH12-P374095.indd 206

5/26/2008 2:54:56 PM

Expressing Emotional Responses to Injustice of Others

207

This “moral superiority” of observers may account for the generally low levels of guilt experienced by those who are overpaid. By focusing only on projections about how satisfied other actors might be with their outcomes, the study does not directly address observers’ actual emotional responses to another’s injustice. Thus, while these two distinct studies make inroads regarding the study of emotional responses to injustice, specific extensions regarding what observers feel and how they express those feelings about unfair outcomes received by others remain warranted.

EMOTION MANAGEMENT Most studies of emotional reactions to one’s own injustice presume emotions to be “a relatively short-term evaluative response essentially positive or negative in nature involving distinct somatic (and often cognitive) components” (Kemper 1978, p. 47). In effect, an emotion is something that is felt in response to situational stimuli, in this case, an objectively unfair outcome. Situational circumstances not only shape the feelings individuals might experience but they may also affect the extent to which individuals might express those feelings. Hochschild (1979) introduces feeling rules to represent cultural standards pertaining to what people ought to feel – in terms of the type, intensity, and duration of an emotion – in a particular situation. When individuals feel something different than that expected on the basis of feeling rules, they may engage in emotional management to create or change in degree a more appropriate feeling. Research examines the strategies people use to bring their feelings in line with the appropriate feeling rules and the consequences of such strategies for wellbeing, especially in work situations (e.g., Erickson and Ritter 2001; Hochschild 1983; Smith and Keinman 1989). In addition, display or expression rules, defined as the often unarticulated rules of what one should express in social interaction, circumscribe the expression of emotions, given situational conditions. For example, it may be inappropriate for subordinates to express anger in response to a superior’s behavior, because to do so may threaten their job security; superiors, in contrast, may have no such constraints on their displays of anger. As Erickson and Ritter (2001) note, emotional management strategies may, “… require workers to mask, hide, or suppress emotions they feel in order to create a suitable emotional display” (p. 150). Several studies in work organizations show that workers do suppress their expression of emotions but the extent to which they do so depends upon their status or position as well as the nature of the emotion (Gibson 1997; Pugliesi and Shook 1997). Using responses to a vignette describing an unfair pay situation, Johnson et al. (2000) study the extent to which actors’ relative power positions and the legitimacy of an authority affects the likelihood of the experience and expression of satisfaction, anger, and resentment in response to denial for a pay raise. Results show that individuals report the highest levels of feelings of

CH12-P374095.indd 207

5/26/2008 2:54:56 PM

208

Social Structure and Emotion

anger and resentment when they are in more power advantaged positions. Also, if subordinates’ peers support the authority, they are less likely to report feelings of resentment. Likewise, subordinates who are relatively advantaged in terms of their alternatives and whose peers do not support the superior, are more likely to express anger and resentment. Thus situational conditions shape both feelings and displays of emotions – sometimes differentially, depending upon the emotion. Here we extend the work of Johnson et al. (2000) by examining how situational conditions affect the emotional reactions of observers to another person’s unjust outcome.

THE EFFECTS OF SOCIAL FACTORS ON OBSERVERS’ EXPERIENCE AND EXPRESSION OF EMOTIONAL RESPONSES TO ANOTHER’S INJUSTICE Undoubtedly, the emotional response to injustice is more potent when an individual directly experiences the injustice. But work from field studies of layoff survivors’ responses to the experiences of layoff victims (e.g., Brockner et al. 1994) and experimental studies on observers’ reactions to an authority who unfairly treats their peers (e.g., Kray and Lind 2002; van den Bos and Lind 2001) show that observers often do respond emotionally to another’s injustice. Situational circumstances, however, may enhance or diminish these emotional responses which, in turn, may spur their behavioral reactions. To the extent that observers experience an emotional response akin to those of the sufferer of an injustice, they may be more likely to act on the victim’s behalf. Here we examine how procedural justice, interpersonal ties, and legitimacy of an authority influence the positive and negative emotional responses to another person’s unfair outcome. We circumscribe our hypotheses to the following conditions. First, the situation pertains to interaction that occurs in an hierarchical organization. Second, the focal instance involves a distributive disadvantage, which may be described objectively as unfair, to a group member other than the observer. Third, the disadvantage has no direct effect on the observer’s own standing or rewards in the group. This condition should mitigate the likelihood of fears that the injustice may also befall the observer. Fourth, the observer is embedded in the group and thus would expect continuing relationships with other group members, both peers and authorities. And fifth, the distributive disadvantage created by the authority is the first instance of such behavior on the authority’s part. Because of the focus on distributive disadvantage, we examine three negative emotional responses: anger, resentment, and frustration. Although these emotions reflect on the outcomes received by the observed party (see Johnson et al. 2000), they are somewhat different. Anger represents strong feelings of displeasure. Resentment captures displeasure tinged with indignation. And, frustration

CH12-P374095.indd 208

5/26/2008 2:54:56 PM

Expressing Emotional Responses to Injustice of Others

209

pertains to the displeasure resulting from the thwarting of valued goals. Kemper (1987) argues that anger is a primary emotion – instinctive and physiologically distinct. In contrast, resentment and frustration are secondary emotions, differentiated by social and cultural factors and without distinct physiological properties. Affect control theory ratings of these emotions (Morgan and Heise 1988) set them apart on three dimensions: evaluation (positive or negative), potency (or powerfulness), and implied activity level (lively or quiet). The three emotions are equally negative but anger is far more potent and active than the other two. Frustration and resentment involve similarly low levels of activity but the potency rating distinguishes the two; frustration typifies a more powerless affective state than resentment. We also examine satisfaction, a common emotional response in justice situations. Generally, the experience of satisfaction in response to another’s disadvantage is likely to be low. SHAPING OBSERVERS’ EMOTIONAL EXPERIENCES

The Effects of Procedural Justice Procedural justice refers to fairness in the rules underlying decision-making or in the treatment of group members. Leventhal, Karuza, and Fry (1980) suggest that fair decision-making rules involve consistency, the suppression of bias and arbitrariness, the accuracy of information, the potential to make corrections for bad decisions, and the representativeness or “voice” of individuals affected by the decision. Tyler and Lind (1992) further indicate that procedurally fair treatment includes demonstration of respect, neutrality, and trust. Most studies of perceptions of another’s injustice focus on procedural injustice and draw upon the group value model (Lind and Tyler 1988). The group value model, stemming from social identity theory (e.g., Tajfel and Turner 1979), argues that individuals want to be well regarded within the groups to which they belong. Fair procedures and treatment signal such regard and as a consequence increase self-esteem. In effect, just procedures highlight social elements of justice processes rather than material ones, which have been the main focus of distributive justice research. Such a highlight is likely to draw attention to the treatment of others and their outcomes. People use information about decision-making procedures or treatment procedures to make assessments of the fairness of outcomes. Labeled the “fair process effect” (van den Bos, Lind, and Wilke 2001), many studies show that procedural justice mitigates the unfairness of a perceiver’s outcomes (see Tyler et al. 1997). Procedural justice also directly affects satisfaction with one’s own outcomes (see Cohen-Charash and Spector 2001). While little work addresses the fairness of procedures and observers’ responses to another’s injustice, the importance of procedures makes it likely that such information may play a role. Procedures directly linked to the outcome may be particularly salient for observers owing to their proximity and implications for all group members. In addition, the fair treatment of others reflects on the group as a whole and may have implications for the development of observers’ pride in their group (Tyler, Degoey, and Smith 1996).

CH12-P374095.indd 209

5/26/2008 2:54:56 PM

210

Social Structure and Emotion

Even if an outcome appears to be objectively unfair, to the extent that fair procedures were used, an observer may feel that the authority upheld the integrity of the group and, by implication, the standing of members of the group. Even though an unfair outcome is likely to be dissatisfying, fair procedures may mitigate that dissatisfaction. Thus: Hypothesis 1: Procedural justice is positively related to an observer’s experience of satisfaction in response to the distributive injustice suffered by another person.

Insofar as procedural justice mitigates the perceived unfairness of one’s own outcomes, it may have a similar effect on the outcomes of others. As such, the negative emotional experience associated with unfair outcomes may dissipate. Indeed, Krehbiel and Cropanzano (2000) show that fair procedures reduced feelings of anger over unfavorable outcomes to oneself. Again, the extent to which fair procedures maintain the integrity of the group may play a role in decreasing negative affect both with regard to own outcomes and others’ outcomes as well. Thus: Hypothesis 2: Procedural justice is negatively related to an observer’s experience of anger, resentment, and frustration in response to the distributive injustice suffered by another person.

When procedures are fair, other situational factors may be of less importance. Given the situation of disadvantage, other factors are unlikely to affect feelings of satisfaction at all. The impact of interpersonal ties and legitimacy of the decision-maker may, however, become more pronounced when procedures leading to the decision are unfair. The Effects of Interpersonal Ties Interpersonal ties pertain to the closeness of group members to each other at a personal level and extend beyond the bounds of particular structural positions (O’Reilly and Chatman 1986). Wenzel (2001) suggests that individuals’ affective connections to fellow group members influence their justice perceptions. In the case of third-party observers, how they feel about other group members is likely to be particularly relevant to how they respond to another’s injustice. Individuals who are close to victims of injustice are more likely to identify with their mistreatment or poor outcomes (Brockner and Greenberg 1990; Skarlicki and Kulik 2005; Turner 1987). Indeed, Huo (2003) specifically notes that, with strong friendships, observers are more likely to empathize with the group member who has received a poor outcome. Such empathy is likely to raise concerns for another’s injustice (Kray and Lind 2002) and to shape the way in which observers respond emotionally to the situation. When group members have strong interpersonal ties, the unfair outcome for a single person may stimulate feelings in observers similar to those experienced by the injustice victim. They would be more likely to experience anger, resentment, and frustration on behalf of their unjustly treated colleague. This effect, however, is more likely to occur when the decision-maker uses unfair procedures because

CH12-P374095.indd 210

5/26/2008 2:54:56 PM

Expressing Emotional Responses to Injustice of Others

211

such actions to create the outcome are inappropriate and unexpected and may in the long run threaten the group. Such implicit threats may reinforce negative feelings about the other person’s unfair outcome. In contrast, when procedures are fair, the observer may presume that the decision-maker acted in good faith despite the negative outcome. Such a presumption would dampen any resulting negative affect. Thus: Hypothesis 3:

When the situation is procedurally just, interpersonal ties will have little impact on an observer’s experience of negative affect. When the situation is procedurally unjust, interpersonal ties is positively related to an observer’s experience of anger, resentment, and frustration in response to the distributive injustice suffered by another person.

The Effects of Legitimacy Complementing affective interpersonal ties, legitimacy pertains to the desire of individuals to uphold the “order” of the group. More specifically, legitimacy involves people’s recognition that they must obey the operating rules of the group, regardless of whether or not they personally believe them to be appropriate (Weber [1924] 1978). Zelditch and Walker (1984; 2000) argue that the views of others in the group enhance the legitimacy of the rules. They identify two collective sources of legitimacy: (1) authorization exists when support for the rules comes from people who occupy higher positions within a group and (2) endorsement exists when support comes from individuals of equal or lower status than the focal person. The anticipation of formal sanctions from authorities or informal sanctions from peers stimulates compliance. Studies show that actors are less likely to attempt to challenge an unfair distribution or structure when it is authorized or endorsed (e.g., Johnson and Ford 1996; Thomas, Walker and Zelditch 1986; Walker, Rogers, and Zelditch 1988; Walker, Thomas, and Zelditch 1986). Little work, however, has examined the effects of legitimacy on justice perceptions or emotional reactions. Following arguments by Hegtvedt and Johnson (2000) that the legitimacy of an authority extends to his or her behaviors and provides the basis for the benefit of the doubt should the behaviors create an undesirable outcome, Mueller and Landsman (2004) show that endorsement and authorization have positive effects on actors’ perceptions of the fairness of their rewards. These effects, however, are indirect, through perceptions of procedural justice. And, as noted above, Johnson et al. (2000) demonstrate that endorsement tends to diminish feelings of resentment over an unfair pay raise, most likely because of the benefit of the doubt created when one’s peers support the decision-maker. The focus on an observer’s emotional reactions to another person’s injustice warrants emphasis on endorsement because the observer and the unjustly treated colleague are situated as peers in the group. Endorsement also generally has stronger effects than authorization on justice perceptions involving groups of colleagues (Johnson et al. 2000; Mueller and Landsman 2004). Endorsement indicates what other group members think of the decision-maker in general and,

CH12-P374095.indd 211

5/26/2008 2:54:56 PM

212

Social Structure and Emotion

by extension, his or her behaviors. Individuals are likely to expect the authority to uphold explicit and implicit rules of the group regarding procedures and treatment of group members. When a decision-maker distributes outcomes using fair procedures, observers may be less likely to respond negatively to the unfair outcome received by a colleague. Fair procedures are likely to be expected. Moreover, by using fair procedures, the authority is behaving in a way that earns the trust of group members and may have initially won the support of subordinates in the group. The use of fair procedures and having support from others may provide a significant buffer against rising negative emotions from subordinates. Both provide a source of stability for the decision-maker that may prevent a negative reaction. Therefore, levels of negative affect are likely to be the lowest when the decision-maker uses fair procedures and is endorsed. In contrast, when a decision-maker uses unfair procedures, observers are likely to wonder why such unexpected behavior occurred (van den Bos et al. 1999). The procedural injustice is inconsistent with existing expectations about the authority and thus may stimulate further analysis of the situation. Observers may readily imagine other possible procedures that the decision-maker could have used. Such referent cognitions (Folger 1986) may reinforce the connection between the unfair procedure and the unfair outcome and also induce a backlash against the authority because, in effect, he or she has violated group members’ trust. Endorsement thus exacerbates observers’ responses to the unfair outcome, produced by unfair procedures. A supported decision-maker who uses unfair procedures creates an even more surprising event – one that is inconsistent with expectations, especially those regarding trust in the authority. The combination of unfair procedures and endorsement intensifies the observers’ negative emotions, creating the highest level of negative affect. Ironically, lack of endorsement coupled with the use of unfair procedures may be less surprising and the inconsistency less worrisome because expectations of trust are weaker. As a result, the observer should still feel negative emotions but less intensely than when an endorsed authority uses unfair procedures. Hypothesis 4:

When the situation is procedurally just, endorsement will have little impact on an observer’s experience of negative affect. When the situation is procedurally unjust, endorsement is positively related to an observer’s experience of anger, resentment, and frustration in response to the distributive injustice suffered by another person.

SHAPING OBSERVERS’ EMOTIONAL EXPRESSIONS

Even if observers experience emotions in response to another person’s unfair outcomes, they may not display those feelings. Many work situations discourage the expression of negative emotions in particular (Heise and Calhan 1995). And to the extent that situational circumstances affect the social costs associated with the expression of emotions, individuals are less likely to let others know how they

CH12-P374095.indd 212

5/26/2008 2:54:56 PM

Expressing Emotional Responses to Injustice of Others

213

feel. Thus, although emotional experiences should be highly related to emotional expressions, they may not fully mediate between situational factors and emotional expressions. Rather, select elements of the social context may exert direct effects on what observers reveal about their feelings regarding the distributive injustice suffered by another. We expect that procedural justice and endorsement will have such effects. In the long run, procedural justice ensures fair outcomes (Lind and Tyler 1988). To the extent that an authority uses fair procedures, an observer may be reluctant to express negative emotions regarding another person’s unfair outcome because in the long run fair procedures will produce fair outcomes. To reveal anger, resentment, or frustration to a decision-maker who has acted in good faith may undermine one’s own position in the group. In such a case, the observer may incur social costs for expressing negative affect. In contrast, when the decision-maker has used unfair procedures, the expression of negative emotions may drive home the point that such procedures fail to ensure the long-term standing of the group and its members. Thus: Hypothesis 5: Procedural justice is negatively related to an observer’s expression of anger, resentment, and frustration in response to the distributive injustice suffered by another person.

Given that procedural justice is likely to affect feelings of satisfaction, it may also promote the expression of that emotion. Again, insofar as procedural justice ensures the integrity of the group and the perceived standing of its members, individuals may feel free to express positive emotions about the behavior of the authority, even if he or she has produced an unfair outcome for a group member. Observers may recognize that conveying their satisfaction about the situation generally (even when the outcome is unfair) is a means to reinforce the use of fair procedures in the future. In the long run, that would promote social benefits of group membership. Observers would be less likely to want to reinforce the use of unfair procedures because to do so might engender social costs in the long run. Under conditions involving unfair procedures, they would be likely to suppress their feelings of satisfaction. Thus: Hypothesis 6: Procedural justice is positively related to an observer’s expression of satisfaction in response to the distributive injustice suffered by another person.

In addition, we argue that when fair procedures are used, endorsement will have little effect on an observer’s expression of anger, resentment, and frustration because such procedures reinforce or stimulate reasons for endorsement. When unfair procedures are used, however, endorsement should affect expression of negative emotions. Johnson et al. (2000) show that to the extent that a decision-maker is supported by subordinates, individuals may find it difficult to express negative emotions for fear of informal sanctions or social costs from peers. Thus people are likely to suppress expressing negative emotions about their own outcomes when the decision-maker is endorsed but not when he or she lacks endorsement.

CH12-P374095.indd 213

5/26/2008 2:54:56 PM

214

Social Structure and Emotion

Observers, however, may feel less constrained by the decision-maker’s legitimacy. Expressing negative emotions may be less costly for observers than for a victim. Moreover, they may perceive that an endorsed decision-maker who uses unfair procedures has violated the expectations people hold for him or her and they may presume that others share that perception. These lower costs, combined with perceptions that the violation of expectations is shared, may have an effect opposite to that found by Johnson et al. (2000). Such conditions may empower observers to speak out on behalf of the victim, to express their anger, resentment, and frustration. And again as in the case of the experience of these negative emotions, a decision-maker who uses unfair procedures and lacks endorsement may simply confirm expectations and lead to less expression of negative affect. Therefore, we expect that the likelihood of expressing negative emotions should be the least when fair procedures are used and the decision-maker is endorsed, and the greatest when unfair procedures are used and the decision-maker is endorsed. Thus: Hypothesis 7:

When the situation is procedurally just, endorsement will have little impact on an observer’s expression of negative affect. When the situation is procedurally unjust, endorsement is positively related to an observer’s expression of anger, resentment, and frustration in response to the distributive injustice suffered by another person.

METHODS OVERVIEW

To test our hypotheses, we created a vignette describing a distributively unjust situation. The vignette involved a decision-maker (Joanne) in a volunteer organization (National Coalition to End Cancer) who does not choose a fully qualified applicant (Marie) to be on a prestigious fundraising committee. Subjects imagined themselves as NCEC volunteers like Marie. Using a 2  2  2 design, we manipulated the procedural justice (high versus low) of the committee members’ selection process, interpersonal ties (strong versus weak) among volunteers, and endorsement (high versus low) of the superior. We expected that these factors would directly or as combinations impact subjects’ emotional reactions to this unfair situation. We employed analysis of variance (ANOVA) to assess the vignette manipulations. We used OLS, including sex of subject as a control, to test our hypotheses. SUBJECTS

Subjects were recruited from introductory sociology classes at a large southeastern university. They were told the purpose of the study was to investigate organizational dynamics. They were asked to spend 15–20 minutes reading a

CH12-P374095.indd 214

5/26/2008 2:54:56 PM

Expressing Emotional Responses to Injustice of Others

215

one page story and answering questions that followed. Subjects were assured that participation was strictly voluntary and was not tied to their grade for the course. One-hundred and sixty students (56 male, 104 female) completed one of the eight randomly distributed vignettes. Cell sizes ranged from 18–26. THE VIGNETTE AND QUESTIONNAIRE

Manipulations in the Vignette Procedural Justice Procedural justice was manipulated through the contents of an email sent to Marie from Joanne (the decision-maker). In the email, Joanne explained how she chose members for the fundraising committee. Her procedures varied in terms of consistency, bias suppression, and informational accuracy (Leventhal et al. 1980). In the high procedural justice condition, Joanne wrote “I know that Directors in the past have followed NCEC guidelines, so I did too. I read all of the applications thoroughly. I created a checklist so I could accurately compare the applicants. Ultimately, I based my decision on this information, not what I already knew about the applicants. I hope you try again next year.” In contrast, in the low procedural justice condition, Joanne wrote “I know that Directors in the past have followed NCEC guidelines. This year, however, I did something different. I read some applications thoroughly and skimmed others. I created a checklist so I could compare the applicants but, ultimately, I based my decision on what I already knew about them. You can try again next year.” Interpersonal Ties We manipulated interpersonal ties by focusing on elements of the situation that might engender affective connections between group members. In the strong interpersonal ties condition, the vignette informed the subject that “You really like your fellow volunteers. You’ve developed some strong friendships, and you often get together after NCEC functions. You know everyone has the same goals and values, and it always feels like you’re working as a team.” In contrast, in the weak interpersonal ties condition the vignette read “You don’t really like the other volunteers. You haven’t developed strong friendships with anyone, and you rarely get together after NCEC functions. You sometimes wonder if everyone has the same goals and values, because it doesn’t always feel like you’re working as a team.” Endorsement According to Johnson and Ford (1996), a person is endorsed if others believe that he or she is the right (or appropriate) person for the job and support him or her in the role. Following this conceptualization, subjects in the high endorsement condition read “The other volunteers think Joanne is the right person for the job and fully support her as the Director of Fundraising.” In contrast, low endorsement

CH12-P374095.indd 215

5/26/2008 2:54:56 PM

216

Social Structure and Emotion

was conveyed by “The other volunteers question whether Joanne is the right person for the job and don’t fully support her as Director of Fundraising.” Dependent Variable Measures in the Questionnaire Emotional Experience To assess the extent to which subjects felt satisfaction, anger, resentment, and frustration after hearing that Marie was not selected for the committee, we asked subjects to indicate on a scale ranging from 1 (not at all) to 9 (a great deal) how much they felt these emotions. Emotional Expressions We measured emotional expressions in a similar manner. Subjects were asked “How likely would you be to express to Joanne (the decision-maker) the following emotions about the situation?” Again, satisfaction, anger, resentment, and frustration were assessed. Answers ranged from 1 (not at all likely) to 9 (very likely). Manipulation Checks To test the adequacy of our procedural justice manipulation, we asked subjects how arbitrary, biased, and inconsistent Joanne seemed in her selection of the committee members. Subjects indicated their responses on a scale ranging from 1, indicating not at all arbitrary, biased, or inconsistent, to 9, indicating very arbitrary, biased, or inconsistent. To assess our interpersonal ties manipulation we asked subjects “To what extent are NCEC volunteers friendly with each other?” and “How well do the NCEC volunteers work together?” Responses ranged from 1, indicating not at all friendly or not at all well, to 9, indicating very friendly or very well. Finally, to examine the endorsement manipulation, we asked subjects whether or not the other volunteers agreed Joanne was the right person for the job and whether or not they supported Joanne as the Director of Fundraising. The response scales ranged from 1 (other volunteers did not agree Joanne was the right person/did not fully support her) to 9 (Joanne was fully supported/thought to be the right person for the job). To make sure we had successfully created a situation seen as distributively unjust, we asked subjects whether or not they believed Marie to be qualified for the position. Subjects who thought Marie was qualified, but not chosen, were expected to perceive the unfairness of the situation. Answers ranged from 1 (not at all qualified) to 9 (very qualified). To assess our role-play method we asked subjects how easy (1) or difficult (9) it was to understand the situation. We also inquired: “How involved did you feel while playing the role of a volunteer at NCEC?” “Having taken the role of the NCEC volunteer, how realistic do you think your responses to the questionnaire were?” “How confident are you that your responses to the questionnaire reflect what you would be likely to perceive and how you would behave in such a situation?” For all scaled responses, 1 indicated a lack of involvement, realism, or confidence, and 9 represented sufficient involvement, realism, and confidence.

CH12-P374095.indd 216

5/26/2008 2:54:56 PM

Expressing Emotional Responses to Injustice of Others

217

RESULTS MANIPULATION CHECKS

ANOVA results generally indicate the success of our manipulations. Subjects in the low procedural justice condition were more likely than those in the high condition to see the procedures used by Joanne as arbitrary (M  6.79 versus 5.62; F (1, 138)  15.90, p  0.001), biased (M  8.03 versus 4.77; F (1, 154)  114.82, p  0.001), and inconsistent (M  7.15 versus 4.67; F (1, 153)  46.15, p  .001). Subjects in the strong interpersonal ties condition perceived other volunteers as more friendly with one another (M  6.98 versus 4.29; F (1, 154)  84.53, p  0.001) and as able to work better with one another (M  7.00 versus 4.82; F (1, 154)  65.90, p  0.001) than those in the weak interpersonal ties condition. And, subjects in the high endorsement condition were more likely than those in the low condition to perceive Joanne as supported by other group members (M  6.92 versus 3.73; F (1, 152)  135.06, p  0.001) and as the right person for the job (M  6.25 versus 3.61; F (1, 154)  61.69, p  0.001). No other situational factors affected these items.2 Analyses also revealed that we successfully created a distributively unjust situation and a successful role-play. The mean rating for how qualified subjects perceived Marie to be was 7.95. This rating did not vary significantly across conditions. Subjects also found the situation easy to understand (M  3.51), felt fairly involved (M  5.12), felt that the scenario was realistic (M  6.65), and most importantly, reported they were confident that their answers reflected their perceptions and anticipated behavior (M  7.02). These means did not vary across condition, but two gender differences were found. Females were significantly more likely than males to find the vignette easy to understand (M  3.24 versus 4.00; F (1, 156)  7.97, p  0.01) and to express more confidence in their answers (M  7.24 versus 6.62; F (1, 152)  4.70, p  0.05). GENERAL EMOTIONAL REACTIONS

Overall, observers were not satisfied with Joanne’s decision to decline Marie’s application for the fundraising committee (M  2.89). Observers were also moderately angry (M  5.04), resentful (M  4.70), and frustrated (M  5.46) regarding the outcome. This was expected as the vignette was designed to capture a distributively unjust situation. Regarding emotional expressions, observers were least likely to express satisfaction (M  2.49) and most likely to express frustration

2 Results, however, indicated that females (M  6.38) were more likely than males (M  5.70) to perceive the procedures as arbitrary (F (1, 138)  4.88, p  0.05). Notes left on the questionnaire and comments from subjects lead us to believe that these differences do not reflect actual differences in evaluation, but rather problems understanding the definition of the word “arbitrary.”

CH12-P374095.indd 217

5/26/2008 2:54:56 PM

218

Social Structure and Emotion

(M  5.84). The likelihood of expressing anger and resentment was similar to the intensity of feeling those emotions (M  4.86 and 4.60, respectively). This too was expected as emotional experiences were predicted to drive emotional expressions.

EFFECTS OF SITUATIONAL FACTORS ON EMOTIONAL EXPERIENCE

Table 12.1 presents OLS regression results for the effects of our independent variables on feelings of satisfaction, anger, resentment, and frustration. Results confirm Hypothesis 1, which predicted that observers are more likely to feel satisfaction when procedures are fair (M  3.67) than when procedures are unfair (M  1.95). Partial support is provided for Hypothesis 2. Observers are more likely to experience anger when procedures are unfair (M  5.79) than when they are fair (M  4.37), but procedural justice exerts no significant effects on feelings of resentment and frustration. Hypothesis 3 predicts an interaction effect between interpersonal ties and procedural justice, suggesting that interpersonal ties are positively related to an

TABLE 12.1 Unstandardized Effects of Procedural Justice (PJ), Interpersonal Ties (IT), and Endorsement (E) on the Experience of Satisfaction, Anger, Resentment, and Frustration Satisfaction

Anger

Resentment

Frustration

Constant

2.192*** (0.346)

5.138*** (0.488)

4.711*** (0.487)

5.155*** (0.517)

PJ

1.771*** (0.284)

1.25* (0.588)

0.471 (0.587)

0.197 (0.623)

IT

0.075 (0.285)

0.050 (0.512)

0.303 (0.511)

0.056 (0.542)

E

0.423 (0.284)

0.153 (0.512)

0.775 (0.511)

0.183 (0.543)

PJ  IT

–

0.040 (0.703)

1.060 (0.702)

0.702 (0.745)

PJ  E

–

0.761 (0.709)

1.600* (0.708)

1.465* (0.752)

Gender

0.607* (0.298)

1.075** (0.373)

0.638 (0.372)

1.323*** (0.395)

Adjusted R2

0.207

0.136

0.163

0.145

Notes: Numbers in parentheses are standard errors; N  160; gender is coded 0  male, 1  female; p-values are for two-tailed tests. * p  0.05 ** p  0.01 *** p  0.001

CH12-P374095.indd 218

5/26/2008 2:54:56 PM

Expressing Emotional Responses to Injustice of Others

219

observer’s experience of anger, resentment, and frustration only when the situation is procedurally unjust. No support emerged for this hypothesis. In contrast, the data provide some support for Hypothesis 4, which predicts an interaction effect between endorsement and procedural justice such that when procedures are unfair, endorsement will enhance negative feelings whereas fair procedures suppress such feelings regardless of level of endorsement.3 As expected, observers reported experiencing the most resentment (M  6.06) and frustration (M  6.23) when procedures were unfair and the decision-maker was endorsed. Also as predicted, those experiencing fair procedures and an endorsed authority compared to those experiencing unfair procedures and an illegitimate authority felt significantly less resentment (M  3.42 versus 5.27; t  6.17, p  0.001) and frustration (M  4.19 versus 6.04; t  3.49, p  0.001). Unexpected results, however, show that under conditions of unfair procedures, the means for frustration and resentment do not significantly differ across high and low endorsement (M  6.06 versus 5.27 and 6.23 versus 5.45, respectively). Instead, when the decision-maker is supported, frustration and resentment vary across high and low procedural justice. With high endorsement, those who experience unfair procedures are more likely than those who experience fair procedures to feel resentment (M  6.06 versus 3.42; F (1, 80)  31.41, p  0.001) and frustration (M  6.23 versus 4.19; F (1, 80)  17.48, p  0.001). Although this finding is slightly different than predicted, it highlights the fact that the combination of endorsement and procedural justice is important for understanding emotional experiences. The combined contextual factors, however, do not affect feelings of anger as predicted. Gender differences in emotional experiences show that females were less likely than males to experience satisfaction (M  2.69 versus 3.25), but more likely to experience anger (M  5.42 versus 4.34) and frustration (M  5.95 versus 4.55). For the experience of resentment, the pattern is similar to that of the other negative emotions, but the difference does not reach statistical significance (M  4.93 for females and 4.37 for males). EFFECTS OF SITUATIONAL FACTORS ON EMOTIONAL EXPRESSIONS

Table 12.2 shows the effects of the independent variables on emotional expressions alone and with the inclusion of the effect of emotional experience. Results provide no evidence for Hypothesis 5 regarding the effects of procedural justice on the expression of negative emotions. A marginally significant effect of procedural justice on satisfaction shows that when a decision-maker uses fair procedures, observers are likely to express more satisfaction than when he or 3

Tables of the means for all interaction effects are available from the authors. Means reported in the text for interaction effects involving procedural justice and endorsement are adjusted for gender of the respondent.

CH12-P374095.indd 219

5/26/2008 2:54:56 PM

CH12-P374095.indd 220

0.420 (0.426) 1.311* (0.596)

0.628 (0.472)

1.152 (0.660)

0.152

Gender

Gender  PJ

Adjusted R2 0.101

–

0.585 (0.412)

1.76* (0.784)

0.572 (0.777)

0.327

–

0.016 (0.366)

1.338* (0.681)

0.549 (0.673)

0.212 (0.490)

0.510 (0.490)

0.538 (0.566) 0.297 (0.566)

0.088 (0.571)

0.559*** (0.077)

2.452* (0.613)

A

0.786 (0.650)

–

5.33*** (0.539)

A

0.130

–

0.485 (0.414)

2.12** (0.788)

0.255 (0.781)

0.910 (0.569)

0.508 (0.568)

0.584 (0.653)

–

4.52*** (0.542)

R

0.327

–

0.2091

–

1.90*** (0.386)

1.82** (0.734)

–1.267 (0.704) 0.144 (0.368)

0.004 (0.728)

0.698 (0.530)

0.133 (0.529)

0.471 (0.608)

–

5.056*** (0.505)

F

0.312 (0.692)

0.496 (0.504)

0.346 (0.500)

–0.332 (0.575)

0.535*** (0.079)

2.002*** (0.605)

R



Notes: Numbers in parentheses are standard errors; N  160; gender is coded 0  male, 1  female; p-values are for two-tailed tests. p  0.10 * p  0.05 ** p  0.01 *** p  0.001

0.308

–

–

PJ  E

–

–0.134 (0.287)

–

0.059 (0.315)

E

0.152 (0.283)

0.009 (0.505)

0.478*** (0.080)

0.905* (0.425)

S

PJ  IT

0.121 (0.314)

0.960 (0.531)

–

1.904*** (0.433)

IT

PJ

Experience of S/A/R/F

Constant

S

0.488

–

1.131*** (0.322)

0.963 (0.598)

0.413 (0.587)

0.591 (0.426)

0.166 (0.426)

0.356 (0.489)

0.583*** (0.063)

2.050*** (0.522)

F

TABLE 12.2 Unstandardized Effects of Procedural Justice (PJ), Interpersonal Ties (IT), and Endorsement (E) on the Likelihood of Expressing Satisfaction (S), Anger (A), Resentment (R), and Frustration (F)

220 Social Structure and Emotion

5/26/2008 2:54:57 PM

Expressing Emotional Responses to Injustice of Others

221

she uses unfair procedures, consistent with Hypothesis 6 (M  3.30 versus 1.66; t  1.81, p  0.10). The inclusion of the experience of satisfaction, however, reduces this effect suggesting that emotional experience mediates the effects of the situational condition. Results partially confirm Hypothesis 7. As predicted, with unfair procedures and an illegitimate authority, observers were significantly more likely than those who saw the enactment of fair procedures by an endorsed authority to express anger (M  5.45 versus 3.47; t  3.60, p  0.001), resentment (M  5.08 versus 3.18; t  3.33, p  0.01), and frustration (M  6.23 versus 4.63; t  3.05, p  0.01). Further analyses show that when the decision-maker is supported, procedural justice has a particularly large effect. Under conditions of high endorsement, unfair procedures compared to fair procedures produce greater expressions of anger (M  5.72 versus 3.47; F (1, 80)  16.00, p  0.001), resentment (M  6.02 versus 3.18; F (1, 80)  23.41, p  0.001, and frustration (M  6.92 versus 4.63; F (1, 80)  19.61, p  0.001). And, the inclusion of emotional experience in the analysis demonstrates that it partially mediates the effects of the contextual factors on the expression of resentment and frustration, but not on the expression of anger. The size of the effect of emotional experience, however, certainly attests to its strong relationship with the expression of emotions. Gender differences also emerged in the expression of emotion. Results show that females are much more likely than males to express frustration, regardless of the inclusion of emotional experiences (M  6.54 versus 4.54). And the unexpected gender by procedural justice interaction effect demonstrates that females were more likely to express satisfaction under conditions of high procedural justice (M  3.47) than under conditions of low procedural justice (M  1.36). Males expressed satisfaction similarly whether or not procedures were fair or unfair (M  2.96 versus 2.00).

DISCUSSION This study examines how situational factors shape observers’ emotional reactions to the unfair treatment of another as well as encourage or suppress the display of such emotional reactions. Two general conclusions emerge from the analyses. First, observers do experience emotional responses to the unfair treatment of others and situational factors do affect the type and intensity of each felt emotion. Second, although emotional experiences drive expressions, they do not completely mediate the effects of situational factors on emotional displays. Specifically, as expected, procedural justice is a key determinant of observers’ experience of satisfaction and anger. Also, as predicted, observers who experience fair procedures by an endorsed decision-maker feel significantly less resentment and frustration than those who experience unfair procedures by an illegitimate decision-maker. Unexpectedly, however, endorsement qualifies the effects of procedural justice. Procedural justice influences resentment and frustration only when

CH12-P374095.indd 221

5/26/2008 2:54:57 PM

222

Social Structure and Emotion

the decision-maker is endorsed, and exerts little effect when the decision-maker lacks endorsement. When the decision-maker is supported, observers who experience unfair procedures are much more likely to experience resentment and frustration than those who experience fair procedures. In effect, endorsement, coupled with unfair procedures, heightens the experience of resentment and frustration because acting unfairly is unexpected and inappropriate by an endorsed authority. In addition, as predicted, observers who witness unfair procedures by an unsupported decision-maker are more likely to express anger, resentment, and frustration than those who see an endorsed decision-maker enact fair procedures. And, similar to the results pertaining to emotional experiences, an unexpected pattern emerges: with an endorsed decision-maker, observers are more likely to express all three negative emotions when the decision-maker used unfair procedures rather than fair procedures. Observers may view the inconsistency of a supported decision-maker using unfair procedures as an affront that needs to be addressed or informally sanctioned by the expression of negative emotions. Emotional experiences also have a very strong effect on the expression of emotions, but only partially mediate the interaction effects of procedural justice and endorsement on their expression. Although we predicted that interpersonal ties would encourage the experience and display of negative emotions when procedures were unfair, no significant differences emerged. Even though manipulation checks confirmed that subjects in the strong interpersonal ties condition recognized the friendliness of group members and their ability to work together, subjects may have had difficulty envisioning the experience of affective ties to these imaginary fellow volunteers. In addition, we did not specifically state that the participant likes the peer who receives the unfair outcome but only presumed it to be inferred. As a consequence, our manipulation may have failed to tap into the impact of affective ties on justice responses in the manner suggested by other researchers (e.g., Brockner and Greenberg 1990; Lind and Kray 2002). Behavioral studies of existing groups within organizations should provide a stronger test of the implications of interpersonal ties for another’s injustice. Finally, study results show gender differences in emotional experiences and expressions. Females indicated experiencing all of the emotions more intensely than did males. This finding is consistent with other research documenting that females report more frequent and intense feelings of both positive and negative emotions than do males (see Brody and Hall 2000). Also, females were more likely than males to express frustration (in all conditions) and satisfaction (under conditions of procedural justice). There were no gender differences, however, in the likelihood of observers expressing anger or resentment. Thus it appears that situational factors are more important than gender in determining emotional expression in this volunteer context. The role-play methodology only allows measurement of anticipated rather than actual emotional experiences and expressions. Future research should move beyond the vignette role-play method to studies that focus on respondents in

CH12-P374095.indd 222

5/26/2008 2:54:57 PM

Expressing Emotional Responses to Injustice of Others

223

existing organizations. Such research may find more evident effects of interpersonal ties on emotional reactions because allegiances have developed over time and over a variety of different types of interaction. In addition, by moving into organizations, comparisons of emotional management between volunteer organizations and organizations in which all members are paid could be made. In the latter type of organization, individuals may vary more explicitly in terms of the extent to which they are dependent upon an authority figure. In the volunteer situation examined here, observers may not have felt subordinate to the decision-maker. Yet, as Johnson et al. (2000) have shown, such dependency has a constraining effect on emotional expressions. Thus work should more explicitly examine the combined effects of power, procedural justice, and legitimacy on emotional experiences and expressions. In a pay situation involving a more definitive power structure or set of power dynamics, endorsement may have a suppressing effect, regardless of procedural justice. Finally, as this study finds that situational factors shapes emotion management, future research should investigate the specific types of tactics used to manage emotions. By assessing both experienced and expressed emotions, this study provides a glimpse of how unfairness and inequality are perpetuated in organizations. Future studies should do the same.

ACKNOWLEDGMENT We would like to thank Leslie Brody and Krysia Wrobel Waldron for their assistance in early stages of this project and the editors of this volume for their constructive comments. We also acknowledge the support of NSF grant SES0136747.

REFERENCES Adams, J. Stacy. 1965. “Inequity in Social Exchange.” Advances in Experimental Social Psychology 2:267–299. Brockner, Joel and Jerald Greenberg. 1990. “The Impact of Layoffs on Survivors: An Organizational Justice Perspective.” In: J. S. Carroll (ed.), Applied Social Psychology and Organizational Settings, pp. 45–75. Hillsdale, NJ: Erlbaum. Brockner, Joel, Mary Konovsky, Rochelle Cooper-Schneider, Christopher Martin, and Robert J. Bies. 1994. “Interactive Effects of Procedural Justice and Outcome Negativity on Victims and Survivors of Job Loss.” The Academy of Management Journal 37:397–409. Cohen-Charash, Yochi and Paul E. Spector. 2001. “The Role of Justice in Organizations: A MetaAnalysis.” Organizational Behavior and Human Decision Processes 86:278–321. Colquitt, Jason A., Donald E. Conlon, Michael J. Wesson, Christopher O. L. H. Porter, and K. Yee Ng. 2001. “Justice at the Millennium: A Meta-Analytic Review of 25 Years of Organizational Justice Research.” Journal of Applied Psychology 3:425–445. Erickson, Rebecca J. and Christian Ritter. 2001. “Emotional Labor, Burnout, and Inauthenticity: Does Gender Matter?” Social Psychology Quarterly 64:146–163.

CH12-P374095.indd 223

5/26/2008 2:54:57 PM

224

Social Structure and Emotion

Folger, Robert. 1986. “Re-Thinking Equity Theory: A Referent Cognitions Model.” In: Hans Bierhoff, Ronald L. Cohen, and Jerald Greenberg (eds), Justice in Social Relations, pp. 145–163. New York: Plenum Press. Gibson, Donald A. 1997. “The Struggle for Reason: The Sociology of Emotions in Organizations.” In: Rebecca J. Erickson and Beverley Cuthbertson-Johnson (eds), Social Perspectives on Emotion, Vol. 4, pp. 211–256. Greenwich, CT: JAI. Hegtvedt, Karen A. 1990. “The Effects of Relationship Structure on Emotional Responses to Inequity.” Social Psychology Quarterly 53:214–228. Hegtvedt, Karen A. and Karen S. Cook. 2001. “Distributive Justice: Recent Theoretical Developments and Applications.” In: Joseph Sanders and V. Lee Hamilton (eds), Handbook of Justice Research in Law, pp. 93–112. Kluwer Academic/Plenum. Hegtvedt, Karen A. and Cathryn Johnson. 2000. “Justice Beyond the Individual: A Future with Legitimation.” Social Psychology Quarterly 63:298–311. Hegtvedt, Karen A. and Caitlin Killian. 1999. “Fairness and Emotions: Reactions to the Process and Outcomes of Negotiation.” Social Forces 78:269–303. Hegtvedt, Karen A. and Barry Markovsky. 1995. “Justice and Injustice.” In: Karen Cook, Gary Allen Fine, and James House (eds), Sociological Perspectives in Social Psychology, pp. 257–280. Allyn & Bacon. Heider, Fritz. 1958. The Psychology of Interpersonal Relationships. New York: Wiley. Heise, David R. and Cassandra Calhan. 1995. “Emotion Norms in Interpersonal Events.” Social Psychology Quarterly 58:233–240. Hochschild, Arlie R. 1979. “Emotion Work, Feeling Rules, and Social Structure.” American Journal of Sociology 85:551–575. Hochschild, Arlie R. 1983. The Managed Heart: Commercialization of Human Feeling. Berkeley: University of California Press. Homans, George C. 1974. Social Behavior: Its Elementary Forms. Hartcourt Brace Jovanovich. Huo, Yuen J. 2003. “Procedural Justice and Social Regulation Across Group Boundaries: Does Subgroup Identity Undermine Relationship-Based Governance?” Personality and Social Psychology Bulletin 29:336–348. Johnson, Cathryn and Rebecca Ford. 1996. “Dependence Power, Legitimacy, and Tactical Choice.” Social Psychology Quarterly 59:126–139. Johnson, Cathryn, Rebecca Ford, and Joanne Kaufman. 2000. “Emotional Reactions to Conflict: Do Dependence and Legitimacy Matter?” Social Forces 79:107–137. Kemper, Theodore D. 1978. A Social Interactional Theory of Emotions. New York: Wiley. Kemper, Theodore D. 1987. “How Many Emotions Are There? Wedding the Social and Autonomic Components.” American Journal of Sociology 93:263–289. Kray, Laura and E. Allan Lind. 2002. “The Injustices of Others: Social Reports and the Integration of Others’ Experiences in Organizational Justice Judgments.” Organizational and Human Decision Processes 89:906–924. Krehbiel, Patricia J. and Russell Cropanzano. 2000. “Procedural Justice, Outcome Favorability, and Emotion.” Social Justice Research 13:339–360. Leventhal, Gerald S., J. Karuza, Jr., and W. R. Fry. 1980. “Beyond Fairness: A Theory of Allocation Preferences.” In: Gerold Mikula (ed.), Justice and Social Interaction, pp. 167–218. New York: Springer-Verlag. Lind, E. Alan, Laura Kray, and Leigh Thompson. 1998. “The Social Construction of Injustice: Fairness Judgments in Response to Own and Others’ Unfair Treatment by Authorities.” Organizational Behavior and Human Decision Processes 75:1–22. Lind, E. Alan and Tom R. Tyler. 1988. The Social Psychology of Procedural Justice. Plenum. Lupfer, Michael, Kelly Weeks, Kelly Doan, and David Houston. 2000. “Folk Conceptions of Fairness and Unfairness.” European Journal of Social Psychology 30:405–428. Morgan, Natasha, Leslie Brody, Krysia Wrobel, Karen A. Hegtvedt, and Cathryn Johnson. 2005. “When Will the Unaffected Seek Justice for Others? Perceptions of and Responses

CH12-P374095.indd 224

5/26/2008 2:54:57 PM

Expressing Emotional Responses to Injustice of Others

225

to Another’s Injustice.” Presented at the Meetings of the Southern Sociological Society, Charlotte, NC. Morgan, Rick and David R. Heise. 1988. “Structure of Emotions.” Social Psychology Quarterly 51:19–31. Mueller, Charles W. and Miriam J. Landsman. 2004. “Legitimacy and Justice Perceptions.” Social Psychology Quarterly 67:189–202. O’Reilly, C., III and J. Chatman. 1986. “Organizational Commitment and Psychological Attachment: The Effects of Compliance, Identification, and Internalization on Pro-Social Behavior.” Journal of Applied Psychology 71:492–499. Peters, Susanne L., Kees van den Bos, and D. Ramona Bobocel. 2004. “The Moral Superiority Effect: Self Versus Other Differences in Satisfaction with Being Overpaid.” Social Justice Research 17:257–274. Pugliesi, Karen and Scott L. Shook. 1997. “Gender, Jobs, and Emotional Labor in a Complex Organization.” In: Rebecca J. Erickson and Beverley Cuthbertson-Johnson (eds.), Social Perspectives on Emotion, Vol. 4, pp. 283–316. Greenwich, CT: JAI. Scher, Steven J. and David R. Heise. 1993. “Affect and the Perception of Injustice.” Advances in Group Processes 10:223–252. Scherer, Klaus R. 1992. Justice: Interdisciplinary Perspectives. Cambridge: Cambridge University Press. Simon, Robin W. and Leda E. Nath. 2004. “Gender and Emotion in the United States: Do Men and Women Differ in Self-Reports of Feelings and Expressive Behavior?” American Journal of Sociology 109:1137–1176. Skarlicki, Daniel P. and Carol T. Kulik. 2005. “Third-Party Reactions to Employee (Mis)Treatment: A Justice Perspective.” Research in Organizational Behavior 26:183–229. Skarlicki, Daniel, John Ellard, and Brad Kelln. 1998. “Third-Party Perceptions of a Layoff: Procedural, Derogation, and Retributive Aspects of Justice.” Journal of Applied Psychology 83:119–127. Smith, Allen C., III and Sherryl Keinman. 1989. “Managing Emotions in Medical School: Students’ Contacts with the Living and Dead.” Social Psychology Quarterly 52:56–69. Sprecher, Susan. 1992. “How Men and Women Expect to Feel and Behave in Response to Inequity in Close Relations.” Social Psychology Quarterly 55:57–69. Tajfel, Henri and John Turner. 1979. “An Integrative Theory of Intergroup Conflict.” In: William G. Austin and Stephen Worchel (eds), The Social Psychology of Intergroup Relations, pp. 33–37 Brooks/Cole. Thomas, George M., Henry A. Walker, and Morris Zelditch, Jr. 1986. “Legitimacy and Collective Action.” Social Forces 65:378–404. Turner, John C. 1987. Rediscovering the Social Group: A Self-Categorization Theory. New York: Basil Blackwell. Tyler, Tom R., Robert J. Boeckmann, Heather J. Smith, and Yuen J. Huo. 1997. Social Justice in a Diverse Society. Westview Press. Tyler, Tom R., Peter Degoey, and Heather J. Smith. 1996. “Understanding Why the Justice of Group Procedures Matter.” Journal of Personality and Social Psychology 70:913–930. Tyler, Tom R. and E. Alan Lind. 1992. “A Relational Model of Authority in Groups.” Advances in Experimental Social Psychology 25:115–191. van den Bos, Kees, Jan Bruins, Henk A. M. Wilke, and Elske Dronkert. 1999. “Sometimes Procedures Have Nice Aspects: On the Psychology of the Fair Process Effect.” Journal of Personality and Social Psychology 77:324–336. van den Bos, Kees and E. Allan Lind. 2001. “The Psychology of Own Versus Others’ Treatment: Self-Oriented and Other-Oriented Effects on Perceptions of Procedural Justice.” Personality and Social Psychology Bulletin 27:1324–1333. van den Bos, Kees, E. Allan Lind, and Henk A. M. Wilke. 2001. “The Psychology of Procedural and Distributive Justice Viewed from the Perspective of Fairness Heuristic Theory.” In: Russell Cropanzano (ed.), Justice in the Workplace, pp. 49–66. Mahwah, MJ: Lawrence Erlbaum.

CH12-P374095.indd 225

5/26/2008 2:54:57 PM

226

Social Structure and Emotion

Walker, Henry A., Larry Rogers, and Morris Zelditch, Jr. 1988. “Legitimacy and Collective Action: A Research Note.” Social Forces 67:216–228. Walker, Henry A., George M. Thomas, and Morris Zelditch, Jr. 1986. “Legitimation, Endorsement, and Stability.” Social Forces 64:620–643. Walster, Elaine, George W. Walster, and Ellen Berscheid. 1978. Equity: Theory and Research. Allyn & Bacon. Weber, Max. [1924] 1978. Economy and Society. In: G. Roth and C. Wittich (eds), Vol. I and II. Berkeley, CA: University of California Press. Wenzel, Michael. 2001. “What Is Social about Justice? Inclusive Identity and Group Values as the Basis of the Justice Motive.” Journal of Experimental Social Psychology 38:205–218. Zelditch, Morris, Jr. and Henry Walker. 1984. “Legitimacy and the Stability of Authority.” Advances in Group Processes 1:1–25. Zelditch, Morris, Jr. and Henry Walker. 2000. “The Normative Regulation of Power.” Advances in Group Processes 17:155–178.

CH12-P374095.indd 226

5/26/2008 2:54:57 PM

13 The Moral Identity, Status, Moral Emotions, and the Normative Order Jan E. Stets, Michael J. Carter, Michael M. Harrod, Christine Cerven and Seth Abrutyn University of California, Riverside, CA, USA

ABSTRACT In this chapter, we examine how the moral self as expressed in one’s moral identity guides behavior in interaction and influences reports of moral emotions, both of which serve to uphold the normative order. We also investigate absolute status and relative status in the social structure to see how each corresponds to reports of specific moral emotions. We examine three types of moral emotions: the “other-critical” emotion of anger, the “other-suffering” emotion of empathy, and the “self-critical” emotions of shame and guilt (Turner and Stets 2006). We identify the conditions under which individuals feel these moral emotions following exposure to situations in which actors engage in normative or counter-normative behavior. Data is collected from over 500 undergraduate students at a large southwestern university. The results help inform us as to how actors’ moral identities, together with the structural position of the self and others in the situation, influence actors’ emotional reactions to their behavior. Key Words: Emotions, Identity Theory, Moral Identity, Normative Order, Status

Social Structure and Emotion

CH13-P374095.indd 227

227

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 4:05:59 PM

228

Social Structure and Emotion

We attempt to link emotions to the social structure by showing how the moral emotions that actors experience serve as an important mechanism that simultaneously preserves and reinforces the self–society relationship. We maintain that while there are patterned regularities in human action, there are also patterned regularities in people’s responses to their actions as reflected in their emotional reactions to what they do. Because individual action is oriented toward the accomplishment of individually held goals, the pattern of which constitutes social structure, the valence of one’s emotional arousal signals how well an actor is doing vis-à-vis his or her goals. To the extent that a person feels positive emotions, we can say that the self and the social structure are mutually sustained. To the extent that a person experiences negative emotions, the self is disrupted, and these effects can reverberate at the social structural level. We examine the role of emotions in the self–society dynamic by studying the moral aspect of social life. We investigate how individuals as moral actors produce actions for their own goals, thereby generating the very patterns that form the moral order. We use identity theory to frame this research. We conceptualize the moral identity as operating within a perceptual control system where the identity serves as the goal state at the principle level, influencing moral conduct and moral emotions at the program level (Stets and Carter 2006). We measure the meanings of the moral identity and the salience of the moral identity to see whether each reliably predicts what actors do and how they subsequently feel in situations. What is important about investigating both how the moral identity is characterized (meanings) and the likelihood of invoking the moral identity across situations (salience) is that while the former has been emphasized in the perceptual control version of identity theory (Burke 1991), the latter has been emphasized in the social structural version of identity theory (Stryker 1980). We think it is important to study both the internal dynamics of the moral identity as well as its social structural aspects since both may exert their own independent effects on moral conduct and moral emotions. We bring social structure more directly into our analysis by examining the effect of one’s structural position of gender, race/ethnicity, and income (absolute status) on moral emotions. Gordon (1989, 1990) has argued that one’s social position may differentially expose a person to some emotions over others. Indeed, differential access to resources may influence an individual’s likelihood of being the target or victim of certain emotions. As Gordon (1990:161–162) remarks: “The wealthy and physically attractive must cope with envy directed at them, while the poor and unattractive must adapt to disparaging humor, contempt, and moral indignation directed against them.” Thus, we examine the social distribution of moral emotions. We also examine the structural position of others in the interaction by studying whether the presence of higher status others and equal status others (relative status) facilitates or inhibits moral emotional reactions. This is important because recently it has been argued that identity theorists need to acknowledge that emotional arousal may not only emerge from identity processes, but also from other processes that are simultaneously occurring in the situation that may also

CH13-P374095.indd 228

5/26/2008 4:05:59 PM

The Moral Identity, Status, Moral Emotions, and the Normative Order

229

influence one’s affective experience (Stets 2004, 2006; Stets and Burke 2005). If multiple meanings are transmitted in an encounter, and one of these meanings is the status of the actors in the situation including their level of power, resources, and prestige (Berger, Cohen, and Zelditch 1972), then this set of meanings also likely influences moral affective responses. For example, we may be less likely to express the other-directed emotion of anger to a higher status other because, however justified we may feel about the anger, we may fear retaliation given the greater power of higher status others; alternatively, we may be more likely to report the inner-directed emotion of shame because we may feel that we have failed to live up to their standards of higher status others (Ridgeway and Johnson 1990). If identity meanings persist in the face of other meanings such as status meanings, then the power of identity theory becomes clearer. In general, this research examines the social psychological forces (identity meanings and salience) and social structural forces (absolute status and relative status) that help shape moral emotions. The expression of these moral emotions helps maintain the normative order at the microlevel and the moral order at the macrolevel.

MORALITY AND THE MORAL IDENTITY Before we discuss the self as a moral entity, we must first define morality. Broadly speaking, morality is the “evaluative cultural codes that specify what is right or wrong, good or bad, acceptable or unacceptable” (Turner and Stets 2006). Following Turner and Stets, as one moves from the macro to the microlevel of analysis, correspondingly, moral codes decrease somewhat in their evaluative content on the good–bad meaning dimension. At the macrolevel, moral codes are the values that influence proper or improper behavior across situations. Values that characterize our society include personal control, individualism, and competition. Moral codes that identify what is good or bad within a specific institutional domain are ideologies (Turner and Stets 2006). For example, the value of control is translated into the institution of the family through our beliefs about parents having the right to control their children. Moral codes at lower levels in the social structure such as face-to-face interaction are norms. Norms dictate appropriate and expected behavior in situations. Norms are not as heavily laden with evaluative content as to what is good/bad or right/wrong as values in the larger social structure, but social actors emotionally react when they violate them. For example, when a person cheats on an exam, the feeling following the act of cheating may be guilt. Thus, we label emotions that are “aroused in reference to cultural codes that contain evaluative content” as moral emotions (Turner and Stets 2006:556). THE MORAL IDENTITY

The overall self is organized into multiple parts known as identities (Stryker and Burke 2000). A person has “an internal positional designation” (Stryker 1980:60) for each of the different groups to which he or she belongs (group

CH13-P374095.indd 229

5/26/2008 4:05:59 PM

230

Social Structure and Emotion

identities), role relationships in which he or she participates in (role identities), and kind of person he or she claims to others (person identities). An identity is a set of meanings one holds for the self as a group member, role holder, or as a person (Stets 2006). For example, the role identity of mother and part of the content as to what it means to be a mother includes being nurturing and caring. These meanings guide the activity of the mother role as revealed in feeding and bathing the child and engaging in warm and intimate interactions with the child. Thus, identity meanings are the standards or goal states for behavior. The moral identity is a person identity in which the meanings involve sustaining the self as a good/bad entity. Recently, it has been argued that the moral identity has two underlying facets: (1) a concern for justice and the preservation of human rights (Kohlberg 1981) and (2) a concern for care and the preservation of mutual relationships (Gilligan 1982; Stets and Carter 2006). There may be other aspects to this meaning dimension, and there may be multiple dimensions to the moral identity, but these issues we leave for future research. This starting point is based on a large body of research that has developed over the past 20 years and has examined the justice and care orientations, primarily in terms of variation by gender (Jaffee and Hyde 2000). We are not interested in gender differences in the meaning structure of the moral identity. Indeed Jaffee and Hyde’s meta-analysis does not provide strong support that men and women differ in care and justice orientations. Additionally, more recent research indicates no gender difference in the meaning structure of the moral identity (Carter 2005). We also are not interested in relating each of these orientations to moral behavior because Gilligan, who made the distinction between justice and care, did not identify how each motivates different types of moral behavior (Jaffee and Hyde 2000). Further, Gilligan may have oversimplified Kohlberg’s view of morality in terms of justice since evaluations of what is right and fair (justice) are made in the context of issues of social responsibility and concern for others (care) (Kohlberg, Levine, and Hewer 1983). Thus, if both justice and care simultaneously guide moral action, they constitute potentially core features of the moral identity. Researchers have been using these two facets in their measure of being moral (Aquino and Reed 2002; Walker and Hennig 2004), and we do as well. We conceptualize the moral identity at one of the highest levels (the principle level) in the overall hierarchy of identities (Stets and Carter 2006). As a person identity, the moral identity operates like a “master” identity in that it influences the selection of other identities such as role and group identities and the corresponding behaviors associated with these identities (Burke 2004). In this way, the moral self operates across various roles and within different groups and thus may figure prominently into many interactions because it is always on display (Burke 2004). For our purposes, we argue that the moral identity importantly guides normative behavior in interaction. Thus, we offer the following hypothesis: Hypothesis 1:

CH13-P374095.indd 230

The more individuals report a high moral identity, the more those individuals will enact normative behavior.

5/26/2008 4:05:59 PM

The Moral Identity, Status, Moral Emotions, and the Normative Order

231

Since individuals have as many identities as there are groups to which they belong to in society, roles that they occupy in the social structure, and characteristics that distinguish them from others they interact with, these identities are conceptualized in identity theory in terms of a salience hierarchy (Stryker 1980). The salience hierarchy reflects which identities are more likely to be activated or played out across situations. Identities higher in the hierarchy are more likely to be enacted across situations, thus they are more salient. In general, the identity salience hierarchy captures the idea that as society is complex, differentiated, yet organized hierarchically, so too is the self. The salience hierarchy portrays the self as linked to the social structure in multiple ways, particularly through its many group memberships and roles. Thus, while moral identity meanings discussed above address the internal dimension of this identity, its content, the salience of the moral identity attends to the external or social structural features of this identity by comparing the ranking of one’s moral identity to other identities that the self simultaneously holds in the social structure. In the same way that we anticipate the moral identity to influence normative conduct, we also expect the relative salience of the moral identity to influence normative behavior. Thus, we hypothesize that: Hypothesis 2:

The higher the salience of the moral identity for individuals, the more those individuals will enact normative behavior.

MORAL EMOTIONS We study three types of moral emotions: the other-critical emotion of anger (Rozin et al. 1999), the other-suffering emotion of empathy (Hoffman 2000), and the self-critical emotions of shame and guilt (Tangney and Dearing 2002). While there are other moral emotions such as the other-praising emotions of awe and elevation (Haidt 2003), and while even the above commonly identified moral emotions might be expanded to include a range of other moral emotions (see Turner and Stets 2006 on this expansion), we begin an empirical investigation into moral emotions, sociologically, by focusing on a select few. Importantly, this research begins to move sociological analysis of emotional responses beyond the gross categories of negative and positive by grouping some primary emotions and secondary emotions into the category of moral emotions (Turner and Stets 2005). The other-directed emotions of anger and empathy emerge under slightly different situations: while anger appears when one sees that someone or something is responsible for one’s bad state of affairs, empathy surfaces when one sees that something bad has happened to another. Of the four moral emotions we investigate, anger is the only “primary” emotion; the others are secondary emotions. Primary emotions, that is, happiness, fear, anger, and sadness, serve as the foundation for all other emotions (Turner and Stets 2005). Secondary emotions emerge in the context of experiencing one or more of the primary emotions

CH13-P374095.indd 231

5/26/2008 4:05:59 PM

232

Social Structure and Emotion

(Kemper 1987). For example, shame and guilt emerge from differing combinations of sadness, fear, and anger, with anger stronger when people feel shame and fear stronger when people experience guilt (Turner 2000). Anger is often expressed when one’s goals have been blocked. In response to the obstruction, one may strike out at another, particularly if another is judged responsible, and this helps restore justice (Averill 1982). Rozin et al. (1999) argue that anger is associated with the violation of the moral code of autonomy. This moral code involves respect for individuals’ rights, freedoms, and choices. The researchers maintain that anger typically emerges when another’s actions are seen as infringing upon one’s rights and liberties. The others’ actions are judged as unjust and unfair. Empathy as an “emotional” experience involves either sharing another’s emotional state or experiencing some emotional state such as anger in response to another’s state (Davis 2006). This is different from the conceptualization of empathy as a “cognitive” experience in which one identifies with another’s emotional state without a corresponding emotional change in oneself. There is some debate as to whether empathy is an emotion at all or simply the ability to feel what another is feeling which, in turn, may lead to the arousal of other emotions (Haidt 2003; Turner and Stets 2006). We think that what is important about empathy as a moral emotion is that, as a feeling state, it is laced with moral consequences for social relations. To the extent that one feels what another is feeling, it promotes helping behavior and tends to inhibit aggressive behavior (Davis 2006). Given the shared feeling state, it facilitates solidarity among actors. As Tangney and Dearing (2002:79) write, “Empathy is the ‘good’ moral affective capacity or experience, leading us in moral directions and diverting us from paths of vice and perdition. Experiences of empathy help us to accurately ‘read’ or interpret interpersonal events, allowing us to respond sensitively to the feelings of others.” Of the four emotions we examine, shame and guilt have been more likely to be labeled moral emotions (Tangney 1991, 2003; Tangney and Dearing 2002). However, the two emotions are affectively different, and according to Tangney, “not equally ‘moral’ emotions” (2003:388). Shame is a more intense feeling than guilt because when the self is involved in the evaluation of a violation of normative codes, the “whole self ” is implicated. One is seen as a “horrible agent” who has caused a misdeed. The self feels small and worthless. One wants to hide, escape, or disappear. In contrast to shame in which the self is seen as a “bad person,” guilt emerges when the self has done a “bad thing.” Because the entire self is not implicated in the evaluation of one’s misdeed, guilt is less painful than shame. The self feels remorse or regrets for the offending behavior, and this motivates the self to confess, apologize, and engage in reparative action. Of the two emotions, Tangney maintains that guilt rather than shame is the moral emotion (Tangney 1991, 2003; Tangney and Dearing 2002). Because the entire self is not “under siege” in feelings of guilt, the self can remain engaged in the encounter rather than wanting to flee. In remaining interpersonally involved,

CH13-P374095.indd 232

5/26/2008 4:05:59 PM

The Moral Identity, Status, Moral Emotions, and the Normative Order

233

the self is able to focus on the other including taking his or her role, empathizing with him or her, and in response to both, do something so that the transgression may be forgiven. Thus, through empathy and guilt one adjusts his or her behavior in a way that facilitates cooperation with others and fosters interpersonal attunement – in short, the self is morally redirected. Tangney argues that there is a special link between empathy and guilt; shame actually interferes with feeling empathy (Tangney 1991, 2003; Tangney and Dearing 2002). Because shame directs negative feelings inward, onto the self, this self-focus is incompatible with the other-focus that is characteristic of empathy. Being consumed with feeling that one is a “horrible person” interrupts attending to the pain that another person may be feeling. In contrast, guilt appears to foster empathy. In focusing on one’s offending behavior rather than the entire self, the person avoids the egocentric preoccupation characteristic of shame, thereby allowing attention to drift to a distressed other and the harm that one has inflicted. In response, offenders make moves to restore goodwill. We are not in full agreement that guilt and empathy are the moral emotions and shame is not. Shame need not always be devastating to the whole self (Turner and Stets 2006). When a particular identity becomes activated in a situation, the shame that one feels for failing to live up to the standards of the identity may only “attack” that identity. All other identities that comprise the self may remain unaffected by the attack. Thus, in the encounter, the self may still be able to take the role of the other, cooperate, and steer itself in the direction of a more appropriate moral response. Therefore, while we recognize that shame may be a more painful feeling than guilt, particularly if one is not living up to a salient identity, we still consider it a moral emotion because, certainly at low intensity levels, shame can “operate as an internal control mechanism, keep individuals ‘in line,’ and in turn, sustain commitments to others and produce solidarity” (Turner and Stets 2006:553). While we expect the moral identity to guide behavior, we also expect it to influence one’s reports of moral emotions. Consistent with identity theory, when individuals with a high moral identity behave in a manner that is inconsistent with normative expectations, they should experience distress (Burke 1991). In the current research, we expect that this distress will translate into the expression of moral emotions. For example, doing something that blocks one’s goals should lead to the response of anger or doing a bad thing should produce feelings of guilt. In turn, these moral feelings should redirect behavior to what is normatively expected. Therefore, we hypothesize that when a norm is violated: Hypothesis 3:

The more individuals report a high moral identity, the more they will express moral emotions.

We anticipate a similar effect for the salience of the moral identity in response to the violation of a norm. Those with a more salient moral identity should be more likely to play out this identity across situations. Thus, when these individuals behave counter-normatively in situations, they should be more likely than

CH13-P374095.indd 233

5/26/2008 4:05:59 PM

234

Social Structure and Emotion

those with a less salient moral identity to report moral emotions such as shame or guilt. Thus, we expect that when a norm is violated: Hypothesis 4:

The higher the salience of the moral identity for individuals, the more those individuals will report moral emotions.

While a discrepancy between a high moral identity and a salient moral identity and “bad” behavior should result in the expression of moral emotions, we also expect one’s behavior to exert its own influence on reports of moral emotions. Specifically, we expect that normative acts should decrease reports of moral emotions since the act, itself, is upholding the moral order. Alternatively, counter-normative behavior should increase reports of moral emotions. The moral emotions following counter-normative behavior should help redirect behavior in a manner that is more consistent with the expectations of the situation. Therefore, we hypothesize that: Hypothesis 5:

The more individuals enact normative behavior, the less those individuals will report moral emotions.

SOCIAL STRUCTURE AND MORAL EMOTIONS Moral emotions help maintain the social order. Since social structures are systems of status positions, we study how status positions influence the expression of moral emotions. We examine the absolute status of respondents in terms of gender, race, and income and the relative status of actors in an interaction, that is, whether others present are of a higher status or equal status to the actor. If anger is a moral emotion, it is also an emotion associated with power that people in positions of high status are more likely to use, and others considered it legitimate for them to do so (Schieman 2006). In terms of our status characteristics, we would expect men to display more anger than women (Shields et al. 2006), and those of high status such as whites and those with high income to report more anger (Conway, Di Fazio, and Mayman 1999). Consistent with this, we would expect that in the presence of a higher status other, individuals would be less likely to report anger. It is not clear how the presence of an equal status other would influence reports of anger as they could facilitate or inhibit the expression of anger, thus we make no predictions about this relationship. Assuming that men, whites, and those with a high income are of high status and women, nonwhites, and those with a low income are of low status, we hypothesize that: Hypothesis 6: High status actors will report more anger than low status actors. Hypothesis 7: Individuals will report less anger in the presence of a higher status other.

If anger is an emotion of the powerful, we anticipate that empathy is an emotion of the powerless. Those who experience hardships may be more sensitive to the plight of others. Looked at another way, if empathy involves taking the role

CH13-P374095.indd 234

5/26/2008 4:05:59 PM

The Moral Identity, Status, Moral Emotions, and the Normative Order

235

of the other, there is evidence that in an interaction lower status actors are more likely to engage in role taking than higher status actors in order to best fit their line of action with those who have more power (Franks 1985). We do not know how relative status in the situation will influence reports of empathy, thus we make no predictions here. Consequently, where women, nonwhites, and those with a low income are of a low status, and their counterparts of men, whites, and those with a high income are of a high status, we expect that: Hypothesis 8: Low status actors will report more empathy than high status actors.

Finally, we anticipate that low status actors will be more likely to report shame and guilt than high status actors. Low status actors may be more likely to fear a negative evaluation when they violate some moral code. In shame, a low status actor may fear exclusion from the group, and in guilt, they may fear retaliation of their “bad” act by the victim. Consistent with this, for example, Tangney and Dearing (2002) report that over many of their own studies, women report greater shame and guilt than men. We also expect that in the presence of a higher status other, individuals will be more likely to report shame and guilt. It is not clear how the presence of an equal status other would influence reports of shame and guilt, thus we make no predictions about this relationship. Consequently, we anticipate that: Hypothesis 9: Hypothesis 10:

Low status actors will report more shame than high status actors. Individuals will report more shame in the presence of a higher status other.

Similarly, for guilt we expect that: Hypothesis 11: Hypothesis 12:

Low status actors will report more guilt than high status actors. Individuals will report more guilt in the presence of a higher status other.

METHOD SAMPLE

We administered a survey to undergraduates enrolled in upper- and lowerdivision sociology classes at a large southwestern university during the spring of 2005. Of the total 699 students sampled across four classes, 519 filled out the survey. Assuming perfect attendance, this is a 74% response rate. Since perfect attendance is rare in the classroom, a conservative estimate of an 80% attendance rate increases our response rate to 93%. More females (60%) than males (40%) filled out the survey. The average age of the respondents is 22. Respondents’ parents’ average income is between $35,000 and $50,000. Respondents’ racial/ ethnic background is diverse consisting of Asian Americans (39%), Latinos/ Chicanos (24%), Whites (22%), African Americans (5%), and other racial/ethnic backgrounds (10%).

CH13-P374095.indd 235

5/26/2008 4:05:59 PM

236

Social Structure and Emotion

SURVEY DESIGN AND PROCEDURE

There are two versions of the survey and two parts to each survey version. In each version, individuals respond to five vignettes. The five vignettes in version 1 are different from the five vignettes in version 2. We split the sample into two equal parts with half responding to version 1 vignettes and the other half responding to version 2 vignettes. Thus, there are 10 vignettes with roughly 260 individuals responding to five vignettes and 260 individuals responding to another five vignettes. This facilitates examining a wide array of normative situations. Each vignette outlines a specific situation in which the moral code is to respond normatively, that is, do the “good” behavior/”right” thing. The respondents are to think about the last time they were in these situations. We use the period of the “last time” for ease of recall. Respondents are to report on their conduct in each of the situations and indicate how they felt following their actions. If respondents had never experienced the situation, they are to imagine themselves in this situation and respond as to what they would do and how they would feel following their actions. Below, (1)–(5) appear in version 1 of the survey and (6)–(10) appear in version 2:

VERSION 1

(1) “Just before entering a store, you are asked to donate to a charity organization.” You “provided a donation” or you “did not provide a donation.” (2) “You receive an exam from your friend. The exam is from a class your friend previously took with the same professor in which you are now enrolled. By acquiring the exam, you know the answers for an upcoming test.” You “reviewed the exam for the answers” or you “did not review the exam for the answers.” (3) “A friend of yours is drunk and wants to drive himself/herself home.” You “let your friend drive home” or you “did not let your friend drive home.” (4) “You are short on cash and you find some money that belongs to someone in your household.” You “took the money” or you “did not take the money.” (5) “You dented someone’s car in the parking lot.” You “left a note identifying yourself as the responsible party” or you “did not leave a note identifying yourself as the responsible party.”

VERSION 2

(6) “A cashier returns more money to you than what is owed.” You “returned the money” or you “did not return the money.” (7) “You hear a person in a group say something that is unkind about your friend.” You “said something kind in defense of your friend” or you “did not say something kind in defense of your friend.”

CH13-P374095.indd 236

5/26/2008 4:06:00 PM

The Moral Identity, Status, Moral Emotions, and the Normative Order

237

(8) “You are in a committed relationship and you have the opportunity to have a ‘one-night stand’ with someone else.” You “had the ‘one-night stand’” or you “did not have the ‘one-night stand.’” (9) “You receive money from your parents only if you are doing well in school. At the end of the term, your parents ask you how you did, and only you know you did poorly.” You “told your parents how poorly you did” or you “did not tell your parents how poorly you did.” (10) “You are driving with another person in your car, and you are late for an appointment. The light turns red at the intersection.” You “ran the red light” or “did not run the red light.” Several things about the vignettes are worth noting. First, the underlying theme for each situation is parallel, in form, across the two survey versions; however, the situation is expressed in a slightly different way in version 1 compared to version 2 to capture a different dimension of the underlying theme. For example, (1) and (6) both involve money, but in the former the person is making a donation to assist the less fortunate while in the latter one is returning a surplus that the person received. In (2) and (8) cheating is involved, but one of the items entails cheating on an exam while the other entails cheating on one’s partner. Both (3) and (7) have to do with helping a friend, but the former is a strategy to protect a friend while the latter is a strategy that defends a friend. Unfairly taking from another is involved in (4) and (9) with the former involving unfairly taking money from any household member while the latter involves unfairly taking from one’s parents. Finally, (5) and (10) have to do with traffic violations or the potential to commit them. Second, an earlier and larger set of vignettes were pretested on 125 undergraduate sociology students enrolled in upper- and lower-division sociology classes at a large southwestern university to see if the respondents perceived the situations like those described above as laced with a moral code. We gave half of the sample one set of vignettes and the other half another set of vignettes. After respondents read the hypothetical situation, they were told how the actor in the situation behaved, and then on the basis of how the actor behaved, the respondent was to evaluate the actor on a set of 12 bipolar characteristics that operationalized the good–bad dimension (for example, the actor was honest–dishonest, kind–unkind, truthful–untruthful, and so forth). A principle components factor analysis of the bipolar characteristics revealed that 10 vignettes had the strongest factor structures in reflecting the underlying dimension of morality. When the actor in the situation either responded in the right/moral way or neglected to respond in the right manner, respondents described the actor in heavily moral (or immoral) terms. We use these 10 vignettes in the current analysis. Third, we use the methodological procedure of vignettes as has been done in other studies on emotions, particularly jealousy (Ellestad and Stets 1998; Hansen 1991) for two reasons. If we were to ask people to report on their own experiences of negative emotions such as anger, shame, or guilt, they may be less reticent to do so for reasons of social desirability, that is, they would want to project a positive image and report positive emotions and not a negative image

CH13-P374095.indd 237

5/26/2008 4:06:00 PM

238

Social Structure and Emotion

that would be suggestive of negative emotions. Additionally, since negative emotions are more likely to be repressed (Turner 2006); they may be more difficult to access by the self, thus making vignettes a useful methodological tool. In seven of the ten vignettes, most respondents reported having experienced the situation that was described, thus they are relying on the last time this happened to them (rather than their imagination) to respond. These situations included “asking to donate to a charity” (93%), “having the opportunity to review an exam for the answers” (55%), “having a friend who is drunk and wants to drive himself/herself home” (72%), “finding money that belongs to another in the household” (62%), “having a cashier return more money than what is owed one” (76%), “hearing a group say something unkind about a friend” (94%), and “having the light turn red at an intersection when one is late for an appointment” (87%). Respondents are less likely to experience the following three situations: “denting someone’s car in a parking lot” (37%), “having the opportunity for a ‘one-night stand’” (39%), and “telling one’s parents about one’s poor performance” (44%). MEASURES

Dependent Variables Normative Behavior Respondent’s behavior in each of the vignettes was coded 0–1 with 0 reflecting the “wrong,” “bad,” or normatively inappropriate behavior and 1 reflecting the “right,” “good,” or normatively appropriate behavior. Thus, across the vignettes, the following behaviors would be coded 1: “providing a donation,” “not reviewing the exam for the answer,” “not letting one’s friend drive home drunk,” “not taking the money from a household member,” “leaving a note identifying oneself as the responsible party to a dented car,” “returning the money to the cashier,” “saying something in defense of a friend,” “not having a ‘one-night stand,’” “telling one’s parents about how poorly one has done,” and “not running the red light.” The alternative to each of the behaviors listed above was coded 0. Moral Emotions Following respondents’ reports of how they behaved (or imagined they would behave) in each situation, they indicated how their behavior made them feel. For ease of responding, they were provided an “emotion wheel” similar to that used by Heise and Calhan (1995). Respondents were provided with a circle, and on the circle were eight emotions. On the top of the wheel was listed happiness, and in a clockwise fashion appeared sadness, anger, fear, pride, empathy, shame, and guilt (coded 0–8). We asked respondents to choose the emotion they felt (or would have felt) after engaging in the behavior they reported. If respondents recalled multiple emotions during the situation, they were to choose the emotion they felt the strongest.

CH13-P374095.indd 238

5/26/2008 4:06:00 PM

The Moral Identity, Status, Moral Emotions, and the Normative Order

239

For our purposes, we are interested in individuals’ responses across the vignettes on the four moral emotions: anger, empathy, shame, and guilt. Each emotion was coded 0–1, where 1 represented the specific emotion and 0 otherwise. Of the four moral emotions, guilt was reported most often (44%) followed by anger (21%), empathy (18%), and shame (17%). Independent Variables Moral Identity To measure the meanings of the moral identity, we gave respondents a list of 12 bipolar characteristics: honest/dishonest, caring/uncaring, unkind/kind, unfair/ fair, helpful/not helpful, stingy/generous, compassionate/hardhearted, untruthful/truthful, not hardworking/hardworking, friendly/unfriendly, selfish/selfless, and principled/unprincipled. We asked them to think about how they saw themselves given each characteristic and to identify where they would place themselves between each bipolar characteristic. Responses ranged from 1 to 5, where 1 reflected agreement with one bipolar characteristic, 5 reflected agreement with the other bipolar characteristic, and 3 placed the respondent in between the two bipolar characteristics. Items 2 and 3, 5–7, and 10 and 11 measured “care,” and the remaining items measured “justice” (Walker and Hennig 2004). These items were based on two recent studies where respondents identified the characteristics that a moral person possesses (Aquino and Reed 2002; Walker and Hennig 2004). Nine out of the twelve items were used in a recent study that investigated the moral identity (Stets and Carter 2006). Using a bipolar design is consistent with the Burke/Tully method of measuring the self-meanings of an identity (Burke and Tully 1977). Like other identity measures such as gender identity (Stets and Burke 1996), the student identity (Burke and Reitzes 1991), and the environment identity (Stets and Biga 2003), this method attempts to capture the multiple aspects of a dimension to which individuals can respond to themselves as an object. Since the focal point in responding to the characteristics is the person rather than a role one holds or a group to which one is a member, this measurement procedure is consistent with how the moral identity is conceptualized, that is, as a person identity. Compared to how identities are usually measured, we did not have at least two criterion categories for the moral identity (a moral and nonmoral person) from which we could discriminate. For example, for gender identity, we have the criterion categories of being male or female from which to discriminate between masculinity and femininity. In the absence of criterion categories for the moral identity, we could not use discriminant function analysis. We used principle components factor analysis as an appropriate alternative. In factor analysis, we identify a cluster of meanings that appear on one factor, and this becomes the dimension of meaning for the moral identity. The results of our principle components factor analysis are provided in Table 13.1. The items form a single factor with an omega reliability of 0.89. Negatively worded characteristics were reverse

CH13-P374095.indd 239

5/26/2008 4:06:00 PM

240

Social Structure and Emotion

TABLE 13.1

Principle Components Factor Analysis

for Moral Identity Items

Factor loading

Honest

0.59

Caring

0.65

Kind

0.63

Fair

0.54

Helpful

0.61

Generous

0.52

Compassionate

0.53

Truthful

0.60

Hardworking

0.52

Selfless

0.47

Friendly

0.58

Principled

0.52

Eigenvalue

3.19



0.89

coded, the items were summed with a high score representing a high moral identity, and the scale was standardized (mean  0; SD  1). Moral Identity Salience This was measured using the procedure suggested by Stryker and Serpe (1994) by asking respondents to identify how they would describe themselves to another when meeting for the first time. We presented three situations: (1) meeting someone at a party for the first time, (2) meeting a friend of a close friend, and (3) going on a date for the first time. Among a list of descriptors (friend, student, moral person, worker, religious person, boyfriend/girlfriend), respondents indicated which they would tell the person first, which they would tell the person second, and so on in each of the situations. Responses ranged from 1–6. For the responses to being a moral person in each situation, the responses were highly correlated (r  0.65 for meeting another at a party and meeting a friend; r  0.54 for meeting another at a party and going on a date; and r  0.61 for meeting a friend and going out on a date; all correlations are significant at p  0.05).We summed the items across the three situations, with a high score representing a more salient moral identity. The omega reliability for this measure was 0.83. Absolute Status Gender was coded 0 for females and 1 for males. The racial/ethnic background of respondents was coded 0 for nonwhite and 1 for white. Income was coded 0 for low income ($50,000) and 1 for high income ($50,000).

CH13-P374095.indd 240

5/26/2008 4:06:00 PM

The Moral Identity, Status, Moral Emotions, and the Normative Order

241

Relative Status For each situation, we asked respondents additional questions surrounding the context. They were to identify whether: (1) a person in a position of authority was nearby (coded 0 for no; 1 for yes) and (2) whether peers were present (coded 0 for no; 1 for yes). ANALYSIS

Rather than the unit of analysis being the individual, the unit of analysis is an individual’s response to a vignette. Since each individual responded to five vignettes, there are five responses/cases per person for about 2600 responses. For the analysis of moral emotions, we are examining respondents’ emotional reactions of anger, empathy, shame, and guilt across all the vignettes. Since respondents’ behavior and their emotional reaction to their behavior are each coded 0–1, we estimate these equations using the maximum likelihood procedure of logit analysis. Since individuals in our analysis are responding to multiple vignettes, the error terms for the multiple responses for each respondent are not independent, and thus allowed to be correlated. To accomplish this, we used the cluster option in STATA. The logit estimates we provide are in terms of changes in the odds ratio for a one-unit change in the independent variable. Departures from an odds ratio of 1 in a positive direction indicate the proportional increase in the odds of engaging in the normative behavior (or reporting a moral emotion) for a one-unit increase in the independent variable. Similarly, departures from an odds ratio of 1 in a negative direction indicates the proportional decrease in the odds of engaging in the normative behavior (or reporting a moral emotion) for a one-unit increase in the independent variable.

RESULTS Table 13.2 presents the frequency distribution of the four moral emotions by the ten vignettes. Looking down each of the columns, the results reveal that each of the moral emotions is reported differently across each situation. Anger is most likely to be reported in the traffic situation of being stopped at a red light when late for an appointment (75%). Less frequent responses of anger occur in the situations having to do with helping a friend, either by defending a friend after something unkind has been said about him/her (53%) or protecting the friend by not letting him/her drive home drunk (47%). Empathy is most often reported in the donation to charity situation (48%) although guilt is almost as frequently reported in this situation (42%). Shame is most often reported in the scenario of telling one’s parents about one’s poor school performance (59%). Guilt frequently appears across four other scenarios: cheating (either for an upcoming exam (68%) or on one’s partner (68%)), the traffic situation of denting another’s car (72%), and returning money to a cashier (62%). The only other situation in

CH13-P374095.indd 241

5/26/2008 4:06:00 PM

242

Social Structure and Emotion

TABLE 13.2

Percentages of Moral Emotional Reactions to Vignettes (N  1031)

Vignette Give to charity Do not cheat Do not let friend drive home drunk Do not take money from another Leave note on car Return money to cashier

Anger 4

Empathy

Shame

Guilt

48

6

42 68

4

9

19

47

38

7

8

5

28

20

47

11

4

13

72

0

28

10

62

53

23

6

18

Do not have one-night stand

8

7

17

68

Tell parents you have done poorly

6

2

59

33

Say something to defend friend

Do not run red light

75

5

5

15

Total

21

18

17

44

2  744.81* *

p  0.05

which guilt has a tendency to be reported is in taking money from a household member (47%). Table 13.3 presents the means and standard deviations for each of the variables used in the analysis. We find that most respondents report doing the good, right, or normatively acceptable behavior across situations. A higher status other is usually not present in the situation. Equal status others/peers have a tendency to be present more than a higher status other. Most people (55%) experience these situations without a higher status other or equal status other present. In Table 13.4, we present the logit estimates for engaging in normative behavior. We find that men are less likely to report engaging in behavior that maintains the moral code than women. The odds of doing a “good” act decrease by 30% for men. Respondents are also less likely to engage in normative behavior when an equal status other is present. There is an almost 25% decrease in the odds of doing a “good” act when peers are present. One’s moral identity increases the odds of enacting normative behavior. A one-unit increase in the moral identity increases the odds of engaging in an act that is laced with moral codes by almost 10%. This supports Hypothesis 1. However, Hypothesis 2 is not supported. A more salient moral identity does not increase the odds of engaging in normative behavior. Table 13.5 provides the logit estimates for the moral emotions. We first point out that an earlier analysis revealed that when a norm is violated, those who report a high moral identity or a salient moral identity are not more likely to report moral emotions. These interaction effects were not significant across the

CH13-P374095.indd 242

5/26/2008 4:06:00 PM

The Moral Identity, Status, Moral Emotions, and the Normative Order

TABLE 13.3

243

Means and Standard Deviations of Variables

Variables

Mean

SD

N

Anger

0.04

0.20

5190

Empathy

0.04

0.19

5190

Shame

0.03

0.18

5190

Guilt

0.09

0.28

5190

Gender

0.41

0.49

5160

Race

0.22

0.42

4990

Income

0.48

0.50

5130

Moral identity

0.00

0.99

5160

Salient moral identity

0.00

0.85

5120

Normative behavior

0.62

0.48

2510

Higher status other

0.10

0.30

2533

Equal status other

0.39

0.49

2526

TABLE 13.4 Logit Estimates of Normative Behavior (N  2326) Independent variables Gender

Odds ratio 0.70*

Race

1.09

Income

0.92

Moral identity

1.09*

Salient moral identity

1.06

Higher status other

0.84

Equal status other

0.76*



2

35.69* p  0.05

*

emotions equations thereby disconfirming Hypothesis 3 and 4.1 Consequently, Table 13.5 omits these interaction effects from the logit analyses. When one looks across all of the equations, we find that one’s moral identity decreases the odds of reporting anger and shame by about 20%. Also looking across the four equations, we find that normative behavior is related to reports of moral emotions (Hypothesis 5), but not consistently across all of the equations (normative behavior does not significantly influence reports of empathy) and sometimes not in the manner we expected (the effect is in the opposite direction 1

These results are available upon request.

CH13-P374095.indd 243

5/26/2008 4:06:00 PM

244

Social Structure and Emotion

TABLE 13.5

Logit Estimates of Anger, Empathy, Shame, and Guilt (N  2326) Odds ratio

Independent variables

Anger

Empathy

Shame

Guilt

Gender

1.23

1.11

0.78

0.60*

Race

1.58*

0.70*

1.21

1.07

Income

0.84

0.71*

1.11

1.06

Moral identity

0.81*

1.13

0.80*

1.01

Salient moral identity

0.91

1.03

0.97

0.98

Normative behavior

4.50*

0.80

0.52*

0.12*

Higher status other

0.44*

0.59

3.60*

1.09

Equal status other

3.13*

1.22

0.48*

0.58*

17.47*

81.02*

276.12*

2

115.66* p  0.05

*

for anger). In the first column, we find that the odds of reporting anger increases at least four times when respondents behave normatively rather than counternormatively such as stopping at a red light rather than running through it or defending or protecting a friend rather than abandoning the friend. This is unexpected. Enacting normative behavior does reduce the moral emotions of shame and guilt. In column three, we see that normative behavior reduces reports of shame by almost 50%, and in column four, normative behavior reduces the expression of guilt by almost 90%. In general, these findings reveal that identity processes influence moral emotions not by way of a discrepancy between a high moral identity and salient moral identity and counter-normative behavior as anticipated in Hypotheses 3 and 4, but by other means. A higher moral identity directly reduces reports of anger and shame, and it indirectly reduces feelings of shame and guilt but increases feelings of anger through normative action. Absolute status and relative status influence moral emotions, but the effects are not uniform across all of the equations. For anger, when a respondent is white, the odds of reporting anger increase by almost 60%. Thus, Hypothesis 6 is supported for the status characteristic of race, but not for gender and income. The presence of a higher status actor reduces the expression of anger by almost 60%. This confirms Hypothesis 7. Additionally, we find that respondents are three times more likely to report anger in the presence of an equal status other. For empathy, when a respondent is white, the odds of reporting empathy decrease by 30% and having a high income decreases the odds of reporting empathy by almost 30%. Therefore, Hypothesis 8 is supported for race and income but not for gender.

CH13-P374095.indd 244

5/26/2008 4:06:00 PM

The Moral Identity, Status, Moral Emotions, and the Normative Order

245

None of the status characteristics influences reports of shame, thereby disconfirming Hypothesis 9. The presence of a higher status other makes respondents almost four times as likely to report this moral emotion. This confirms Hypothesis 10. Further, the presence of equal status others reduces reports of shame by over 50%. For guilt, when a respondent is male, the odds of reporting guilt decrease by 40%. No other status characteristics are significant. Consequently, Hypothesis 11 is supported for gender but not for race and income. The presence of a higher status actor does not increase reports of guilt. Thus, Hypothesis 12 is not supported. However, the presence of equal status others has the effect of decreasing feelings of guilt by over 40%.

DISCUSSION We have investigated how social psychological forces in the form of one’s moral identity and social structural forces in terms of absolute status and relative status facilitate or inhibit the expression of moral emotions. Both the moral identity and status characteristics of the self and the other in the situation directly influence reports of moral emotions and indirectly influence reports of moral emotions, through normative behavior. Further, it is the meanings of the moral identity that has an effect on moral emotions and not its structural feature, that is, whether the moral identity is salient given the host of other identities that one claims. Therefore, identity meanings and status characteristics of self and other coalesce to produce the expression of moral emotions in encounters. In Table 13.2, we found that anger was most likely to occur in the traffic situation of being stopped at a red light, and to a lesser degree, in situations having to do with helping a friend either by coming to the friend’s defense after something unkind is said about him or her or protecting the friend from harming himself or herself by not letting him or her drive home drunk. When we examine these situations more closely, we find that most anger in the traffic situation emerges following stopping at the red light (91%) rather than running through it. In the scenario defending the friend, 82% of the anger follows from defending the friend compared to not defending the friend. In addition, in the drunkdriving case, all of the anger follows from not letting the friend drive home drunk. What this suggests is that doing behavior that is laced with moral codes makes one upset. It may be that individuals’ internal focus such as the accomplishment of their own goals is interrupted, and they have to shift to an external focus as in attending to traffic rules or friendship codes. This interruption may be particularly frustrating and anger producing for individuals. The positive moral emotion of empathy primarily emerges not after giving to charity but after not giving to charity (73%). This is unexpected since empathy fosters helping behavior. This may be an anomaly because in all other situations in which empathy is also often reported, the moral emotion follows from having done a “good” act, for example, not letting one’s friend drive home drunk,

CH13-P374095.indd 245

5/26/2008 4:06:00 PM

246

Social Structure and Emotion

not taking money from another household member, or returning money to the cashier. However, respondents do not report empathy often. Modernity may have something to do with this finding (Clark 1997). With increasing residential mobility, urbanization, and pressures to compete and succeed, individuals may be less sensitive to the problems of others and not respond benevolently. Thus, as a positive moral emotion, empathy may be in short supply. This is unfortunate since empathy serves to build solidarity among social actors. Shame is often experienced when people perceive that they are being evaluated by others – when they have not lived up to the standards of important others (Barrett 1995). Consistent with this, the primary situation in which shame emerged in this study was in confessing to one’s parents about one’s poor school performance. When respondents confess, they are more likely to report shame (72%) than if they did not confess. Like empathy, respondents are unlikely to report the moral emotion of shame. Part of this low report may be due to the “low visibility of shame” (Scheff 1990). On the one hand, people do not want to admit shame and invite negative sanctions from others even if the feelings of shame are at low levels (Turner and Stets 2005). On the other hand, given that shame is a painful emotion, particularly at high levels, it may go unacknowledged or become repressed (Scheff 1987). Repression could lead to the “shame– anger” cycle discussed below (Scheff 1987). According to Scheff, when a person is negatively sanctioned in a situation and acknowledges shame, this can lead to greater interpersonal attunement in the situation. When a shamed person recognizes his or her failures in the situation, admits feelings of shame, and expresses the desire to behave alternatively in the future, such moves can repair the social bonds that have been fractured. However, when shame is not acknowledged, denied, or repressed, it can activate a shame–anger cycle that can be disruptive to social relationships. The anger emerges in response to having been shamed. The anger is directed either at the self because he or she has done a shameful thing or at another who has shamed the self. In response to the anger, one may feel further shame, perhaps feeling ashamed for being angry. Anger can then re-emerge. The shame–anger sequence can continue indefinitely and may increase in intensity over time leading to humiliated fury and violence. At the core, the self is experiencing an “emotional reaction to one’s emotional reactions” (Scheff 1987:112) that can produce an indefinite cycle of negative emotions in rapid succession. This makes smooth social relationships almost impossible to maintain. In this study, some of the anger that respondents may be reporting may be due to being locked in the shame–anger cycle. Of all of the moral emotions examined in this study, guilt is the most frequently reported. In guilt, one is concerned with how one’s behavior potentially harms another. Consistent with this, our results revealed that guilt generally occurs in response to not having done a “good” act as in not giving to charity (98%), cheating (89%), not leaving a note after denting another’s car (77%), and not returning money to the cashier (95%). There is one exception. When

CH13-P374095.indd 246

5/26/2008 4:06:00 PM

The Moral Identity, Status, Moral Emotions, and the Normative Order

247

respondents report not engaging in a “one-night stand,” they are more likely to report guilt (61%) than if they did commit the “one-night stand.” Perhaps some of the guilt in this situation is due to individuals thinking about the possibility of a “one-night stand.” Indeed, specific behaviors do not have to be committed for one to feel guilty (Morris 1987). For example, one can feel guilty for wishing harm to another (guilt from a “state of mind”), for benefiting from something that cannot be defended as fair (fortuitous and unjust enrichment), or from thinking of the wrongful acts that have been enacted by others with whom one identifies (vicarious guilt). In general, the results of this research reveal that guilt is reported the most and shame is reported the least. Researchers have studied whether people in the United States experience more shame or guilt compared to other countries. In a cross-cultural study in 37 countries, Wallbott and Scherer (1995) found that in collectivist cultures such as Brazil, Mexico, and Greece shame appeared to be the more predominant emotion while in individualist societies, guilt seemed to be the more common emotion. The researchers point out that in culture where the Protestant ethic of personal responsibility prevails likely encourages feelings of guilt compared to shame. Alternatively, shame may be more repressed in individualistic societies than collectivistic societies. Whether guilt is the moral emotion because it fosters responsible, normative, moral behavior while shame promotes illicit, self-destructive behavior (Tangney 2003; Tangney and Dearing 2002) is not clear. At low levels of intensity, shame can serve to keep behavior in line with normative expectations as well. In examining the identity process and status process in this research, we found that both played a role in influencing reports of moral emotions. And for both processes, the role is direct and indirect. For the moral identity, having a high moral identity decreases reports of anger and shame. Additionally, having standards of a “good” person influences “good” acts and good acts induce less self-critical emotions (shame and guilt), but more other-critical emotions (anger). Interestingly, those who engage in “good” acts are more likely to be women than men, and women report a higher moral identity than men. The higher status of men compared to women may make them feel more entitled to violate social expectations. In other words, they can behave counter-normatively because they have the power to get away with it. Absolute status in terms of gender, race, and income directly influence the expression of moral emotions. However, their effects are not consistent across the four moral emotions we studied. High status actors (specifically, whites) are more likely to report the other-critical moral emotion of anger. Their sense of entitlement given their high status may make them more sensitive to violations of their rights and freedoms. Low status actors (nonwhites and persons with a low income) are more likely to report the other-suffering emotion of empathy, and low status actors (women) are more likely to report the self-critical emotion of guilt. The moral emotions that low status actors report are suggestive of them being more sensitive to others in situations. Tangney and Dearing (2002)

CH13-P374095.indd 247

5/26/2008 4:06:00 PM

248

Social Structure and Emotion

point out that empathy and guilt work together to produce positive and rewarding interpersonal relationships. Both empathy and guilt direct one’s attention to the other in the situation rather than the self, as in the case of shame. While guilt is a self-critical emotion like shame, guilt is originally brought on by seeing the effects that one’s behavior has on another. Empathy helps foster this orientation. Low status actors may be more likely to report more empathy and guilt than high status actors because they do not have power and thus need to remain attuned to the other in the situation. Relative status influences reports of anger and shame. Specifically, the presence of a higher status actor reduces the other-directed emotion of anger, but in the presence of peers, this anger is increased. Lower status actors are less likely to direct negative affect, particularly anger, at higher status actors because they may fear retaliation from the higher status actor (Ridgeway and Johnson 1990). This fear is unlikely to be present when peers are in the situation. In fact, peers may be viewed as the source of one’s goals being blocked. If higher status actors inhibit anger, they facilitate feelings of shame; and if peers facilitate feelings of anger, they inhibit shame. The presence of a higher status actor likely reminds lower status actors as to the standards that he or she must live up to. This selfevaluation may activate feelings of inadequacy and worthlessness vis-à-vis the higher status other. However, given that we live in a society where shame is often unacknowledged (Scheff 1990), when peers are present, individuals may feel ashamed of their shame, and others may respond with negative sanctions of their own. All of this would add to their existing painful feeling of shame, making it more likely that such shame will be repressed and go unreported (Turner and Stets 2006). In general, a higher status actor facilitates feelings of shame in situations, but if these feelings get repressed it can lead to anger which a lower status actor is unlikely to express in the presence of a higher status actor. Peers appear to encourage anger, but inhibit shame. Peers also inhibit reports of guilt. Thus, peers are discouraging reports of the self-critical emotions in respondents. This is disheartening since shame and guilt have traditionally been viewed as “the moral emotions.” If these moral emotions are less likely to be activated in the presence of peers, then it suggests that individuals fear negative sanction from them.

CONCLUSION The best way to see the moral order at the microlevel is to observe individuals’ moral responses when they violate normative expectations. It is important that these moral emotions are routinely expressed in situations because they tell the self and others what the self values. And, reaffirming what the self values helps to sustain the larger moral order within which the self and his or her standards are embedded. In this research, we found that both identity processes and status processes contribute to the expression of moral emotions in encounters.

CH13-P374095.indd 248

5/26/2008 4:06:00 PM

The Moral Identity, Status, Moral Emotions, and the Normative Order

249

Actors’ identity standards guide moral behavior in situations, but simultaneously operating in situations are status and power inequalities that correspond to the distribution of expressed moral emotions. High status actors are more likely to express anger, perhaps reflecting greater sensitivity to their own goals, rights, and freedoms, while low status actors are more likely to express empathy and guilt, suggesting greater sensitivity to others in the situation. Further, higher status actors in a situation are less likely to witness anger by lower status actors, but they are more likely to witness shame by lower status actors. Equal status actors appear to encourage other-critical moral emotions such as anger, but they discourage self-critical moral emotions such as shame and guilt. The latter finding suggests that peers may weaken the self-evaluative process that is important in assessing whether one is responsible for “bad” outcomes in a situation. The self-referential nature of moral emotions such as shame and guilt help in the development of the self to the extent that people learn about themselves, including modifying future actions. Ultimately, feeling moral emotions reminds actors of society’s standards vis-à-vis one’s own actions. The moral emotions monitor one’s own behavior when others are not able to monitor it for the person. This makes social order, the moral order, possible.

ACKNOWLEDGMENTS We would like to thank the Academic Senate of the University of California, Riverside, which supported the research for this paper to the first author. We also would like to thank the members of the Social Psychology Research Seminar at the University of California, Riverside for their comments on an earlier draft of this paper.

REFERENCES Aquino, Karl and Americus II Reed. 2002. “The Self-Importance of Moral identity.” Journal of Personality and Social Psychology 83:1423–1440. Averill, James R. 1982. Anger and Aggression: An Essay on Emotion. New York: Springer-Verlag. Barrett, Karen Caplovitz. 1995. “A Functionalist Approach to Shame and Guilt.” In: J. P. Tangney and K. W. Fischer (eds), Self-Conscious Emotions: The Psychology of Shame, Guilt, Embarrassment, and Pride, pp. 25–63. New York: The Guilford Press. Berger, Joseph, Bernard P. Cohen, and Morris Zelditch, Jr.. 1972. “Status Characteristics and Social Interaction.” American Sociological Review 37:241–255. Burke, Peter J. 1991. “Identity Processes and Social Stress.” American Sociological Review 56:836–849. Burke, Peter J. 2004. “Identities and Social Structure: 2003 Cooley–Mead Award Address.” Social Psychology Quarterly 67:5–15. Burke, Peter J. and Donald C. Reitzes. 1991. “An Identity theory Approach to Commitment.” Social Psychology Quarterly 54:239–251. Burke, Peter J. and Judy Tully. 1977. “The Measurement of Role/Identity.” Social Forces 55:881–897.

CH13-P374095.indd 249

5/26/2008 4:06:00 PM

250

Social Structure and Emotion

Carter, Michael J. 2005. “The Moral identity and Group Membership.” Master’s Thesis, University of California, Riverside. Clark, Candace. 1997. Misery and Company: Sympathy in Everyday Life. Chicago: University of Chicago Press. Conway, Michael, Roberto Di Fazio, and Shari Mayman. 1999. “Judging Others’ Emotions as a Function of the Others’ Status.” Social Psychology Quarterly 62:291–305. Davis, Mark H. 2006. “Empathy.” In: J. E. Stets and J. H. Turner (eds), Handbook of the Sociology of Emotions, pp. 443–466. New York: Springer. Ellestad, June and Jan E. Stets. 1998. “Jealousy and Parenting: Predicting Emotions from Identity Theory.” Sociological Perspectives 41:639–668. Franks, David D. 1985. “Role-Taking, Social Power and Imperceptiveness: The Analysis of Rape.” Studies in Symbolic Interaction 6:229–259. Gilligan, Carol. 1982. In a Different Voice: Psychological Theory and Women’s Development. Cambridge, MA: Harvard University Press. Gordon, Steven L. 1989. “The Socialization of Children’s Emotions: Emotional Culture, Competence, and Exposure.” In: C. Saarni and P. L. Harris (eds), Children’s Understanding of Emotion, pp. 319–349. Cambridge: Cambridge University Press. Gordon, Steven L. 1990. “Social Structural Effects on Emotions.” In: T. D. Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 145–179. Albany, NY: State University of New York Press. Haidt, Jonathan. 2003. “The Moral Emotions.” In: R. J. Davidson, K. R. Scherer, and H. H. Goldsmith (eds), Handbook of Affective Sciences, pp. 852–870. New York: Oxford University Press. Hansen, Gary L. 1991. “Jealousy: Its Conceptualization, Measurement, and Integration with Family Stress Theory.” In: P. Salovey (ed.), The Psychology of Jealousy and Envy, pp. 211–230. New York: The Guilford Press. Heise, David R. and Cassandra Calhan. 1995. “Emotion Norms in Interpersonal Events.” Social Psychology Quarterly 58:223–240. Hoffman, Martin L. 2000. Empathy and Moral Development: Implications for Caring and Justice. New York, NY: Cambridge University Press. Jaffee, Sara and Janet Shibley Hyde. 2000. “Gender Differences in Moral Orientation: A MetaAnalysis.” Psychological Bulletin 126:703–726. Kemper, Theodore D. 1987. “How Many Emotions are There? Wedding the Social and Autonomic Components.” American Journal of Sociology 93:263–289. Kohlberg, Lawrence. 1981. The Philosophy of Moral Development. San Francisco: Harper and Row. Kohlberg, Lawrence, Charles Levine, and Alexandra Hewer. 1983. “Moral Stages: A Current Formulation and a Response to Critics.” Contributions to Human Development 10:174. Morris, Herbert. 1987. “Nonmoral Guilt.” In: F. Schoeman (ed.), Responsiblity, Character, and the Emotions, pp. 220–240. New York: Cambridge University Press. Ridgeway, Cecilia L. and Cathryn Johnson. 1990. “What Is the Relationship between Socioemotional Behavior and Status in Task Groups?” American Journal of Sociology 95:1189–1212. Rozin, Paul, Laura Lowery, Sumio Imada, and Jonathan Haidt. 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. Scheff, Thomas J. 1987. “The Shame–Rage Spiral: A Case Study of an Interminable Quarrel.” In: H. B. Lewis (ed.), The Role of Shame in Symptom Formaqtion, pp. 109–149. Hillsdale, NJ: Lawrence Erlbaum. Scheff, Thomas J. 1990. “Socialization of Emotion: Pride and Shame as Causal Agents.” In: T. D. Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 281–304. Albany: State University of New York. Schieman, Scott. 2006. “Anger.” In: J. E. Stets and J. H. Turner (eds), Handbook of the Sociology of Emotions, pp. 493–515. New York: Springer.

CH13-P374095.indd 250

5/26/2008 4:06:00 PM

The Moral Identity, Status, Moral Emotions, and the Normative Order

251

Shields, Stephanie A., Dallas N. Garner, Brooke Di Leone, and Alena M. Hadley. 2006. “Gender and Emotion.” In: J. E. Stets and J. H. Turner (eds), Handbook of the Sociology of Emotions, pp. 63–83. New York: Springer. Stets, Jan E. 2004. “Emotions in Identity Theory: The Effects of Status.” Advances in Group Processes 21:51–76. Stets, Jan E. 2006. “Identity Theory and Emotions.” In: J. E. Stets and J. H. Turner (eds), Handbook of the Sociology of Emotions, pp. 203–223. New York: Springer. Stets, Jan E. and Chris F. Biga. 2003. “Bringing Identity Theory into Environmental Sociology.” Sociological Theory 21:398–423. Stets, Jan E. and Peter J. Burke. 1996. “Gender, Control, and Interaction.” Social Psychology Quarterly 59:193–220. Stets, Jan E. and Peter J. Burke. 2005. “New Directions in Identity Control Theory.” Advances in Group Processes 22:43–64. Stets, Jan E. and Michael J. Carter. 2006. “The Moral Identity: A Principle Level Identity.” In: K. McClelland and T. J. Fararo (eds), Purpose, Meaning, and Action: Control Systems Theories in Sociology, pp. 293–316. New York: Palgrave. Stryker, Sheldon. 1980. Symbolic Interactionism: A Social Structural Version. Menlo Park, CA: Benjamin Cummings. Stryker, Sheldon and Peter J. Burke. 2000. “The Past, Present, and Future of an Identity Theory.” Social Psychology Quarterly 63:284–297. Stryker, Sheldon and Richard T. Serpe. 1994. “Identity Salience and Psychological Centrality: Equivalent, Overlapping, or Complementary Concepts?” Social Psychology Quarterly 57:16–35. Tangney, June Price. 1991. “Moral Affect: The Good, the Bad, and the Ugly.” Journal of Personality and Social Psychology 61:598–607. Tangney, June Price. 2003. “Self-Relevant Emotions.” In: M. R. Leary and J. P. Tangney (eds), Handbook of Self and Identity, pp. 384–400. New York: The Guilford Press. Tangney, June Price and Rhonda L. Dearing. 2002. Shame and Guilt. New York: The Guilford Press. Turner, Jonathan H. 2000. On the Origins of Human Emotions: A Sociological Inquiry into the Evolution of Human Affect. Stanford: Stanford University Press. Turner, Jonathan H. 2006. “Psychoanalytic Sociological Theories and Emotion.” In: J. E. Stets and J. H. Turner (eds), Handbook of the Sociology of Emotions, pp. 276–294. New York: Springer. Turner, Jonathan H. and Jan E. Stets. 2005. The Sociology of Emotions. New York: Cambridge University Press. Turner, Jonathan H. and Jan E. Stets. 2006. “Moral Emotions.” In: J. E. Stets and J. H. Turner (eds), Handbook of the Sociology of Emotions, pp. 544–566. New York: Springer. Walker, Lawrence J. and Karl H. Hennig. 2004. “Differing Conceptions of Moral Exemplarity: Just, Brave, and Caring.” Journal of Personality and Social Psychology 86:629–647. Wallbott, Harald G. and Klaus R. Scherer. 1995. “Cultural Determinants in Experiencing Shame and Guilt.” In: J. P. Tangney and K. W. Fischer (eds), Self-Conscious Emotions: The Psychology of Shame, Guilt, Embarrassment, and Pride, pp. 465–487. New York: The Guilford Press.

CH13-P374095.indd 251

5/26/2008 4:06:00 PM

PA R T

V Emotion Labor

CH14-P374095.indd 253

5/26/2008 4:07:31 PM

14 Commentary Christopher D. Moore Lakeland College, Sheboygan, WI, USA

Sociologists have long recognized that emotions are shaped by social context and the rituals of social interaction (Collins 1975, 1981, 2004; Durkheim [1912] 1965; Goffman 1959, 1967; Kemper 1991; Kemper and Collins 1990; Scheff 1988). Yet, a coherently integrated sociology of human emotions, particularly in relation to emotion labor, has yet to emerge from the various theoretical approaches that have dominated the past several decades (for reviews, see Smith-Lovin 1995; Smith-Lovin and Molm 2000; Stets 2003; Turner and Stets 2006). While previous chapters in this volume have focused on the relationships between emotion and processes of power, status, justice, and self/identity, the two papers in this section reflect the ongoing shift in focus of research on emotion labor to one that considers the distinctive, yet interconnected roles of local culture, social structure, and emotion management/emotion work. This increased and nuanced attention to social context, particularly when emotion work occurs in the workplace (e.g., emotion labor), is especially evident in Erickson’s chapter. In addition, the chapter by Francis and Heise offers an effective illustration of what occurs when poor African American patents’ expectations for personal interactions with caregivers differs, or is perceived to differ, from those held by health care professional in their community. Hochschild’s groundbreaking work in the 1970s and 1980s led sociologists to better understand the normative emotional structures and socialization processes of everyday life, and the workplace in particular (Hochschild 1979, 1983a,b). Hochschild argues that the experience of emotion, as well as the norms that define what we should feel and express, are situated in particular cultural and structural settings. Both the emotions we feel, and the communication of emotions though social displays, informs us and others of our relative places in situations. Emotions are thusly managed through contextually contingent norms for what one should feel as well as what one should express. The sometimes Social Structure and Emotion

CH14-P374095.indd 255

255

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 4:07:31 PM

256

Social Structure and Emotion

dual nature of social structure and cultural content is emphasized by Erickson’s efforts at reframing this discourse as one concerned with how these combine to create gendered and racialized performance expectations for emotional labor performances that in turn produce different emotional experiences across and within the same structural positions. Hochschild’s notion of “feeling rules” and “expression rules” refers to the norms pertaining to the appropriate feelings and behaviors in which one should engage given certain situations and actors (also called “display rules” (Ekman, 1972)). Individuals manage their emotions though “emotion work” (or “emotion labor” if in the context of one’s vocation) in order to appear to others as though they are experiencing the normative emotion (even if they are not). Workers, particularly those employed in customer care/service industries, are often expected to engage in emotion labor as part of their job – often without explicit compensation. Such emotional displays are commonly intended to satisfy the expectations of clients, customers, and organizational executives concerned about the opinions (and patronage) of these groups. This kind of emotion work often takes the form of “surface acting” to merely display appropriate emotions (even if the person doesn’t actually feel those emotions). “Deep acting,” on the other hand, is a more laborious process through which a person attempts to additionally bring his emotional state in-line with what he perceives is the expected behavioral expression. While the display of unfelt emotions (faking) can be generalized as work performed in “bad faith” (surface acting) or “good faith” (deep acting), these acts may be laden with meanings that can affect individuals’ self-concept and perceptions of authenticity. Thoits (1990) demonstrates the power of such normative structures by suggesting that individuals who recognize that their “true” emotions are chronically mismatched to the emotions they know they ought to be feeling, may come to believe they are (or become) mentally ill. Further, the effects of deep acting can sometimes be immediate as it may be sometimes necessary to enable workers to overcome anger, fear, or even empathy, to adequately perform their other job-related tasks. Relative structural position of actors is particularly relevant in workplace emotion management as the local social context (cultural content and relative structural position of actors) provides for the interpretation of the exchange value of the emotional labor performed by individuals (who is expected to perform this work, and what if any compensation is provided). Hegtvedt (1990) argues that actors in power-advantaged positions generally experience more positive emotions than do actors in power-disadvantaged positions by virtue of their enhanced ability to achieve and/or maintain power and status (see also Molm 1997; Willer, Lovaglia, and Markovsky 1997). Collins (1990) takes this work one step further by loosely incorporating Durkheim’s notion of “collective effervescence” to offer another reason higher status actors are likely to experience more positive emotions than lower status actors. Specifically, Collins argues that higher status actors draw their power and status from the existence and success of the social structure in which they are embedded. Because of this, higher status actors have

CH14-P374095.indd 256

5/26/2008 4:07:31 PM

257

Commentary

a greater stake in whether the organization succeeds or fails. Collins argues that this fact, coupled by the fact that higher status actors (e.g., executives, managers, foremen) are in positions to achieve compliance from lower status actors (workers), creates situations in which higher status actors can enjoy knowing that they are responsible for the success of the organization and subsequently the ability to ensure and reinforce their power and status positions. Of course this argument may be reversed to suggest that when organizations fails to achieve their goals, higher status actors may feel (and be expected to feel) more intense negative emotions because they failed the organization and themselves. Ridgeway and Johnson (1990) show that not only do higher status individuals experience more positive emotions because they are high status, but if a lower status group member gives a higher status group member negative feedback (i.e., he doesn’t’ agree with one of the higher status member’s suggestions), the high status member will direct negative emotion towards that group member to “remind him of his (lower) place.” Interestingly, other recent work suggests that emotion structures can affect back onto status structures. Specifically, Shelly (2001) argues that sentiment structures (who initially likes/dislikes whom in a group), can influence performance expectations (“likes” are viewed as more competent than “dislikes”), which can in turn affect the resulting status structure. Shelly’s (2001) work reflects a recent trend characteristic of a more integrated constructionist/structural approach towards emotion management. Control process theories such as affect control theory, place comparatively greater emphasis on human agency, and posit that individuals actively manage social situations based on (1) internalized culturally rooted sentiments, and (2) emergent localized impressions, by choosing to act and/or interpret events in ways that facilitate equilibrium between these meanings for self and other(s)’ identities, behaviors, and the setting (Heise 2007; Robinson 2007). The research presented in the following chapters point towards the gains that are possible when efforts are made to combining perspectives sensitive to different aspects of the reciprocal nature of internalized cultural content, structural position, subjective experience of emotion (feeling and expression) rules, and the implications for “surface” and “deep” acting emotion performances.

REFERENCES Collins, Randall. 1975. Conflict Sociology: Toward an Explanatory Science. Academic Press. Collins, Randall. 1981. “On the Microfoundations of Macrosociology.” American Journal of Sociology 86:984–1010. Collins, Randall. 1990. “Stratification, Emotional Energy, and the Transient Emotions.” In T.D. Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 27–57. Albany, NY: SUNY Press. Collins, Randall. 2004. Interaction Ritual Chains. Princeton, NJ: Princeton University Press. Durkheim, Emile. [1912] 1965. The Elementary Forms of Religious Life. New York: Oxford University Press.

CH14-P374095.indd 257

5/26/2008 4:07:31 PM

258

Social Structure and Emotion

Ekman, Paul. 1972. “Universals and Cultural Differences in Facial Expressions of Emotion.” In: J. Cole (ed.), Nebraska Symposium on Motivation 1971, Vol. 19, pp. 207–283. Lincoln, NE: University of Nebraska Press. Goffman. 1959. The Presentation of Self in Everyday Life. Anchor Books: New York. Goffman. 1967. Interaction Ritual. New York: Pantheon Books. Hegtvedt, Karen A. 1990. “The Effects of Relationship Structure on Emotional Responses to Inequity.” Social Psychology Quarterly 53:214–228. Heise, D. R. 2007. Expressive Order: Confirming Sentiments in Social Actions. New York: Springer. Hochschild, Arlie R. 1979. “Emotion Work, Feeling Rules, and Social Structure.” American Journal of Sociology 85:551–575. Hochschild, Arlie R. 1983a. “Comment on Kemper’s Social Constructionist and Positivist Approaches to the Sociology of Emotions.” American Journal of Sociology 89:432–434. Hochschild, Arlie R. 1983b. The Managed Heart: Commercialization of Human Feeling. Berkeley, CA: University of California Press. Kemper, Theodore D. 1991. “Predicting Emotions from Social Relations.” Social Psychology Quarterly 54:330–342. Kemper, T. David and Randall Collins. 1990. “Dimensions of Microinteraction.” American Journal of Sociology 96:32–68. Molm, Linda D. 1990. “Structure, Action and Outcomes: The Dynamics of Power in Social Exchange.” American Sociological Review 55:427–447. Molm, Linda D. 1997. “Risk and Power Use: Constraints on the Use of Coercion in Exchange.” American Sociological Review 62:113–133. Ridgeway, Cecilia and Cathryn Johnson. 1990. “What is the Relationship Between Socio-emotional Behavior and Status in Task Groups?” American Journal of Sociology 95:1189–1212. Robinson, Dawn T. 2007. “Control Theories in Sociology.” Annual Review of Sociology 33:157–174. Scheff, Thomas J. 1988. “Shame and Conformity: The Deference-Emotion System.” American Sociological Review 53:395–406. Shelly, Robert K. 2001. “How Performance Expectations Arise from Sentiments.” Social Psychology Quarterly 64:72–87. Smith-Lovin, Lynn. 1995. “The Sociology of Affect and Emotion.” In: K. Cook, G. A. Fine, and J. S. House (eds), Sociological Perspectives on Social Psychology, pp. 118–148. Boston: Allyn & Bacon. Smith-Lovin, Lynn and Linda Molm. 2000. “Introduction to the Millennium Special Issue on the State of Sociological Social Psychology.” Social Psychology Quarterly 63(4):281–283. Stets, Jan E. 2003. “Emotions and Sentiments.” In: J. D. Delameter (ed.), Handbook of Social Psychology, pp. 309–335. New York: Kluwer/Plenum. Thoits, Peggy A. 1990. “Emotional Deviance: Research Agendas.” In: T. D. Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 180–203. Albany, NY: SUNY Press. Turner, John H. and Jan E. Stets. 2006. “Sociological Theories of Human Emotions.” Annual Review of Sociology 32:25–52. Willer, David, Michael Lovaglia, and Barry Markovsky. 1997. “Power and Influence: A Theoretical Bridge.” Social Forces 76:571–603.

CH14-P374095.indd 258

5/26/2008 4:07:31 PM

15 The Context of Care: Reconsidering Culture, Structure, and the Performance of Emotional Labor Among Registered Nurses Rebecca J. Erickson University of Akron, Akron, OH, USA

ABSTRACT The emotion-management perspective first introduced by Hochschild (1979; 1983b) has been consistently characterized as a “cultural” approach. This classification is not surprising given the emphasis placed on how emotion norms and feeling rules enter into, and help to construct, emotional experience. In this chapter, however, I reconsider the cultural label in light of Hochschild’s original theoretical formulation and recent sociological theorizing about the structure–culture distinction. In so doing, I propose that theory and research on emotion, and particularly emotion management, would benefit from reframing the approach as “contextual” as opposed to cultural or structural. Through an analysis of the emotion management experiences of 1158 registered nurses, I provide support for this proposal by showing how structure (e.g., gender, race, job structure) and culture (e.g., feeling rules) combine to differentially affect nurses’ experience of emotion and their performance of both surface and deep Social Structure and Emotion

CH15-P374095.indd 259

259

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 4:08:59 PM

260

Social Structure and Emotion

acting. Consistent with Hochschild’s original statement that the emotion-management perspective be used as a lens through which to examine social structure, as well as self and interaction, the results provide new insights into the gendered and racialized performance of emotional labor and the multidimensional context that shapes emotional experience. Key Words: Emotional Labor, Emotion, Nursing, Culture, Social Psychology

Forgetfulness of the questions that are the starting points of inquiry leads us to ignore the substantive assumptions “buried” in our concepts and commits us to a one-sided view of reality. —Dennis Wrong (1961)

In their introduction to the Millennium Special Issue of Social Psychology Quarterly, Smith-Lovin and Molm (2000:281) noted that a number of papers pointed to “ways in which sociological social psychology should reconnect with the fundamental questions and interests of sociology as a whole.” Hollander and Howard (2000) further observe that because social psychology has proceeded relatively independently from sociology as a whole, the broader discipline has not benefited as much as it might from social psychological insight (also Collins 2004:103). Although some of the blame for this lack of integration should certainly be placed on non-social psychological sociologists, I hope to advance further discussion of these issues by exploring one potential “fault line” located within emotional labor theory and research. Specifically, I suggest that we need to reconsider the “cultural” label consistently attached to studies of emotional labor. In so doing, I illustrate some of the ways that our acceptance of this label may have led to the relative neglect of Hochschild’s (1979) initial points of inquiry and, as a result, may have limited the ability of emotional labor researchers to ask the questions and pursue the answers that ground our identities in sociological social psychology. To explore this possibility further, I empirically examine the extent to which diffuse status characteristics and indicators of proximate occupational structure and culture influence the experience and management of emotion among registered nurses. The findings of this study suggest that social scientists’ ability to understand the effects of the context of care requires that more systematic research attention be paid to the ways in which proximate structures, as well as diffuse status characteristics, affect the performance of emotional labor.

COMMON POINTS OF INQUIRY: RECONSIDERING THE STRUCTURE–CULTURE DISTINCTION There are no issues more central to sociology than those of social reproduction and social change. Understanding the simultaneous operation of these

CH15-P374095.indd 260

5/26/2008 4:08:59 PM

Reconsidering culture, structure, and emotional labor

261

processes has motivated sociological theory and research since the discipline’s inception and continues to motivate much of it today (Calhoun, Rojek, and Turner 1995; Comte 1896). If sociologists charge their social psychological colleagues with not addressing the central questions of the discipline, then the fault must lie with our ability to further social scientific understanding of reproduction and change. Sociologists of emotion know this is far from the case. From the earliest statements in the 1970s to the American Sociological Association Emotion Section’s most recent contribution awards, emotions scholars have been centrally concerned with examining how emotions serve to both bring people together and tear them apart or how emotions figure into processes of social reproduction and social change (Collins 2004; Turner and Stets 2005; see also Chapter 18 by Turner). Nonetheless, it is open for debate whether emotion researchers have consistently characterized their contributions in such general sociological terms (Shilling 2002). To be sure, one can find a number of recent claims regarding the general sociological importance of emotion that make use of the dualisms commonly found in more general theoretical statements. Barbalet (2002:3) identifies that the statement “emotions link structure and agency” serves as a crucial sociological proposition. Shilling (2002:11) also notes that emotions are central to “the key sociological issues of social order and action,” and Collins (2004:103) argues that emotions hold society together and mobilize conflict (also see Chapter 18 by Turner). The use of dualisms such as agency–structure and action–order clearly parallel statements found in the writings of social theorists more generally. The tendency to frame the micro–macro relation in this way parallels the work of social theorists in another way as well – that is, in the pronounced absence of references to culture (Hays 1994). I raise this point not because I think sociologists of emotion are not attentive to culture (quite the opposite), but merely to raise the possibility that examining how social theorists have framed the structure–culture–agency relation may provide unique insights into the strengths and weaknesses of how emotions scholarship has developed over the past three decades. The development of separate research strands within the sociology of emotions has certainly led to theoretical and empirical gains (Smith-Lovin 2005). However, this developmental approach has also led to some unintended negative consequences. Instead of being far removed from central sociological concerns, however, these developments parallel recent theoretical trends, outcomes, and concerns within sociology (and sociological social psychology) as a whole. Classic sociological theorists often emphasized the role of social structure in understanding processes of social reproduction and social change (Fligstein 2001). Concisely captured by Wrong’s (1961) depiction of the “oversocialized conception of man in modern sociology,” this structural emphasis remained a prominent feature of much sociological theory produced during the twentieth century; its dominance finally sparking theoretical movements explicitly aimed at reclaiming the power of individual agency (Berger and Luckmann 1966; Blumer

CH15-P374095.indd 261

5/26/2008 4:08:59 PM

262

Social Structure and Emotion

1969). During the past two decades, social theorists have spent a great deal of time demonstrating that these countervailing tendencies are not new (Giddens 1991; Hewitt 1989). Nonetheless, there remains little theoretical consensus about how to connect agents and structures, or even on how best to conceptualize the terms of the debate (Alexander et al. 1987; Coleman 1986; Fligstein 2001; Huber 1991; Sewell 1992). Part of the problem may be due to assumptions underlying how structure and agency are conceptualized and to how both of these concepts are theorized in regard to culture. Hays (1994) argues, for example, that writers often define social structure through a series of contrasts with both agency and culture. Because of this definitional tendency, the interconnections between the concepts tend to be lost and the differences between them heightened. Structure is systematic, patterned, constraining, static, and collective while agency is contingent, random, active, and individual. In maintaining a tradition of distinguishing culture from structure (Hays 1994; MacLeod and Lively 2003), theorists simultaneously reinforce a privileged image of structure and structural accounts by characterizing them as objective, material, external, and available for scientific observation. In contrast, these same discourses portray culture as subjective, internal, interpretive, and operating as a weaker counterpart to the more solid and scientific structural approaches. This tendency within more general social theory is similar to the discursive process found within the sociology of emotions. A process that, in seeking to specify the scope of its subject matter over the past three decades, has reinforced the distinctions between approaches rather than seeking out points of overlap. Paralleling similar changes within social theory (Hays 1994; Rubinstein 2001) and sociological social psychology (MacLeod and Lively 2003), emotions scholars have started to break down some of the assumed divisions between approaches (Lawler and Thye 1999; Simon and Nath 2004; Turner and Stets 2005). The motivation to move forward with deconstructing presumed differences, along with the development of more synthetic approaches, may not continue, however, if the empirical utility and theoretical significance of such endeavors is not forthcoming. In what follows, I reconsider theory and research related to emotion management to illustrate some of the theoretical, methodological, and empirical implications of the cultural label having been accepted as a sufficient characterization of Hochschild’s (1979, 1983b) original points of inquiry. In then analyzing data on the emotional labor performed by registered nurses, I conduct the first quantitative examination of the structural and cultural correlates of surface and deep acting. These results indicate that further efforts must be made to understand the structural antecedents of emotion management processes and provide further support for Hochschild’s (1990) statement that how we manage emotion can provide unique insights into how larger systems of stratification are sustained and changed.

CH15-P374095.indd 262

5/26/2008 4:08:59 PM

Reconsidering culture, structure, and emotional labor

263

FROM STRUCTURE TO CULTURE AND BACK AGAIN: (RE)FRAMING EMOTION MANAGEMENT RESEARCH Over the past two decades, every author reviewing the sociology of emotions has characterized the emotion-management perspective as “cultural” (Kemper 1993; Smith-Lovin 1995; Stets 2003; Thoits 1989; Turner and Stets 2005). This label has gone largely unchallenged despite the fact that Hochschild entitled her first major piece on the subject “Emotion Work, Feeling Rules, and Social Structure” (1979, emphasis added) and began The Managed Heart (1983b) with an example drawn from Marx’s Das Kapital. In remarks commemorating the twentieth anniversary of the publication of The Managed Heart, Smith-Lovin (2003) also suggests that Hochschild’s most important contribution may lie in having shown us how emotions are cultural phenomenon. However, SmithLovin goes on to note that sociologists of emotion have failed to develop the full promise of Hochschild’s insights. Specifically, emotions scholars have largely failed to study “social structural sources” of variation in emotion management and emotional labor (Smith-Lovin 2003:7; also see Gibson 1997). As suggested above, one reason for this failure may lie with the relatively unquestioned acceptance of the claim that studying the performance of emotion management is a cultural enterprise (Copp 1998; Gordon 1990).1 While adopting a cultural approach would not seem to preclude an examination of structural variation or influence, the theoretical and empirical meanings implied by the cultural label (i.e., internal, hidden, requiring interpretation) may have limited research on emotion management in just this way. As Smith-Lovin (2003:7) observes, we have learned a great deal about emotion culture but we have not moved past Hochschild’s “initial Marxist-inspired insight – that emotional labor exists – to find out more about how it is related to social structure.” This gap in our understanding is somewhat surprising given Hochschild’s (1983a) insistence that her approach is concerned with the effects of social structure. Hochschild consistently expressed this view across more than a decade of written work on emotion management (Hochschild 1979, 1983b, 1990). She clearly emphasized this point in 1990 when identifying the path that future researchers ought to pursue, stating that connecting studies of emotion to broader sociological concerns of reproduction and change requires the examination of how diffuse status characteristics (e.g., gender, race, and class) relate to individual 1

As Lively (2000) notes, emotions theorists have been more likely to discuss the importance of structural influences on emotion management processes. However, these theoretical statements have not translated into structural characteristics being investigated as direct influences on surface or deep acting processes. For example, studies examining the antecedents of emotional labor within both sociology and organizational psychology have emphasized the influence of culture (i.e., emotion norms or rules) rather than structure (Diefendorff, Croyle, and Gosserand 2005; Grandey 2000; Lively 2000; Morris and Feldman 1996).

CH15-P374095.indd 263

5/26/2008 4:08:59 PM

264

Social Structure and Emotion

strategies for action and feeling (Hochschild 1990; also see Collins 2004). Despite the consistency of her statements, authors describing Hochschild’s perspective have downplayed this facet of her work in favor of emphasizing the role of cultural beliefs, emotion norms, and feeling and display rules guide the emotion management process (Smith-Lovin 1995; Thoits 1989; Turner and Stets 2005; cf. Thoits 2004). Given Hochschild’s statements, the cultural label attached to the perspective inspired by her work is at best incomplete. This situation also leads me to wonder whether the emphasis placed on emotion norms, along with the relative neglect of structural influences, has more to do with the theoretical and methodological implications of the cultural label than with truly believing that this emphasis accurately reflects the social context within which emotion management occurs. Paralleling a tendency that MacLeod and Lively (2003:82) find in sociological social psychology more generally – and quite consistent with the foundational debates of the 1980s (Hochschild 1983a; Kemper 1981) – emotion management scholars have tended to take an either/or approach to structural and cultural influences on emotion. Because the influence of structural variation on emotion management has remained in the background, investigators tend to operate as if a broad normative consensus exists regarding the content of emotion norms and the extent to which they apply to the population being studied (Gordon 1990; Thoits 1989). Hochschild’s (1979, 1983b) early examples probably reinforced this tendency by assuming that all social groups hold similar emotion norms for funerals and weddings and that gendered occupational norms are similarly recognized and internalized by all women performing those occupations.2 As a result, emotion management researchers have been less likely to examine how diffuse status characteristics (other than gender) may, along with proximate structural positions, alter the extent to which actors are held (and hold themselves) accountable for adhering to particular feeling rules through surface and/or deep acting processes.3 So while there have been a number of studies targeting the emotional labor performed within specific occupations (Erickson and Wharton 1997; Leidner 1993; Lively 2000; Pierce 1995; Pugliesi and Shook 1997; Smith and Kleinman 1989; Sutton 1991; Sutton and Rafaeli 1988; see also Chapter 16 by Lively), researchers have not critically examined the underlying assumption that cultural, organizational, or professional norms, values, and beliefs influence emotion management processes more than structure and that, in studying emotion management, culture rather than structure should be of primary concern. 2 The problem may also have been exacerbated by the tendency for authors (including Hochschild 1979, 1983) to frame Goffman as an interactionist rather than emphasizing the more structural implications of Goffman’s work (see Collins 2004; Giddens 1988; Gonos 1977). 3 Given her more recent work (Ehrenreich and Hochschild 2002; Hochschild 2000, 2003), I doubt that Hochschild would universalize emotion norms in the same way today. However, at the time of her initial writing, the third wave feminist movement – that which raised awareness of the profound differences among women and the implications of intersectionality (Hill Collins 1991; Glenn 1992; hooks 1989) had not yet taken hold.

CH15-P374095.indd 264

5/26/2008 4:08:59 PM

Reconsidering culture, structure, and emotional labor

265

One unanticipated result of this tendency is that studies of emotion management are not generally viewed as contributing to structural knowledge of emotion and, simultaneously, “structural” examinations of emotion tend to focus on emotional experience rather than emotion management. The tendency to separate structural and cultural approaches to the study of emotion and emotion management leads to unintended restrictions on the questions being asked and the methods being used to examine how social contexts might relate to emotional experiences along with their management. As social structure and personality researchers are well aware, there are also serious political implications of adopting a strictly cultural approach to social phenomena. When social problems (e.g., the negative outcomes associated with emotional labor performance) are identified as having a cultural explanation, they are linked to behavior patterns emerging from beliefs and values that are transmitted through socialization (House 1981; MacLeod and Lively 2003). As a result, reducing the incidence of negative outcomes requires “fixing” the individual not changing the system (Lewis 1968). Despite Thoits’ (1989) early recognition that a social structure and personality approach to socialization would tend to highlight the influence of exogenous macrostructures, the scholarship on the negative outcomes of emotional labor, though quite diverse, does not usually demonstrate how the (capitalist) logic underlying organizational structures are primarily to blame for these effects. The problematic implications of sociologists’ failure to emphasize the structural factors influencing the performance of emotional labor are particularly evident when the perspective has been employed outside of sociology. Work within the organizational psychology and medical literatures are filled with descriptions of the need to recruit particular personality types for emotionally labor-intensive jobs or to enhance the individual “coping” strategies of emotional laborers who are at greatest risk for negative outcomes (Ducey, Gautney, and Wetzel 2003; Korczynski 2003; Lewis 2005; Sanders 2005; Schweingruber and Berns 2005; cf. Bullock and Waugh 2004). Is this what Hochschild (1983b) envisioned would result from drawing our attention to how feelings have been commodified? In sum, although the performance of emotion management has been consistently linked with holding relatively lower interactional status (Goodrum and Stafford 2003; Hochschild 1983b; Leidner 1993; Lively 2002; Pierce 1995; see also chapter 16 by Lively) and theorists have long argued that emotion management is an integral part of the process through which systems of stratification are reproduced (Clark 1990; Collins 1990), researchers have yet to systematically examine how aspects of both structure and culture may influence this process. The tendency for emotional labor researchers to focus on the types, distribution, and effects of emotional labor performance (Lively 2000) does not have to restrict their ability to move beyond linking these to the feeling rules governing the situation or the emotion norms to which they are held accountable. As sociologists of emotion continue to explore the individual emotion management strategies that emerge from cultural norms, they must also seek to understand

CH15-P374095.indd 265

5/26/2008 4:08:59 PM

266

Social Structure and Emotion

how these strategies are linked to systems of stratification and how they are being used in the reproduction and change of these same systems (Hochschild 1979, 1983b, 1990; see MacLeod and Lively 2003 for a similar argument).

THE CASE FOR CONTEXT Many scholars have argued that everyday interaction helps reproduce social inequality; they identify this interactional link as culture. . . .“Emotion work” links everyday interaction and social structure. —Eliasoph and Lichterman (2003:773)

As suggested by Eliasoph and Lichterman (2003), studies of emotional labor have the potential to contribute to sociological understanding of social reproduction and change. To do so, however, requires a return to the original point of inquiry specified by Hochschild (1979:551) when she stated that an emotion-management perspective serves as “a lens to inspect self, interaction, and structure” (emphasis added). This process would be aided by shifting away from the idea that studies of emotional labor reflect a cultural approach and toward a conceptualization that emphasizes the contextual bases of emotion management processes and outcomes. This may seem to be a mere language game but, as suggested above, the implications of such linguistic choices should not be overlooked. A contextual approach to emotion management would include theoretical and methodological references to systems of social relations (i.e., structure) and systems of meaning (i.e., culture).4 What I am suggesting is similar to the approach advocated by Lawler and Thye’s (1999) use of the term “exchange context” as a linguistic tool that recognizes the simultaneous influence of emotion norms and power/status conditions.5 A contextual approach is consistent with House’s (1995) recommendations that narrower and more precise conceptions of structure and culture be used. As described above, a contextual approach would therefore be consistent with the work of social theorists who have sought to conceptualize culture as structured. This would include Giddens’ (1984) conceptualization of “rules and resources” and Sewell’s (1992)“schemas and resources.” 4

This definition of “social context” is taken from Hays’ (1994:65) reconceptualization of “social structure.” I use the more inclusive “context” to avoid conceptual confusion between the use of the terms “structure” and “social structure.” 5 Lawler and Thye (1999) initially state that the emotional culture surrounding the experience and management of emotion takes place within a context that also includes structural positions. However, they later revert to the common tendency to grant more scientific authority to structural-relational approaches by noting that, as opposed to the cultural-normative approaches that are “more conceptual and interpretive” (225), these approaches make “causal predictions for which emotions emerge given a set of structural conditions.” While I do not deny that this has been the case, what I am suggesting is that adopting a cultural approach should not be characterized as precluding a consideration of structural conditions or of making causal predictions.

CH15-P374095.indd 266

5/26/2008 4:08:59 PM

Reconsidering culture, structure, and emotional labor

267

House (1995:390) notes that these are consistent with his narrower definitions of culture and structure and that “these are two separable components of any social system, that each may affect the others, and that one or the other may be more important as a cause or explanation of any observed individual behaviors or patterns of behavior among members of a social system.” Extending the argument proposed by Lawler and Thye (1999), a contextual approach to studying emotion management would examine the effect of normative expectations on the experience and management of emotion as well as the extent to which social positions impact the emotions people are likely to experience and manage. As these authors state, Collins’ (1975, 2004) order giver–order taker theory of stratification has stimulated exchange theorists to explore a number of different questions about the relationships among power, status, and emotional experience. I would suggest that this is no less possible for studies of emotion management. In this case, researchers would need to explore not only the relative effect of normative expectations and social positions on felt emotion but also on the extent to which these facets of social context may combine to influence the emotion management processes that result. This would suggest that Lawler and Thye’s (1999:223) statement that “structural-relational conditions are fundamental causes of emotions actually felt and emotion norms shape their expression or display” (see Collins 1975, 1990, 2004; Kemper 1978, 1991; Smith-Lovin 1995; Stets 2003; Thoits 1989; Turner and Stets 2005 for similar arguments) should not be viewed as a foundational assumption within the sociology of emotions but should be reframed as a question in need of empirical verification.

THE CURRENT STUDY The current study provides an initial examination of the issues discussed above. In contrast to most quantitative studies of emotional labor that focus on the mental health outcomes of emotional experience and management (Adelmann 1989; Bulan, Erickson, and Wharton 1997; Erickson and Ritter 2001; Pugliesi and Shook 1997; cf. Gibson 1997), I limit the analysis to examining the extent to which aspects of social context are associated with the experience of emotion and the performance of surface and deep acting. In exploring these relationships, I have organized the analysis in a way that is generally consistent with a social structural and personality framework (House 1981; MacLeod and Lively 2003). Given that the sample is limited to registered nurses working within one health care system, I have included social context variables that are most consistent with other empirical research on this population (Aiken et al. 2001; Atencio, Cohen, and Gorenberg 2003; Institute of Medicine 2004; Rambur et al. 2005; Smith 1992). In addition to operationalizing the cultural context in terms of feeling rules and values, I operationalize the structural context in terms of both individual positions in social hierarchies and dimensions of the more proximate occupational context. The latter includes a measure

CH15-P374095.indd 267

5/26/2008 4:09:00 PM

268

Social Structure and Emotion

of work-identity discrepancy that, while not completely consistent with measures used by identity theorists (Burke 1991; Marcussen and Large 2003), supports the underlying logic that emotions tend to emerge from discrepancies between how people perceive themselves in a situation and their identity standards. The longstanding recognition by interactionist scholars that the production and reproduction of status systems involve identity processes (Burke and Hoelter 1988; White and Burke 1987) makes the inclusion of such variables essential for studies of emotion management processes as well (see Ashforth and Humphrey 1993 for a similar argument). A number of specific hypotheses could be listed that reflect previous research and theory on emotional experience and management. These might include: – Those holding higher status positions (men, whites, managers, those with more human capital) are likely to experience more positive emotions (Hochschild 1979; Thoits 1989). – Those holding higher status positions are less likely to manage their emotions (Erickson and Ritter 2001). – The more feeling rules are perceived to apply to situated interaction, the more emotion management will be performed (Grandey 2000; Morris and Feldman 1996). – The more job autonomy one has, the less emotion management will be performed (Morris and Feldman 1996). – The more intense the emotions experienced, the more likely emotion management will be performed (Erickson and Ritter 2001; Grandey 2000). Although each of these could be tested with the current data, for this chapter I am not as interested in these more detailed outcomes as I am with examining how the variables, as indicators of structure and culture, are related to felt emotion and emotion management processes. – To what extent are structure and culture related to emotional experiences? To surface and deep acting? – Is structure more strongly related to emotional experience than culture? Is culture more strongly related to emotion management than structure?

METHODS DATA AND SAMPLE

The data presented here are part of a larger study examining the effect of the emotional context of care on the retention of registered nurses working within acute care hospitals and were collected from registered nurses employed at two urban hospitals located in a midwestern city. The hospitals were owned and operated by a single health care system. A complete listing of nursing personnel was obtained from the health system’s human resources department, and a written

CH15-P374095.indd 268

5/26/2008 4:09:00 PM

Reconsidering culture, structure, and emotional labor

269

questionnaire was distributed to all registered nurse employed within each hospital (n  1461). Eighty percent of nurses returned a completed questionnaire (n  1169). For this study, the experiences of both managers and direct care nurses were examined. Nine individuals were dropped from the analyses as a result of large amounts of missing data; for those missing one or two scale items, mean substitution was used. Of the 1158 registered nurses, 96% were female and 95% were white. In regard to education, 32.6% earned a diploma of nursing, 11.4% had an Associate’s degree, 49.1% had a Bachelor’s degree, and 6.9% had earned a graduate degree in nursing. Seventy-five percent of the respondents were married and sixty-five percent had children living at home. The mean age of respondents was 43 years old with an average of 17.4 years as a registered nurse. These demographics are consistent with national averages for gender and age. However, the current sample under-represents minority nurses, and over-represents those who are married with children. Although these statistics are representative of nurses employed by the health system studied, nurses holding bachelor’s degrees are overrepresented compared with hospitals nationally (Spratley 2000). The average nurse in this sample worked 34 hours per week with 65% working a day shift. Consistent with the hospitals’ employment records, 21% of the nurses in this sample were employed in medical-surgical units, 30% in critical care, 15.7% in operating or recovery units, 16.8% in maternity, and the remaining 17% in psychiatric, outpatient, or other units. Emotional Experiences and Emotional Labor The analyses focus on three types of emotional experiences (i.e., agitated, negative, and positive) and two general forms of emotional labor (i.e., surface and deep acting). Because the performance of emotional labor rests, in part, on the extent to which emotions are felt (Erickson and Ritter 2001), emotional experiences were included as independent variables in the models estimating the performance of emotional labor. The operationalization of emotional experiences builds on prior measurement strategies (Erickson and Ritter 2001) as well as information obtained in preliminary focus groups with registered nurses from both hospitals (n  96). Because the sample was limited to registered nurses, I could not assume that Russell’s (1989) or Plutchik’s (1962) models would necessarily capture the most frequent emotions experienced by nurses during work hours. As a result, these lists of emotion were modified based on information obtained from nurses about their own emotional experiences. The resulting list included the following 15 emotions: afraid, angry, anxious, ashamed, calm, excited, frustrated, guilty, happy, helpless, irritated, proud, relaxed, sad, and surprised. In the written questionnaire, respondents were asked to think about the last week at the hospital and to indicate how strongly or intensely they felt each of the identified emotions on an eight-point scale (0  Not at all; 1  Very Slightly; 4  Moderately; 7  Very Intensely). Factor analyses were then used to combine sets of emotions into

CH15-P374095.indd 269

5/26/2008 4:09:00 PM

270

Social Structure and Emotion

similarly valenced groups (Erickson and Ritter 2001; Russell 1989). Three significant factors emerged from these analyses. The first factor, agitated emotions, included feelings of anger, frustration, and irritation (Cronbach’s alpha  0.89; eigenvalue  5.17). The second factor contained the positive emotions of calm, happy, excited, proud, and relaxed (Cronbach’s alpha  0.78; eigenvalue  2.37). The final factor included the negative emotions of afraid, ashamed, and guilty (Cronbach’s alpha  0.60; eigenvalue  1.13). The factors explained 34%, 16%, and 8% of the variance, respectively. Factor loadings ranged from 0.62 to 0.88 and no items loaded highly (0.40) on more than one factor. One of the contributions of this study is the examination of how social context may influence surface and deep acting differently. Surface acting, or “expressive emotion work” (Hochschild 1979:562) occurs when individuals alter their outward display of emotion in an effort to conform to feeling rules and as a means to potentially modify felt emotion (Hochschild 1983b). Deep acting takes place when a person directs her management efforts toward exhorting authentically felt emotion so that it is in line with applicable feeling rules. Each form of emotional labor requires a certain amount of effort and success is not guaranteed. In addition, while both forms occur as a result of emotional dissonance (i.e., when felt emotions are inconsistent with feeling rules), the amount and type of effort or “intervention” varies according to the specific form of emotional labor performed (Ashforth and Humphrey 1993; Grandey 2003; Hochschild 1983a,b). Hochschild (1983b:33) identified two basic surface acting strategies. The first, disguising what we feel or “suppression,” has been identified empirically as masking or covering up felt emotions that are at odds with the feeling rules governing the interaction (Erickson and Ritter 2001; Erickson and Wharton 1997; Hochschild 1979, 1983b; Lively 2000). The second, pretending to feel what we do not, is a more active process than masking in that the individual attempts to generate visible signs of expected feeling; a process that requires the creation or modification of emotional displays in addition to the masking of inappropriate feelings. Based on these conceptual differences, the first indicator of surface acting asked respondents to indicate on a five-point scale the frequency with which they covered up their true feelings at work with each set of role-related interactional partners: patients, patients’ families, physicians/residents, unit managers/directors, nursing co-workers, and non-RN staff (1  Never, 5  Everyday; adapted from Brotheridge and Lee 2003). These items were then summed to form an index with a range from 6 to 30 where higher scores indicate the extent to which respondents attempted to cover up their true feelings at work. For the second indicator of surface acting, respondents were asked to report the frequency with which they pretended to have feelings that were expected but that they did not really feel when interacting with members of each of the previously identified groups (1  Never; 5  Everyday; adapted from Brotheridge and Lee 2003). These items were also summed to form an index ranging from 6 to 30 where higher scored indicate more attempts to pretend to feel the expected emotions.

CH15-P374095.indd 270

5/26/2008 4:09:00 PM

Reconsidering culture, structure, and emotional labor

271

Where surface acting identifies the effort of trying to appear to be experiencing the required emotion, deep acting requires that individuals work on their emotions in an attempt to authentically feel that which is expected (Hochschild 1983b). To measure deep acting, respondents were asked to indicate the frequency with which they make an effort to actually feel the emotions they are expected to display with each group (1  Never, 5  Everyday; adapted from Brotheridge and Lee 2003). As an index, responses ranged from 6 to 30 with higher scores indicating more frequent deep acting efforts. Social Context Facets of the work environment’s cultural context were measured in terms of the hospitals’ perceived emotion norms and the congruence or “fit” between what nurses perceive the organization to value and their personal values as health care professionals. Positive feeling rules do not measure what nurses actually do but rather the specific feeling rules governing their interactions with others on the job. To provide quality care, nurses (along with other health care professionals) are required to actively work toward the creation of positive interactional contexts and to neutralize negative ones (Boykin and Schoenhofer 2001; Mitchell and Grippando 1993). These emotion norms also reflect the need for nurses to engage in interpersonal emotion management (Thoits 1996) as well as self-management. For this study, positive feeling rules were measured by summing the responses to a question that asked: “To be effective in your job, to what extent are you required to:” (a) reassure patients who are distressed or upset; (b) express feelings of sympathy; and (c) express friendly emotions (Cronbach’s alpha  0.73; adapted from Best, Downey, and Jones 1997). Negative feeling rules included requirements that nurses (a) hide anger or disapproval about something someone has done; (b) remain calm even when they were astonished; (c) hide disgust over something someone has done; and (d) hide fear of someone who is threatening (alpha  0.85; adapted from Best, Downey, and Jones 1997). For both of these scales, higher scores indicate that nurses perceived that more of these requirements must be followed to do their job effectively. Maslach and Leiter (1997) identified value congruence as being one of the key domains of work life associated with the relationship between job stress and burnout. When organizational and personal values are congruent, successes are shared and positive outcomes are more likely. When there are differences between an organization’s values and the values of its staff, or if the organization does not practice its stated values, congruence is low and more negative outcomes tend to occur. In this study, value congruence was measured by summing responses to five items assessing the extent to which respondents agreed (1  Strongly Disagree; 4  Strongly Agree) with each of the following statements: (a) My values and the organization’s are alike; (b) Working here forces me to compromise my values; (c) The organization is committed to quality; (d) The organization’s goals influence my day-to-day work activities; and (e) my personal career goals are consistent with the organization’s stated values (alpha  0.75; Leiter and Maslach 2004).

CH15-P374095.indd 271

5/26/2008 4:09:00 PM

272

Social Structure and Emotion

Consistent with a social structure and personality framework (MacLeod and Lively 2003), structural context was operationalized using “component” indicators of human capital (i.e., education, years of nursing experience), each respondent’s position within prominent social hierarchies (i.e., race, income, sex), and more “proximate” measures of job structure (e.g., managerial position, workload, autonomy). Education was measured by asking respondents to indicate the highest level of education they had achieved (1  School of Nursing Diploma, 2  Associate’s Degree, 3  Baccalaureate Degree, 4  Graduate Degree) and years as an RN reflected the length of time that the respondent had been a registered nurse. Sex was measured with female coded as 1. Race was dichotomized with European American/white coded as 1. Income was coded into 11 categories (1  $1 – $9999; 11  $100,000 or more). Managers were identified by asking respondents to indicate whether or not they held a management position (1  yes). The measures of workload and job autonomy were derived from Karasek’s (1979) job demands-control model using a four-point scale (1  Strongly Disagree, 4  Strongly Agree). The workload scale summed responses to five items asking about the extent to which nurses agreed that they: (a) had too much work to do everything well; (b) never seemed to have enough time to get everything done; (c) were not asked to do excessive amounts of work (reverse coded); (d) had enough time to get the job done (reverse coded); and (e) were free form conflicting demands in their job (reverse coded). The Cronbach’s alpha reliability for the workload scale was 0.82. Job autonomy was measured similarly for the following items: (a) I have the freedom to decide what I do on my job; (b) It is basically my own responsibility to decide how my job gets done; (c) My job requires that I be creative; (d) I have a lot of say about what happens on my job; and (e) I am given a lot of freedom to decide how I do my own work (alpha  0.75). The measure of work-identity discrepancy served as an indicator of person-environment fit (French, Caplan, and Harrison 1982) as well as evaluating the level of discrepancy between what nurses thought others expected them to be like on-the-job and how they actually perceived themselves to be.6 To assess the extent to which these identity-related meanings were discrepant, respondents were given a list of 12 paired adjectives and asked to identify which number along a seven-point continuum most accurately reflected their view of what other people generally expected nurses to be like. Respondents were then asked to indicate how they personally saw themselves as a nurse using the same adjective pairs. The adjective pairs included: Altruistic–Selfish, Calm–Agitated, Caring–Uncaring, Dependent–Autonomous, Empathetic–Unfeeling, Flexible–Inflexible, Immoral– Moral, Inept–Competent, Organized–Unorganized, Responsible–Irresponsible, Unknowledgeable–Knowledgeable, Unsure–Confident. The absolute value of the 6 This concept and its associated measure are similar to constructs presented in identity theory (Burke and Reitzes 1991; Burke and Tully 1977) and discrepancy theories (Higgins, Klein, and Strauman 1985; Marcussen and Large 2003). Despite such similarities, the measures are different enough to preclude a direct test of theoretical propositions drawn from these theories.

CH15-P374095.indd 272

5/26/2008 4:09:00 PM

Reconsidering culture, structure, and emotional labor

273

difference between the expected obligation and the actual self-perception for each item was calculated and the results summed. Finally, two control variables were included that have been shown to influence nurse outcomes, number of hours worked and working in the nightshift (1  yes) (Aiken et al. 2001; Institute of Medicine 2004). Number of hours worked per week was measured by collapsing the hours listed into four groups reflecting the most common nursing shifts (1  1–24 hours; 2  25–36 hours; 3  36–40 hours; and 4  more than 40 hours). Table 15.1 presents the descriptive statistics for all variables as well as their bivariate correlations.

RESULTS The primary goal of these analyses was to examine the extent to which there is empirical justification for a more contextual approach to the experience and management of emotion. To this end, a few of the correlational results are worth noting before moving on to the results of regression analyses. As might be expected, proximate forms of culture and structure were more strongly correlated with the emotion variables than were the diffuse status characteristics and measures of human capital (see Table 15.1). Despite the disproportionate representation of women in nursing, being female is negatively correlated with income and holding a managerial position. Women were also more likely than men to perceive the existence of rules requiring the display of positive emotions and to engage in deep acting. Suggestive of the gendered characteristics of nursing, women reported lower levels of work-identity discrepancy than men. Finally, compared to nurses of color, white nurses were more likely to report experiencing positive emotions on the job and to pretend to have feelings that were expected but that they were not actually experiencing. Table 15.2 presents the results of regression analyses examining the relationships between social context and emotional experiences. These results show that structural positions along with proximate occupational characteristics influence the experience of emotion but do so differently depending on the experience under consideration. As indicated in the first column of Table 15.2, hours worked, having a higher workload, work-identity discrepancy, and negative feeling rules are associated with higher levels of agitation (i.e., anger, irritation, frustration) among nurses. In contrast, lower levels of agitation are related to more years of occupational experience, job autonomy, and value congruence. In the second column of Table 15.2, similar relationships are found between negative emotions (e.g., sadness, fear, guilt, shame) and years of experience, workload, discrepancy, and negative feeling rules. In addition, negative emotions are less likely to be reported among managers and those with more education. Finally, the results presented in the last column of Table 15.2 indicate that white nurses report experiencing more positive emotional experiences (e.g., happiness, pride, excitement) than do nurses of color. Higher workload levels, more work-identity

CH15-P374095.indd 273

5/26/2008 4:09:00 PM

274

Social Structure and Emotion Means, Standard Deviations, and Correlation Coefficients for all

TABLE 15.1 1

2

3

4

5

6

7

8

1 Agitated emotions 2 Negative emotions

0.445**

3 Positive emotions

0.319**

0.217**

4 Cover up feelings

0.427**

0.290**

0.316**

5 Pretend to have feelings

0.401**

0.264**

0.262**

0.707**

6 Make effort to feel

0.095**

0.111**

0.001

0.201**

0.256**

7 Hours worked per week

0.111**

0.037

0.092**

0.035

0.077**

8 Nightshift

0.000

0.037

0.049

0.036

0.025

9 Female

0.008

0.001

0.011

0.002

0.003

0.003

0.026

10 White 11 Income

0.074*

12 Education

0.057

0.063*

0.029

0.065*

0.029 0.011 0.059* 0.037

0.149** 0.097**

0.025

0.020

0.029

0.026

0.034

0.002

0.034

0.007

0.561**

0.027

0.026

0.031

0.032

0.016

0.070*

0.026 0.049

0.103**

0.026

0.019

0.038

0.072*

0.280**

13 Years as an RN

0.130**

0.117**

14 Positive feeling rules

0.170**

0.129**

0.002

0.265**

0.208**

0.164** 0.073*

0.038

15 Negative feeling rules

0.287**

0.235**

0.151**

0.416**

0.385**

0.200**

0.027

16 Value congruence

0.340**

0.187**

0.339**

0.259**

0.282**

17 Manager

0.015

0.041

0.091**

0.024

0.039

0.016

0.015

0.098**

0.132**

0.040

0.390**

0.088**

0.118**

0.112**

0.082**

0.113**

0.085**

0.483**

0.349**

0.320**

0.337**

0.282**

0.225**

0.107**

0.260**

0.221**

0.219**

0.034

0.252**

0.260**

0.238**

0.184**

0.194**

0.037

M

8.91

1.55

18.08

17.44

14.75

17.91

2.20

0.21

SD

5.60

2.64

6.60

4.65

5.28

6.77

1.01

0.41

18 Workload 19 Job Autonomy 20 Work identity discrepancy

0.005

0.060*

p  0.05.

*

p  0.01.

**

discrepancy, and negative feeling rules were all associated with lower levels of positive emotion while job autonomy, positive feeling rules, and value congruence were related to higher levels of these feelings. Although the operationalization of identity meanings is not identical to that used by identity theorists, the results regarding work-identity discrepancy support the underlying logic of the theory. In addition, the results showing strong relationships between identity discrepancy and each set of emotional experiences suggest that identity theorists should continue to think about ways of examining the relationships between identity and emotion. To what extent are structure and culture related to emotional experiences? As the results in Table 15.2 indicate, components of both structure and culture

CH15-P374095.indd 274

5/26/2008 4:09:00 PM

275

Reconsidering culture, structure, and emotional labor Variables (n  1158) 9

10

11

12

13

14

15

16

17

18

19

20

0.015 0.114** 0.055 0.041

0.039

0.081**

0.119**

0.056

0.090** 280**

0.098**

0.030

0.002

0.039

0.022

0.024

0.002 0.005

0.061* 0.051

0.025

0.028

0.021

0.065*

0.017

0.297** 0.118**

0.010

0.016

0.098** 0.012

0.005

0.018

0.106** 0.003

0.070*

0.001 0.033

0.011

0.630**

0.178** 0.038 0.163**

0.194**

0.020

0.034

0.169**

0.290**

0.220**

0.098** 0.093** 0.152**

0.267**

0.002

0.067* 0.161**

0.061*

0.089**

0.183**

0.151**

0.124** 0.229** 0.151** 0.050

0.139**

0.122**

0.96

0.95

5.35

2.30

17.36

16.43

9.52

14.15

.10

13.16

12.87

7.42

0.19

0.22

1.48

1.00

10.46

2.53

2.74

2.29

.30

2.70

2.42

4.61

contribute to explaining the variance in emotional experiences. While the influence of structure might have been expected (Kemper 1991; Lawler and Thye 1999), the findings reported here suggest that feeling rules and values also play a significant role in understanding felt emotion. The differential influence of positive and negative feeling rules is particularly interesting in this regard. These results suggest that rules governing the experience and display of positive emotional states may actually help to increase positive emotional outcomes. While merely suggestive, the results suggest that it is the requirement to suppress negative feelings (not exhort positive ones) that are problematic for well-being. The findings concerning the indicators of structural position and human capital are also interesting given the degree to which their effects vary across the different emotions.

CH15-P374095.indd 275

5/26/2008 4:09:00 PM

276

Social Structure and Emotion

TABLE 15.2 Regression of Emotional Experiences on Structural Context, Cultural Context, Emotional Experience, and Controls Agitated emotions

Negative emotions

b



Controls Hours worked Nightshift

0.636 0.273

0.114*** 0.020

Structural context Structural position/ human capital Female White Income Education Years as an RN

0.887 0.093 0.030 0.126 0.036

0.031 0.004 0.008 0.023 0.067**

0.216 0.379 0.019 0.166 0.021

0.761 0.754 0.195 0.172

0.041 0.364*** 0.084*** 0.142***

Job structure Manager Workload Job autonomy Work-identity discrepancy Cultural context Positive feeling rules Negative feeling rules Value congruence Constant R2 Adjusted R2



b

0.101 0.128



b

0.675 0.021

0.103** 0.001

0.016 0.031 0.011 0.063* 0.083**

1.073 2.080 0.131 0.244 0.026

0.032 0.068** 0.029 0.037 0.041

0.517 0.273 0.011 0.110

0.059* 0.280*** 0.010 0.192***

0.055 0.584 0.371 0.223

0.003 0.239*** 0.136*** 0.156***

0.017

0.017

0.355

0.136***

0.048

0.022

0.212

0.104***

0.134

0.468

0.191***

0.058

2.175 0.35 0.34

Positive emotions

 2.573 0.20 0.19

0.039 0.020

0.139*** 0.050

0.229 0.574

0.095** 0.199***

8.135 0.25 0.24

n  1158. p  0.05. ** p  0.01. *** p  0.001. *

Is structure more strongly related to experience than culture? Yes. Although both structure and culture influence the experience of agitated, negative, and positive emotions, the structural variables were more strongly related to these emotional experiences than were the indicators of culture. For nurses, workload consistently had the strongest relationship to these emotional outcomes. Moreover, when the indicators of job structure were removed from the model (not shown), the variance explained dropped 43% (from 0.35 to 0.20) for agitation, 55% (from 0.20 to 0.09) for negative emotions, and 44% (from 0.25 to 0.14) for positive emotions. In

CH15-P374095.indd 276

5/26/2008 4:09:00 PM

Reconsidering culture, structure, and emotional labor

277

comparison, when the cultural indicators were removed, the variance explained dropped only by 11%, 5%, and 20%, respectively. Finally, work-identity discrepancy was significantly related, in the expected directions, to all emotional experiences. This result is significant in that it suggests the important role that work-related identity processes may play in understanding the well-being of employees. Table 15.3 presents regression results for models predicting the performance of surface and deep acting. The findings suggest the importance of distinguishing between these two forms of emotional labor as well as the relative influence of structure and culture. Looking at the effect of emotional experiences across the forms of emotional labor, the results indicate that the intensity of emotional experience demonstrates a relatively strong influence on the performance of surface acting but not deep acting. The difference in the amount of variance explained across the two forms of emotional labor further suggest that we know relatively little about what influences the tendency to engage in deep acting. Although most of the results for deep acting barely reach the p  0.10 threshold, they raise a number of interesting possibilities. First, note that women are more likely to engage in deep acting than men. This finding is consistent with the fact that nursing is a highly gendered occupation and Hochschild’s (1983b:167) suggestion that deep acting processes may be more closely associated with women’s ways of performing emotion work than men’s. Also noted in column 3 of Table 15.3, years of experience is associated with less deep acting while workload is associated with more. It may be that more experienced nurses find little need to engage in emotional labor of any kind and that a higher workload requires a range of emotion management techniques. (If this latter is truly the case, however, it is somewhat surprising that no such effect is found between workload and pretending to have feelings.) Finally, as with the other forms of emotional labor, the perception that negative feeling rules are expected to govern workplace interactions is positively associated with the performance of deep acting. In the first column of Table 15.3, the results indicate that workload, negative feeling rules, and the experience of agitated and negative emotions are positively related to covering up or masking one’s feelings on the job. As might be expected, job autonomy, value congruence, and the experience of positive emotion tend to be associated with fewer attempts to cover up one’s feelings. The results in column two indicate that the two forms of surface acting reflect somewhat different management processes. First, white nurses are more likely than nurses of color to pretend to feel emotions that they are not actually experiencing. Experiencing greater work-identity discrepancy is also significantly associated with pretending to feel. Other results are somewhat similar to those for covering up. Job autonomy is associated with less pretending as is value congruence and positive emotional experiences. In contrast, negative feeling rules and experiencing agitated and, marginally, negative emotions are related to nurses being more likely to pretend to having feelings they do not actually experience. To what extent are structure and culture related to surface and deep acting? As the results in Table 15.3 suggest, components of both structure and culture

CH15-P374095.indd 277

5/26/2008 4:09:00 PM

278

Social Structure and Emotion

TABLE 15.3 Regression of Surface and Deep Acting on Structural Context, Cultural Context, Emotional Experience, and Controls Surface acting Cover up feelings

Deep acting

Pretend to have feelings

Make an effort to feel

b



b



b



Controls Hours worked Nightshift

0.223 0.560

0.048 0.049

0.463 0.485

0.088** 0.037

0.183 0.298

0.027 0.018

Structural context Structural position/ human capital Female White Income Education Years as an RN

0.475 0.540 0.136 0.144 0.014

0.020 0.025 0.043 0.031 0.031

0.287 1.458 0.047 0.154 0.021

0.011 0.060* 0.013 0.029 0.041

1.941 1.383 0.154 0.267 0.038

0.056† 0.044 0.034 0.039 0.059†

0.112 0.100 0.127 0.040

0.007 0.058* 0.066** 0.040

0.593 0.028 0.170 0.079

0.034 0.014 0.078** 0.069**

0.867 0.151 0.041 0.017

0.038 0.060† 0.014 0.012

0.055

0.012

0.006

Job structure Manager Workload Job autonomy Work-identity discrepancy

Cultural context Positive feeling 0.102 rules Negative feeling 0.421 rules Value congruence 0.114 Emotional experience Agitated emotions 0.171 Negative emotions 0.102 Positive emotions 0.103 Constant R2 Adjusted R2

13.968 0.33 0.32

0.248***

0.136

0.051

0.514

0.266***

0.356

0.144***

0.056*

0.240

0.104***

0.066

0.022

0.206*** 0.058* 0.146***

0.186 0.096  0.069

0.197*** 0.048†  0.086**

0.009 0.132 0.061

0.008 0.051 0.059†

12.417 0.29 0.28

9.504 0.06 0.05

n  1158. †  0.10. * p  0.05. ** p  0.01. *** p  0.001.

contribute to explaining the variance in emotional labor. Here, the influence of culture would have been anticipated (Smith-Lovin 1995; Thoits 1989). However, the findings indicate that aspects of structure also influence the performance of each type of emotional labor, though less consistently than in the models predicting

CH15-P374095.indd 278

5/26/2008 4:09:00 PM

Reconsidering culture, structure, and emotional labor

279

emotional experience. Clearly, feeling rules are not the only factor influencing the performance of emotional labor but proximate structures appear to have more influence on the performance of surface acting than deep acting. Given the small number of minority and male nurses, the results showing that nurses of color tend to “pretend” less frequently than white nurses and female nurses perform more deep acting than male nurses are particularly interesting. Though merely suggestive, these results raise new questions about the extent to which the different forms of emotional labor operate differently across race and gender. These results indicate the importance of examining the structural conditions that may influence the tendency to rely on specific forms of management strategies, the effectiveness of such strategies, and the extent to which these differences may influence well-being. Is culture more strongly related to emotion management than structure? Yes. Although both structure and culture influence the performance of emotional labor, the cultural variables were more strongly related to these performances than were the indicators of structure. For nurses, negative feeling rules were most strongly associated with the performance of each form of emotional labor. When the indicators of the cultural context were removed from the models (not shown), the variance explained dropped 24% (from 0.33 to 0.25) for covering up feelings, 24% (from 0.29 to 0.22) for pretending, and 50% (from 0.06 to 0.03) for deep acting. In comparison, when the job structure variables were removed, the variance explained dropped only by 0.01 for each of the models tested (i.e., 3%, 3%, and 17%, respectively). It should be noted that the intensity of emotional experiences were also significantly related to the performance of emotional labor, particularly in regard to surface acting. Thus, it is likely that the proximate indicators of job structure are influencing emotional labor performance indirectly through their effect on emotional experience.

DISCUSSION AND CONCLUSION The findings reported for the performance of emotional labor indicate that Hochschild (1979) was right to conceptualize the emotion work process as related to feeling rules and structure. Unfortunately, the tendency for many writers who followed to apply a cultural label to Hochschild’s approach has contributed to the relative neglect of structural influences and an overemphasis on the role of feeling rules. Explicitly recognizing the role of structure and culture in further theoretical and empirical work on emotion management process would not only represent a more accurate legacy for Hochschild’s work but would help to clarify the importance of emotional processes for understanding social reproduction and change. The results for emotional labor also lend support to the utility of conceptualizing the causal significance of culture not in terms of ultimate values but more in keeping with Swidler’s (1986) image of culture as a “tool kit” which people may use in varying configurations to solve different types of problems. These tools are used to construct patterned “strategies of action” that differ across both time

CH15-P374095.indd 279

5/26/2008 4:09:01 PM

280

Social Structure and Emotion

and structural position. Taking this approach, culture is not viewed as a unified system that pushes action in one consistent direction. Instead, actors are seen as using features of culture in an active, skilled manner. Based on the findings presented here, the extent to which surface and deep acting are perceived to be tools that are useful are, in turn, structured by race and gender positions as well as by the amount of work-identity discrepancy and the autonomy one has in choosing a course of action. Such an approach would also be consistent with the more traditional use of culture to explain why different groups may behave differently in the same structural situation. Despite frequent mention of the potential for variation in the emotional experiences and management across race, class, and gender (Hochschild 1990), few researchers have moved beyond examinations of gender variation (Mirchandani 2003; cf. Simon and Nath 2004). As Hochschild notes, based on variations in one’s structural location, particular ideologies and feeling rules emerge influencing the development of preferred strategies of action. Here, the question becomes: To what extent do individuals alter their emotional strategies (including strategies related to emotion management) to fit potentially different emotional subcultures related to their positions within the stratification system? Assuming that we view individuals as self-motivated (or agentive), we can conceive these strategies as “active strategic stances” toward stratification systems in which they may be disadvantaged and empirically investigate the emotion management processes that may emerge from each stance. This would be consistent with Cahill’s (1998) view that diffuse status characteristics such as gender or race are likely to influence behaviors related to the more circumscribed, or proximate, statuses found within organizational arenas. In the future, emotion management researchers need to more fully explore the extent to which emotional cultures, experiences, and management processes that may be linked to organizational or professional settings generalize across the diffuse characteristics that sociologists commonly associate with more macro-level inequalities. For example, Thoits (1985, 1990) recognized that status inconsistencies such as being a “male nurse” might lead to discrepant emotional experiences. Given the history of racialization within the nursing profession (Glenn 1992), being a nurse of color might also represent a form of multiple role occupancy and/or subcultural marginality that may be associated with reports of emotional deviance (Thoits 1990). Although the results reported here do not provide direct support for the idea that those with these multiple statuses are more likely to experience emotional deviance, they do suggest that the feeling rules underlying the profession of nursing may not apply consistently across social structural categories. The findings that male nurses were less likely to perform deep acting and that nurses of color were less likely to perform surface acting that required them to pretend to have emotions they did not feel provide indications that diffuse structural characteristics influence the emotional labor process in a way that may (or may not) be unique to the profession of nursing. In addition to raising questions about the theoretical and empirical adequacy of examining the effects

CH15-P374095.indd 280

5/26/2008 4:09:01 PM

Reconsidering culture, structure, and emotional labor

281

of feeling rules as if they operated outside of the influence of structure, these results have implications for how studies of emotional labor may facilitate new insights into agency and social change. Since Hochschild’s initial empirical examination of the effects of emotional labor on employee well-being, researchers have sought to specify the conditions that explain, exacerbate, or buffer such effects (Bulan, Erickson, and Wharton 1997; Erickson and Ritter 2001; Lively 2002; Pugliesi and Shook 1997; Wharton 1993). This research tradition has shown that while emotional labor may not have negative effects on well-being at all times, it tends to be problematic for worker health. At the same time, experiencing emotions that are at odds with specified feeling rules tend to lead to emotion work that will bring feeling back in line with social expectations (Thoits 1985). The results of the current study indicate that those who should be experiencing greater emotional discrepancies (i.e., those with status inconsistencies) are not necessarily engaging in all forms of emotion management available to them. The crucial question thus becomes: How does this lack of emotional labor performance affect the health of these nurses? If males and people of color do not hold themselves accountable to the same occupational emotion norms as white, female nurses, might status inconsistency actually have health benefits? While clearly preliminary, these results do lend support to Mirchandani’s (2003) suggestion that studies of emotional labor have masked important forms of racialized emotion work. Mirchandani’s insights draw attention to the fact that theorists of emotion management have often recognized the gendered expectations in women’s jobs that assume the natural possession of caring skills. Neglected in most of these studies, however, have been the race and class assumptions underlying the privileging of gender. While emotion management scholars may continue to view emotion work as basically an act of “doing gender” in a particular way, the results reported here suggest that we must not lose sight of the fact that these structured expectations are carried through actors whose emotional cultures modify these gendered expectations in subtle and complex ways. By not performing surface and deep acting in the ways that are consistent with how white, female nurses do it, the experiences of minority and male nurses may provide unique insights into how to resist some of the more oppressive and debilitating dimensions of the commodification of care. In adopting a contextual approach that explicitly models how minority positions within institutional systems may provide insight into difference, resistance, and change, emotion management researchers may be able to contribute to more general theoretical attempts to elucidate the enabling or empowering features of social structure (Giddens 1984; Hays 1994). In sum, I suggest that emotions scholars should seriously consider the theoretical and empirical utility of differentiating between “structural” and “cultural” approaches. In particular, I hope that the information provided here has raised questions about the adequacy and implications of applying the cultural label to the approach inspired by Hochschild’s (1979, 1983b) original work. By reconsidering how structure and culture are theorized and studied, scholars of emotional labor are well-positioned to develop a contextual approach that does not

CH15-P374095.indd 281

5/26/2008 4:09:01 PM

282

Social Structure and Emotion

implicitly assume that persons who occupy similar structural positions share common experiences. In so doing, their work will further sociological understanding of the microexperiences of macrostructures and the emotional bases of social reproduction and change.

ACKNOWLEDGMENT This research was supported by a grant from the Health Resources and Services Administration (1 D1D HP 00004-010).

REFERENCES Adelmann, Pamela K. 1989. Emotional Labor and Well-Being. PhD dissertation, Department of Psychology, The University of Michigan. Aiken, Linda H. et al. 2001. “Nurses Report on Hospital Care in Five Countries.” Health Affairs 20:43–53. Alexander, J. C., B. Giesen, R. Munch, and N. Smelser (eds). 1987. The Micro–Macro Link. Berkeley, CA: University of California Press. Ashforth, Blake E. and Ronald H. Humphrey. 1993. “Emotional Labor in Service Roles: The Influence of Identity.” Academy of Management Review 18:88–115. Atencio, Bonnie L., Jayne Cohen, and Bobbye Gorenberg. 2003. “Nurse Retention: Is It Worth It?” Nursing Economics 21(6):262–268. Barbalet, Jack (ed.) 2002. Emotions and Sociology. Oxford: Blackwell Publishing. Berger, Peter L. and Thomas Luckmann. 1966. The Social Construction of Reality. Garden City, NY: Anchor Books. Best, Richard. G., Ronald G. Downey, and R. G. Jones. 1997. “Incumbent Perceptions of Emotional Work Requirements.” Paper presented at the 12th Annual Conference of the Society for Industrial and Organizational Psychology, St Louis, Missouri, April. Blumer, Herbert. 1969. Symbolic Interaction. Englewood Cliffs, NJ: Prentice-Hall. Boykin, Anne and Savina O. Schoenhofer. 2001. Nursing as Caring: A Model for Transforming Practice. Boston: Jones and Bartlett. Brotheridge, Céleste M. and Alicia A. Grandey. 2002. “Emotional Labor and Burnout: Comparing Two Perspectives of ‘People Work’.” Journal of Vocational Behavior 60:17–39. Brotheridge, Céleste M. and Raymond T. Lee. 2003. “Development and Validation of the Emotional Labour Scale.” Journal of Organizational and Occupational Psychology 76:365–379. Bulan, Heather Ferguson, Rebecca J. Erickson, and Amy S. Wharton. 1997. “Doing for Others on the Job: The Affective Requirements of Service Work, Gender, and Emotional Well-Being.” Social Problems 44(2):235–256. Bullock, Heather E. and Irma Morales Waugh. 2004. “Caregiving Around the Clock: How Women in Nursing Manage Career and Family Demands.” Journal of Social Issues 60:767–786. Burke, Peter J. 1991. “Identity Processes and Social Stress.” American Sociological Review 56:836–849. Burke, Peter J. and Jon W. Hoelter. 1988. “Identity and Sex–Race Differences in Educational and Occupational Aspirations Formation.” Social Science Research 17:29–47. Burke, Peter J. and Donald C. Reitzes. 1991. “An Identity Approach to Commitment.” Social Psychology Quarterly 54:239–251. Burke, Peter J. and Judy C. Tully. 1977. “The Measurement of Role Identity.” Social Forces 55:881–897.

CH15-P374095.indd 282

5/26/2008 4:09:01 PM

Reconsidering culture, structure, and emotional labor

283

Cahill, Spencer. 1998. “Toward a Sociology of the Person.” Sociological Theory 16(2):131–148. Calhoun, Craig, Chris Rojek, and Bryan S. Turner. 1995. Handbook of Sociology. London: Sage Publications. Clark, Candace. 1990. “Emotions and Micropolitics in Everyday Life: Some Patterns and Paradoxes of ‘Place’.” In: T. D. Kemper (ed.), Research Agendas in the Sociology of Emotions,, pp. 305–333. Albany, NY: SUNY Press. Coleman, James S. 1986. “Social Theory, Social Research, and a Theory of Action.” American Journal of Sociology 91(6):1309–1335. Collins, Patricia Hill. 1990. Black Feminist Thought: Knowledge, Consciousness, and the Politics of Empowerment. New York: Routledge. Collins, Randall. 1975. Conflict Sociology: Toward an Explanatory Science. New York: Academic Press. Collins, Randall. 1990. “Stratification, Emotional Energy, and the Transient Emotions.” In: T. D. Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 27–57. Albany, NY: SUNY Press. Collins, Randall. 2004. Interaction Ritual Chains. Princeton. NJ: Princeton University Press. Comte, Auguste. 1896. The Positive Philosophy of Auguste Comte (Trans.) Harriet Martineau. London: George Bell & Sons. Copp, Martha. 1998. “When Emotion Work is Doomed to Fail: Ideological and Structural Constraints on Emotion Management.” Symbolic Interaction 21(3):299–328. Diefendorff, James M., Meredith H. Croyle, and Robin H. Gosserand. 2005. “The Dimensionality and Antecedents of Emotional Labor Strategies.” Journal of Vocational Behavior 66:339–357. Ducey, Ariel M., Heather Gautney, and Dominic Wetzel. 2003. “Regulating Affective Labor: Communication Skills Training in the Healthcare Industry.” Research in the Sociology of Work 12:49–72. Ehrenreich, Barbara and Arlie Russell Hochschild. 2002. Global Woman: Nannies, Maids, and Sex Workers in the New Economy. New York: Metropolitan Books. Eliasoph, Nina and Paul Lichterman. 2003. “Culture in Interaction.” American Journal of Sociology 108:735–794. Erickson, Rebecca J. and Christian Ritter. 2001. “Emotional Labor, Burnout, and Inauthenticity: Does Gender Matter?” Social Psychology Quarterly 64:146–163. Erickson, Rebecca J. and Amy S. Wharton. 1997. “Inauthenticity and Depression: Assessing the Consequences of Interactive Service Work.” Work and Occupations 24(2):188–213. Fligstein, Neil. 2001. “Social Skill and the Theory of Fields.” Sociological Theory 19:105–125. French, John R. P. Jr., Robert D. Caplan, and R. Van Harrison. 1982. The Mechanisms of Job Stress and Strain. New York: Wiley. Gibson, Donald E. 1997. “The Struggle for Reason: The Sociology of Emotions in Organizations.” In: R. J. Erickson and B. Cuthbertson-Johnson (eds), Social Perspectives on Emotion, Vol. 4, pp. 211–256. Greenwich, CT: JAI Press. Giddens, Anthony. 1984. The Constitution of Society. Berkeley, CA: University of California Press. Giddens, Anthony. 1988. “Goffman as a Systematic Social Theorist.” In: P. Drew and A. Wootten (eds), Erving Goffman: Exploring the Interaction Order, pp. 250–279. Cambridge: Polity. Giddens, Anthony. 1991. Modernity and Self-Identity. Stanford, CA: Stanford University Press. Glenn, Evelyn Nakano. 1992. “From Servitude to Service Work: Historical Continuities in the Racial Division of Paid Reproductive Labor.” Signs 18(1):1–43. Goodrum, Sarah and Mark C. Stafford. 2003. “The Management of Emotions in the Criminal Justice System.” Sociological Focus 36:179–196. Gordon, Steven L. 1990. “Social Structural Effects on Emotion.” In: T. D. Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 145–179. Albany, NY: SUNY Press. Gonos, George. 1977. “‘Situation’ Versus ‘Frame’: The ‘Interactionist’ and the ‘Structuralist’ Analyses of Everyday Life.” American Sociological Review 42:854–867. Grandey, Alicia A. 2000. “Emotion Regulation in the Workplace: A New Way to Conceptualize Emotional Labor.” Journal of Occupational Health Psychology 5(1):95–100.

CH15-P374095.indd 283

5/26/2008 4:09:01 PM

284

Social Structure and Emotion

Grandey, Alicia A. 2003. “When ‘The Show Must Go On’: Surface Acting and Deep Acting as Determinants of Emotional Exhaustion and Peer-Rated Service Delivery.” Academy of Management Journal 46:86–96. Hays, Sharon. 1994. “Structure and Agency and the Sticky Problem of Culture.” Sociological Theory 12(1):57–72. Hewitt, John P. 1989. Dilemmas of the American Self. Philadelphia: Temple University Press. Higgins, E. Tory, Ruth Klein, and Timothy Strauman. 1985. “Self-Concept Discrepancy Theory: A Psychological Model for Distinguishing Among Different Aspects of Depression and Anxiety.” Social Cognition 3:51–76. Hochschild, Arlie. 1979. “Emotion Work, Feeling Rules, and Social Structure.” American Journal of Sociology 85(3):551–575. Hochschild, Arlie. 1983a. “Comment on Kemper’s ‘Social Constructionist and Positivist Approaches to the Sociology of Emotions.” American Journal of Sociology 89(2):432–434. Hochschild, Arlie. 1983b. The Managed Heart: Commercialization of Human Feeling. Berkeley, CA: University of California Press. Hochschild, Arlie. 1990. “Ideology and Emotion Management: A Perspective and Path for Future Research.” In: T. D. Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 117–142. Albany, NY: SUNY Press. Hochschild, Arlie. 2000. “The Nanny Chain.” American Prospect 11(4). January 3 Hochschild, Arlie. 2003. The Commercialization of Intimate Life. Berkeley, CA: University of California Press. Hollander, Jocelyn A. and Judith A. Howard. 2000. “Social Psychological Theories on Social Inequalities.” Social Psychology Quarterly 63(4):338–351. Hooks, Bell. 1989. Talking Back: Thinking Feminist, Thinking Black. Boston, MA: South End Press. House, James S. 1981. “Social Structure and Personality.” In: M. Rosenberg and R. H. Turner (eds), Social Psychology: Sociological Perspectives, pp. 525–561. New York: Basic Books. House, James S. 1995. “Social Structure, Relationships, and the Individual.” In: K. Cook, G. A. Fine, and J. S. House (eds), Sociological Perspectives in Social Psychology, pp. 387–395. Boston, MA: Allyn and Bacon. Huber, Joan (ed.). 1991. Macro–Micro Linkages in Sociology. Newbury Park, CA: Sage Publications. Institute of Medicine. 2004. Keeping Patients Safe: Transforming the Work Environment of Nurses. Washington, DC: National Academies Press. Karasek, Robert A. 1979. “Job Demands, Job Decision Latitude, and Mental Strain: Implications for Job Redesign.” Administrative Science Quarterly 24:285–306. Kemper, Theodore D. 1978. A Social Interactional Theory of Emotions. New York: Wiley. Kemper, Theodore S. 1981. “Social Constructionist and Positivist Approaches to the Sociology of Emotions.” American Journal of Sociology 87:336–361. Kemper, Theodore D. 1991. “Predicting Emotions from Social Relations.” Social Psychology Quarterly 54:330–342. Kemper, Theodore D. 1993. “Sociological Models in the Explanation of Emotions.” In: M. Lewis and J. M. Haviland (eds), Handbook of Emotions, pp. 41–51. New York: Guilford Press. Korczynski, Marek. 2003. “Communities of Coping: Collective Emotional Labor in Service Work.” Organization 10:55–79. Lawler, Edward J. and Shane R. Thye. 1999. “Bringing Emotions Into Exchange Theory.” Annual Review of Sociology 25:217–244. Leidner, Robin. 1993. Fast Food, Fast Talk. Berkeley, CA: University of California. Leiter, Michael P. and Christian Maslach. 2004. “Areas of worklife: A Structured Approach to Organizational Predictors of Job Burnout.” In: P. L. Perrewe and D. C. Ganster (eds), Research in occupational stress and well being: Emotional and physiological processes and positive intervention strategies, Vol. 3, pp. 91–134. Oxford, UK: JAI Press/Elsevier. Lewis, Oscar. 1968. La Vida: A Puerto Rican Family in the Culture of Poverty – San Juan and New York. New York: Vintage.

CH15-P374095.indd 284

5/26/2008 4:09:01 PM

Reconsidering culture, structure, and emotional labor

285

Lewis, Patricia. 2005. “Suppression or Expression: An Exploration of Emotion Management in a Special Care Baby Unit.” Work, Employment, and Society 19:565–581. Lively, Kathryn. 2000. “Reciprocal Emotion Management: Working Together to Maintain Stratification in Private Law Firms.” Work and Occupations 21(1):32–63. Lively, Kathryn. 2002. “Client Contact and Emotional Labor: Upsetting the Balance and Evening the Field.” Work and Occupations 29:198–225. MacLeod, Jane D. and Kathryn J. Lively. 2003. “Social Structure and Personality.” In: J. Delamater (ed.), Handbook of Social Psychology, pp. 77–102. New York: Kluwer Academic/Plenum Publishers. Marcussen, Kristen and Michael D. Large. 2003. “Using Identity Discrepancy Theory to Predict Psychological Distress.” In: Peter J. Burke, Timothy J. Owens, Richard T. Serpe, and Peggy A. Thoits (eds), Advances in Identity Theory and Research, pp. 151–164. New York, NY: Kluwer Academic/Plenum Publishers. Maslach, Christina and Michael P. Leiter. 1997. The Truth about Burnout. San Francisco, CA: Jossey-Bass. Mirchandani, Kiran. 2003. “Challenging Racial Silences in Studies of Emotion Work: Contributions from an Anti-Racist Feminist Theory.” Organization Studies 24(5):721–742. Mitchell, Paula R. and Gloria M. Grippando. 1993. Nursing Perspectives and Issues, 5th edition. Albany, NY: Delmar Publishers. Morris, J. Andrew and Morris C. Feldman. 1996. “The Dimensions, Antecedents, and Consequences of Emotional Labor.” Academy of Management Review 21:986–1010. Pierce, Jennifer L. 1995. Gender Trials: Emotional Lives in Contemporary Law Firms. Berkeley: University of California Press. Plutchik, Robert. 1962. The Emotions: Fact, Theories, and a New Model. New York: Random House. Pugliesi, Karen and Scott L. Shook. 1997. “Gender, Jobs, and Emotional Labor in a Complex Organization.” In: R. J. Erickson and B. Cuthbertson-Johnson (eds), Social Perspectives on Emotion, Volume 4, pp. 283–316. Greenwich, CT: JAI Press. Rambur, Betty, Barbara McIntosh, Mary Val Palumbo, and Kyndaron Reinier. 2005. “Education as a Determimant of Career Retention and Job Satisfaction Among Registered Nurses.” Journal of Nursing Scholarship 37(2):185–192. Rubinstein, David. 2001. Culture, Structure, and Agency: Toward a Truly Multidimensional Society. Thousand Oaks, CA: Sage Publications. Russell, James A. 1989. “Measures of Emotion.” In: Robert Plutchik and Henry Kellerman (eds), Emotion: Theory, Research, and Experience, Vol. 4, pp. 83–111. San Diego: Academic Press. Sanders, Teela. 2005. “It’s Just Acting: Sex Workers Strategies for Capitalizing on Sexuality.” Gender, Work, and Organizations 12:319–342. Schweingruber, David and Nancy Berns. 2005. “Shaping the Selves of Young Salespeople Through Emotional Labor.” Journal of Contemporary Ethnography 34:679–706. Sewell, William. 1992. “A Theory of Structure: Duality, Agency, and Transformation.” American Journal of Sociology 99:1–29. Shilling, Chris. 2002. “The Two Traditions in the Sociology of Emotions.” In: J. Barbalet (ed.), Emotions and Sociology, pp. 10–32. Oxford: Blackwell. Simon, Robin W. and Leda E. Nath. 2004. “Gender and Emotion in the United States: Do Men and Women Differ in Self-Reports of Feelings and Expressive Behavior?” American Journal of Sociology 109(5):1137–1176. Smith, Pam. 1992. The Emotional Labour of Nursing. London: Macmillan. Smith, Allen C. III and Sheryl Kleinman. 1989. “Managing Emotions in Medical School: Students’ Contacts with the Living and the Dead.” Social Psychology Quarterly 52:56–69. Smith-Lovin, Lynn. 1995. “The Sociology of Affect and Emotion.” In: K. S. Cook, G. A. Fine, and J. S. House (eds), Sociological Perspectives on Social Psychology, pp. 118–148. Boston: Allyn and Bacon. Smith-Lovin, Lynn. 2003. “They Got the Feeling, But They Missed the Marx: Twenty Years after The Managed Heart.” Paper presented at the Annual Meeting of the American Sociological Association, Atlanta, Georgia.

CH15-P374095.indd 285

5/26/2008 4:09:01 PM

286

Social Structure and Emotion

Smith-Lovin, Lynn. 2005. “Forward.” In: J. H. Turner and J. E. Stets (eds), The Sociology of Emotions, pp. xv–xvxvii. New York: Cambridge University Press. Smith-Lovin, Lynn and Linda Molm. 2000. “Introduction to the Millenium Special Issue on the State of Sociological Social Psychology.” Social Psychology Quarterly 63(4):281–283. Spratley, Ernell. 2000. “The Registered Nurse Population, Findings from the National Sample Survey of Registered Nurses.” U.S. Department of Health & Human Services, Bureau of Health Professions, Division of Nursing, Health Resources and Services Administration. Stets, Jan E. 2003. “Emotions and Sentiments.” In: J. Delamater (ed.), Handbook of Social Psychology, pp. 309–335. New York: Kluwer Academic/Plenum. Sutton, Robert I. 1991. “Maintaining Norms About Expressed Emotions: The Case of Bill Collectors.” Administrative Science Quarterly 36:245–268. Sutton, Robert I. and Anat Rafaeli. 1988. “Untangling the Relationship Between Displayed Emotions and Organizational Sales: The Case of Convenience Stores.” Academy of Management Journal 31:461–487. Swidler, Ann. 1986. “Culture in Action: Symbols and Strategies.” American Sociological Review 51:273–286. Thoits, Peggy A. 1985. “Self-labeling Processes in Mental Illness: The Role of Emotional Deviance.” American Journal of Sociology 91:221–249. Thoits, Peggy A. 1989. “The Sociology of Emotions.” Annual Review of Sociology 15:317–342. Thoits, Peggy A. 1990. “Emotional Deviance: Research Agendas.” In: T. D. Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 180–203. Albany, NY: SUNY Press. Thoits, Peggy A. 1996. “Managing the Emotions of Others.” Symbolic Interaction 19:85–109. Thoits, Peggy A. 2004. “Emotion Norms, Emotion Work, and Social Order.” In: A. S. R. Manstead, N. Frijda, and A. Fischer (eds), Feelings and Emotions, pp. 359–378. Cambridge: Cambridge University Press. Turner, Jonathan H. and Jan E. Stets. 2005. The Sociology of Emotions. New York: Cambridge University Press. Wharton, Amy S. 1993. “The Affective Consequences of Service Work: Managing Emotions on the Job.” Work and Occupations 20:205–232. White, Clovis L. and Peter J. Burke. 1987. “Ethnic Role Identity Among Black and White College Students.” Sociological Perspectives 30:310–331. Wrong, Dennis H. 1961. “The Oversocialized Conception of Man in Modern Sociology.” American Sociological Review 26(2):183–193.

CH15-P374095.indd 286

5/26/2008 4:09:01 PM

16 Status and Emotional Expression: The Influence of “Others” in Hierarchical Work Settings Kathryn J. Lively Dartmouth College, Hanover, NH, USA

ABSTRACT To date, most studies of emotion management in hierarchical work settings have focused on the emotional labor of workers. Recently, however, there has been increasing interest in the dual processes of emotion management and emotional expression. For the most part, these studies have either been qualitative or experimental in nature and, therefore, have relied on ethnography or small convenience samples. Moreover, the majority of studies have focused primarily on individuals’ cognitive attempts to bring their feelings and expressions in line with both culturally and corporately defined feeling rules. Since the release of the General Social Survey’s (GSS) (1996) emotion module, a handful of survey studies of emotion have revealed that others in the workplace may play an important role in these processes. Although studies based on surveys have particular advantages over other methods, extant survey data are necessarily limited by what they ask and do not ask. In this paper, I will address some of the questions raised by recent quantitative work by revisiting a set of in-depth interviews to illustrate the utility of a more multi-method approach to the study of emotion. Social Structure and Emotion

CH16-P374095.indd 287

287

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 4:11:57 PM

288

Social Structure and Emotion

Key Words: Emotion Management, Emotion Expression, Hierarchical Setting, Workplace, Similar/Dissimilar Others

INTRODUCTION In recent years there have been an increasing number of studies detailing the roles that emotion management and emotional expression play in hierarchical work settings. For the most part, these studies have centered on the issue of emotional labor, especially in terms of what it is (Hochschild 1983), how it is distributed (Steinberg and Figart 1999), its role in the reification of existing status structures (Lively 2000; Pierce 1995), and its psychological and physical consequences (Wharton 1996, 1999). However, a number of recent studies have also begun to examine emotional outcomes more generally: in particular, emotion management and emotional expression. While some have examined emotional expression as a form of emotion management (Schieman 2000; Sloan 2004; Stets and Tsushima 2001; and Thoits 1986), others (Lively and Powell 2006) have framed emotional expression as a marker of “social place,” in keeping with research on small groups and status hierarchies (Lovaglia and Houser 1996; Ridgeway and Johnson 1990; also see Clark 1990 and Hochschild 1983). In this chapter, I will discuss the concepts of emotional expression and emotion management (particularly as they have been utilized in survey research) and the relationship between the two. I will also illustrate how a more multi-method approach to the study of emotion management and emotional expression may further our understanding of how emotion operates in hierarchical settings by bringing together findings from ethnographic and survey based research. I will end by suggesting new lines of scholarship that make more central the influence of others on both emotion management and emotional expression in hierarchical settings.

EMOTIONAL EXPRESSION (AS EMOTION MANAGEMENT) Emotion management refers to the cognitive, behavioral, and expressive strategies that individuals use to bring their emotional experiences and expressions in line with culturally mandated feeling rules (Hochschild 1979). If individuals are unable to bring their emotions within appropriate normative bounds, they may be labeled deviant, or may come to label themselves as being mentally ill (Thoits 1986). When emotion management is sold for a wage in order to create a corporately mandated feeling state in oneself in order to elicit a desired response from a customer or client for the benefit of a corporation, it is called emotional labor (Hochschild 1983). Emotional labor is often associated with “service with a smile,” and contains elements of both coercion (corporations co-opt workers’

CH16-P374095.indd 288

5/26/2008 4:11:58 PM

Status and Emotional Expression

289

personal feelings for corporate gain) and manipulation (the emotional laborers attempt to create a change in the customer or client vis-à-vis changes in themselves). Most research on emotional labor and emotion management more generally has focused on self-directed emotion management strategies, with less attention paid to interpersonal emotion management strategies that workers use to influence others’ emotions in more direct ways. In a theoretical paper designed to trace the linkages between the literatures on emotion and stress research, however, Thoits (1986, also see 1995) has argued that emotion management could (and should) be viewed as a form of coping behavior, just as interpersonal emotion management can be viewed of as a form of social support (also see Simon 2006). Building upon this idea, there has been an increasing number of attempts to understand the role that others play in emotion management. The majority of these studies, however, have centered on social interactions among close intimate partners (Staske 1996, 1999) or in therapeutic settings that offer specialized forms of social support (Thoits 1996) and identity transformations (Francis 1997). Ironically, far fewer studies of this type have been conducted in the workplace (but see Gatta 2002 and Lively 2000) or other hierarchical settings. Extending theoretical discussions of the relationship between coping and emotion management and qualitative studies of intrapersonal and interpersonal emotion management (as coping and social support, respectively), some researchers have used survey data to tap into the degree to which individuals use emotional expression as a form of emotion management (or as a form of coping, e.g., to talk to the target of their anger, or support, e.g., to talk to someone else in lieu of the rightful target). In particular, scholars relying primarily on data from the GSS (1996) emotions module have used items tapping into emotional expression (“I spoke to the person with whom I was angry” and “I spoke with someone else”) as forms of intra- and interpersonal emotion management and/or self- and other-directed coping (Schieman 2000; Sloan 2004; Stets and Tsushima 2001).1

EMOTIONAL EXPRESSION (AS A MARKER OF STATUS) Others, however, view emotional expression less as a coping strategy and more as a marker of status or as a symbolic display that reflects social norms that are rooted in both interpersonal and institutional patterns of power and 1

In fact, the wording of these questions leads one to this interpretation. In the series of questions referred to in this chapter, respondents were asked to think of a time in the last month where they had been really irritated, angry, or annoyed. They were then prompted to tell the interviewer the incident and were probed for the appropriate target of their anger, including their social relationship. Finally respondents were shown a list of 19 coping strategies and were asked: “I’m going to show you a list of things that people sometimes do to change their feelings. Did you do any of these things after you got angry or annoyed?”

CH16-P374095.indd 289

5/26/2008 4:11:58 PM

290

Social Structure and Emotion

status (Collins 1990; Kemper 1978; also see Ridgeway and Smith-Lovin 1999). Indeed, most structural accounts of emotion suggest that individuals express their emotions not simply as a means to manage them (and their concomitant stressors; Lazarus and Folkman 1984), but also as a means to mark or enhance their social positions within a social hierarchy (Clark 1990; also see Lovaglia and Houser 1996). For example, Clark argues that individuals holding higher status may purposefully exhibit anger as a means of maintaining or even increasing the social distance between themselves and the targets of their anger (also see Lovaglia and Houser 1996). That emotional expression may be something that is used routinely to mark one’s place in a social hierarchy implies that emotional expression may not always be genuine, but rather a product of emotion management (see Hochschild 1983 and Sutton 1991). Although most scholars acknowledge that emotional expression is subject to acts of emotion management, the relationship between these two concepts remains surprisingly underspecified (a point to which I return below). Using the same data as those studying emotional expression as a form of coping (Schieman 2000; Sloan 2004; Stets and Tsushima 2001), Lively and Powell (2006) recently examined the extent to which emotional expression (in this case, the expression of anger) is influenced by both the domain in which the emotion occurs as well as the relative status of the actors involved, net of the characteristics of the respondent and, in some cases, the gender of the anger’s rightful target. Unlike previous studies that used binary logistic models to document the likelihood of engaging in particular strategies (Did you engage in expressive coping or not? Did you engage in non-expressive coping or not? Sloan 2004), Powell’s and my use of multinomial logistic models allowed us to address not only the likelihood of engaging in a particular strategy but also the likelihood of engaging in a particular strategy instead of (or at the expense of) another. This approach also allowed us to identify how the expression of negative emotion flows both down and across social hierarchies within two of the most important social domains (work and family) by assessing the likelihood that individuals would (1) speak to the target of his or her anger only, (2) speak to someone else, (3) speak to both the target and someone else, or (4) speak to no one at all (see Lively and Powell 2006 for more detail).

THE ROLE OF OTHERS IN HIERARCHICAL WORK SETTINGS Whether scholars using the GSS emotions module have framed their studies in terms of emotion management or emotional expression, their attempts, as a whole, have served to further sociological understanding of how emotion operates generally within the society, as well as within specific social domains. In particular, these studies have challenged our understanding of differences in emotional experience and expression based on diffuse status characteristics (Lively forthcoming;

CH16-P374095.indd 290

5/26/2008 4:11:58 PM

Status and Emotional Expression

291

Lively and Heise 2004; Mabry and Kiecolt 2005; Simon and Nath 2004) yet confirmed our understanding of emotional expression as a function of more institutionalized markers of status (Lively and Powell 2006; Schieman 2000; Sloan 2004; also see Ridgeway and Smith-Lovin 1999). These studies have also furthered the link between emotion and other important social psychological constructs, namely identity, equity, and justice (Hegtvedt 1990; Lawler and Thye 1999; Lively and Heise 2004; Lively, Steelman, and Powell 2004, Stets and Tsushima 2001). For those who have either compared the social domains of work and family (Lively and Powell 2006; Schieman 2000; Stets and Tsushima 2001) or focused specifically on work (Sloan 2004), one of the surprising findings is the seeming importance that others play in both emotion management and emotional expression of workplace anger. This finding is surprising because most studies of emotion in the workplace (e.g. emotional labor) focused strictly on intrapersonal emotion management (Hochschild 1983). However, Lively and Powell (2006), find that individuals who are angered by someone at work are significantly more likely to speak to people other than the target of their anger than individuals who are angered by someone within the family. Moreover, individuals who are angered by individuals within the family are significantly more likely to say nothing at all (or to engage in non-expressive strategies only) than individuals angered by someone at work (also see Sloan 2004 and Schieman 2000). Although it is clear from quantitative analyses that others play an important role in the management of workplace emotion (Sloan 2004), their participation in the processes of emotion management and expression is less clear. It becomes even more problematic when you take into consideration Lively and Powell’s (2006) analyses, which considers the possibility of speaking to the target and someone else –a strategy that was a more likely workplace strategy than not speaking to anyone at all, but less likely than speaking to someone else only. These findings raise the following questions: Who are the others? What roles do they play in the emotion management process? And how does their participation subsequently affect the quality, rather than the frequency, of emotional expression? Using data from qualitative in-depth interviews with paralegals working in privately owned law firms, I find that individuals employed in a hierarchical work settings engage in emotion management strategies with both similar and dissimilar others rather than express themselves to the rightful target of their anger. These strategies, many of which are reciprocal in nature (Lively 1999) seem not only to change the likelihood of speaking directly to the target, but also to change the tenor of the subsequent expression.

DATA AND METHODOLOGY Focusing on service workers who deal less frequently with the public, or who do so over a more extended period of time, I studied the social support and emotion management strategies of paralegals working in private law firms

CH16-P374095.indd 291

5/26/2008 4:11:58 PM

292

Social Structure and Emotion

(see Lively 2000, 2001).2 Paralegals, by definition, are members of a satellite occupation who assist attorneys in the delivery of legal services (Johnstone and Wenglinski 1985; NALA 2001; Pierce 1995). Because I was interested in the emotion management dynamics that existed within a single occupational category, I excluded secretaries and attorneys from my sample; however, paralegals discussed attorneys and secretaries both as sources of anger and as alternate sources of social support and emotion management. I purposefully sampled paralegals from differently sized law firms in order to see whether the emotion management and emotional labor strategies enacted by paralegals in large firms differed from those enacted in small firms (Lively 2002). The respondents in this study were selected through snowball sampling, a method by which one increases the number of respondents by asking each participant already in the study to recommend others (Weiss 1994). The sample upon which this study is based consists of 43 women, 5 of whom were AfricanAmerican and the rest white. The age of the paralegals ranged from 24 to 58.3 Given that paralegals occupy a fairly nebulous occupational niche, which is partially due to the vagueness of their occupational definition and the inability of paralegal associations to control occupational boundaries and standardize educational requirements, the range of education was from a high school diploma with 25 years of prior legal secretary experience to a Master of Arts in Theology and a 2 year paralegal certification. Salaries also ranged from $16,000–27,000 starting to $35,000–47,500 with 5 to 10 years’ experience. While some of the paralegals worked together, no more than five paralegals were sampled from the same firm. Some of these firms were large generalist firms in which most types of law were practiced, while others specialized in single areas of law such as medical malpractice, consumer bankruptcy, domestic law, workers’ compensation, etc. Some paralegals worked for one particular attorney, but the vast majority was assigned to work for two or more at any given time. The interviews, which typically lasted between 1 and 2 hours, followed a loosely structured outline that elicited the respondents’ stories regarding the demands of their jobs, situations on the job in which they experienced stress, and the strategies that they used to combat stress. I specifically asked: “Can you think of a time when a client or a co-worker had made you angry or disgusted at work and you could not let your feelings show?” If they answered yes (almost all did), I asked, “Why weren’t you able to show what you really felt?” followed

2

For a complete discussion of my selection of privately owned law firms as a site for emotional labor and emotion management and my focus on paralegals in particular, see Lively (2001), also see Pierce (1995). 3 Although I originally interviewed eight male paralegals whose experiences have been recorded in other projects (Lively 2000, 2001, 2002), I have chosen to exclude them from this analysis. Male paralegals tend to have different work experiences than female paralegals which result in qualitatively different emotional experiences and management strategies, if not actual outcomes.

CH16-P374095.indd 292

5/26/2008 4:11:58 PM

293

Status and Emotional Expression

by, “Was there someone else in the office to whom you turned for help in dealing with these feelings?” “Did you do anything else?”4 Through no conscious design, as this study originally went to field in 1995, these questions are very similar to those in the GSS emotions module. While these questions dealt specifically with interpersonal stressors, many paralegals also spoke about engaging in emotion management in conjunction with instrumental and informational stressors when I asked: “Would you describe your job as a high stress job?” and, “What are some of the things about your job that you find stressful?”5 Most of the interviews collected for this study were done in the paralegals’ offices, typically scheduled around their lunch hour. Paralegals’ offices, for the most part, provided a “safe” setting where the respondents could close the door, hold their calls, and speak with me for an extended period with few interruptions. Their offices also provided me with an opportunity to see their workspaces, including their physical distance to other paralegals and to the attorneys with whom they worked, and to experience the general atmosphere within the firm. For those paralegals who did not have their own private workspace but still wanted to meet on the lunch hour, we met inside closed conference rooms that they had reserved for their own use. And for those few who wished to speak to me outside of the office, or beyond the normal work week, I arranged to meet them in a neutral location or in their homes.

RESULTS As noted, sociologists who study emotion in the workplace have paid very little attention to the mechanisms of interpersonal emotion management. Despite the degree to which they are the same phenomenon as social support (Thoits 1995), we know little about how these processes are perceived, how they work, and to what end. In the following sections, I will attempt to answer these questions by highlighting the strategies that paralegals use in backstage areas in order to manage their emotional reactions to a variety of stressors (Lazarus and Folkman 1984), paying particular attention to their experiences of anger, frustration, and irritation when dealing with bosses, customers, co-workers, and subordinates. I will show that they did this by actively seeking others with whom they 4 Although these questions dealt specifically with negative emotions, I also asked if they had ever experienced positive emotions that they felt they could not express. Although a small number recounted instances when they had to hide amusement (especially if it were at the expense of clients or bosses), there seemed to be fewer normative expectations regarding the suppression of positive emotions. 5 Basic demographic information and job descriptions were collected at the beginning of each interview and I ended each interview by asking what type of advice the respondent would give to a new paralegal starting work at her firm. All of the interviews were conducted in person and were tape-recorded so that they could be transcribed at a later date. Names and uniquely identifying details have been changed to assure the confidentiality of the subjects.

CH16-P374095.indd 293

5/26/2008 4:11:58 PM

294

Social Structure and Emotion

routinely reciprocated emotion management assistance, as well as social support more generally, and that their choices of others depended largely on their perception of support availability, the nature of the stressor that caused the emotional reaction to begin with (informational, instrumental, or interpersonal), and the corresponding assistance that they required. SPEAKING TO SOMEONE ELSE

When asked to relate instances in which individuals in her office turned to her for emotional management assistance, Terri Malone, a 39 year old medical malpractice paralegal with 6 years of experience, described an incident involving another paralegal who had been upset by the attorney for whom they both worked. Earlier in her interview, Terri explained that paralegals, at least the ones in her firm, were expected to accept being treated “badly” by attorneys, because paralegals are responsible for assisting attorneys with their negative emotions. Despite their understanding of their social “place” within the firm (Clark 1990), some paralegals had trouble dealing with their attorneys’ misdirected anger, or rudeness. Terri explains: But, a better example would be … not an intentionally mean word, but, a short word [from an attorney] being used toward a legal assistant – and they do that. That happens to all of us – all of the legal assistants, the secretaries. [We’ve] all experienced it, and, this person [co-worker], you know is in tears and comes to me and I say, “Look at the source. Don’t take it to heart.” I mean, but then when it happens to you, you do the same thing, and then they’re the ones … And they’ll be in there, “Like, oh no. What did he say this time?” And it’s not intentional … I said, “Don’t. It’s not intentional. It’s just the way he is.”

As Terri’s comments suggest, she and the paralegals (and some of the more senior secretaries) in her office experienced emotion management not only as an individual level or private process, but as something that is inherently collective or collaborative in nature (Hochschild 1983; Staske 1996, 1998). Based in large part on empathetic understanding, if not personal experience, they turn to one another to express their negative felt emotions and for assistance in managing those emotions. Moreover, Terri’s comments betray a degree of reciprocity and a consideration of time not typically captured in quantitative studies of emotion management or emotional expression: “I mean, but when it happens to you, you do the same thing, and then they’re the ones … [giving help to you].” I also interviewed the paralegal referenced by Terri in the above passage, 27 year old Amy Westphal, who had 3 years of experience. As might be expected, when I asked Amy what it was about her job that made it stressful or anger producing, she cited the same attorney as the root cause of most of her job-related stress; she also mentioned Terri when I asked whether she had ever received or provided emotional assistance to anyone in her workplace regarding issues at work. In many ways, Amy’s response is similar to Terri’s: Most of the time it’s not a procedural problem, or a work-related problem. Most of the time it’s like stress or just … time to blow off and they just want to talk. But I mean that – it’s very obviously work-related. Uhm, basically I just listen and let them talk, and let

CH16-P374095.indd 294

5/26/2008 4:11:58 PM

Status and Emotional Expression

295

them just spout it out, because that’s how I prefer it … So I let them just talk until they’re done and then I try to kind of build up their self-esteem because I would feel that that’s how I would want to feel.

Note that Amy bases the advice that she gives others on the type of advice she would hope to receive (and had already received) in similar situations; but, when pressed for a concrete example, Amy recounted an interaction that she had with Terri the previous month: [Terri] came to me with problems with the same attorney. She was very frustrated because he asked me to do a task that was a duplicate task of one that he had asked her to do the day before. And she construed it as, “I’m inadequate. I’m inept, therefore, he is asking you.” I [just] laughed and said, “Surely you don’t think that he came to me for fear that you wouldn’t do it, you know, correctly. Surely you don’t think that.” [And she said,] “Well, yeah. He doesn’t trust me.” And I’m like, “You’re crazy!!” You know of course I didn’t say that, but I said, “You know, I think you need to take a few steps back. I think you need to re-evaluate your feelings …” I [also] try to … build her up, to try to take her mind off of it and tell her that she needs to just kind of re-evaluate, go home tonight, take a bath, light a bunch of candles, and just think … And to do as she always tells me: “Don’t take it personally.”

Here, Amy begins by allowing Terri to express her frustration, anger, and other-related negative emotions before attempting to help her reconceptualize not only the attorney’s hurtful request, but also her image of herself (Lively and Heise 2004; also see Francis’ 1997 discussion of interpersonal emotion management as identity transformation). In addition to these more immediate cognitive strategies, Amy also suggests a delayed behavioral strategy – a candlelit bath that will also assist Terri in adjusting her physiological reaction or physical tension that often accompany the experience of strong, negative emotions (Thoits 1986). As indicated in both Terri’s and Amy’s comments, paralegals use one another for the articulation and expression of negative emotion, as well as for assistance in moving from an internal attribution of blame (“I’m inadequate. I’m inept …”) to a more external attribution (“Look at the source … It’s not intentional. It’s just the way he is …”). These emotion management processes seem to be separate from more practical types of social support seeking and may in some ways (as I will show below), be necessary in order for more active social support (i.e., instrumental or informational) to take place. Pam Miller, a 58 year old medical malpractice paralegal at a different firm related a very similar incident, in which one of the senior attorneys in the firm instructed her to look over documents that were part of another paralegal’s case: Lou [the other paralegal] became very upset – I knew she would be. But I told her the truth of what went on, and she said, “Then it’s your case.” And I said, “It’s not my case. It’s your case.” And she said, “Well, I don’t want him having you to linger and looking over my shoulder.” I said, “That’s not why he did it – consider who we’re talking about. You know his personality. He only did it because I’m involved in this other case and he doesn’t think I’ve been paying him enough attention …”

In each of these three accounts, the paralegals used their own personal knowledge of the attorneys’ personalities to excuse their rude or demanding behavior and

CH16-P374095.indd 295

5/26/2008 4:11:58 PM

296

Social Structure and Emotion

help one another change the meaning that the situation had for them. By telling each other that the attorney was simply rude or insensitive, they were able to protect each other from internalizing the hurt and the blame that they might have otherwise felt had they believed the attorneys were justified in their behaviors. These three accounts also involved elements of “mothering,” a form of gendered emotional labor identified by Pierce (1995), where the paralegals attempted to build up their co-workers’ self-esteem through the provision of socio-emotional aid, even at their own expense. When asked if she ever went to or received emotional assistance from Lou, Pam laughed, and brought up the subject of venting (emotional expression to someone other than the rightful target) – perhaps the most common type of emotion management conducted by the paralegals in this study. As the following quotations illustrate, venting not only allowed paralegals to assist one another in redefining the situation cognitively, but may also have assisted them in altering their respective levels of physiological arousal, while keeping them in touch with their felt emotions: Oh gosh yes … We close our door [to our office] and just uh, really have our say to one another. If we didn’t have one another to do that with, I don’t know [what we’d do] … we may kick the trashcan … whatever it takes …

Ann Watson, a litigation paralegal working in a mid-sized firm in another city, reported having a similar relationship with Frances Blakemore (another litigation paralegal): they regularly used one another to vent, or to blow off steam. Backstage areas, such as closed offices or restaurants outside of the reach of attorneys or office managers, provided paralegals, in a real sense, with protected spaces in which they could express their negative emotions and share their collective frustration (also see Goffman 1959). Ann explained: Now I would have and I did go to one of the paralegals and generally she’s the one – and she does the same thing. We’re both frustrated. We’re just like, “Ahahahhhhh!” You know, and we’ll go in [and shut the door] and just put our hands around each others’ throat and go, “I’m gonna kill somebody!!!” And then we just laugh, because she’s in the same position … And I think it’s pretty much the same thing when she comes to me [for support]. I mean she’s frustrated because some of the people – some of her staff for whatever reason didn’t get something done or whatever and she’s coming in to me. You know, “Somebody’s just dropping the ball and I think, I’m going to kill them.” Yeah – and I mean, it’s probably an everyday thing …6

Ann’s quotation, like Lou’s, above, represents the collective and collaborative nature of emotion management first introduced by Hochschild (1983) and

6 In this quote, Ann states that Frances also vents the negative emotions that she experiences as a result of events involving workers holding occupational positions subordinate to hers. As will be addressed below, many paralegals are motivated to hide their negative emotions from secretaries or clerks, or to maintain positive emotional relationships with them, as a means to procure instrumental assistance during office speed-ups, i.e., depositions, trials, court filings, bankruptcy mailings, etc.

CH16-P374095.indd 296

5/26/2008 4:11:58 PM

Status and Emotional Expression

297

Staske (1996, 1999), but it also reveals a degree of reciprocity not found in other descriptions of emotion management or emotional expression (see Lively 2000). Indeed, one of the things that make this type of emotion management different from other types of emotion management enacted in law firms (e.g., that enacted by paralegals for the benefit of attorneys), aside from its mutually rewarding properties, is that both parties have a normative right to instigate emotion management, just as both parties have a normative obligation to respond. Moreover, Ann and Frances do not simply turn to one another one time, but they do so regularly over time and, indeed, their ability to turn to one another for emotion management assistance becomes an important foundation upon which their working relationship, as well as their friendship outside of the office, is based. Jennifer Martin, a commercial litigation paralegal who worked in a small work group (or team) with two other paralegals, commented similarly: … Both Morgan [one of the other paralegals in her workgroup] and I work with high maintenance clients where they call all the time and they’re always like, “Did you do this? Did you do that? We got to get this done. Make sure that you get it done.” And sometimes I’ll just walk in [her office] and say, “That jerk.” He’s such a jerk. He said, “Blah, blah, blah – and he makes me so mad.” And she’ll say, “I know. I can’t believe that he said that – he’s really a jerk.” And then I feel better [because] someone agrees with me that he’s a jerk. And then I say, “Okay, I’m going back to work” – I mean, almost every day [this happens] … … And once you say it and get it off your chest, and someone agrees with you that you’re right, which you already knew that you were right … then you just go back to work. That happens, definitely, several times a day.

Once again, Jennifer describes a sequence of steps that she and Morgan reciprocate among themselves. They first seek a relatively safe space; they then assist one another in moving from an internal to an external attribution and construct (or at the very least reaffirm) a shared meaning of themselves as right and the target of their anger as wrong. According to these statements, it seems that the paralegals in this study use one another as safety valves. In other words, even though paralegals hide their negative emotions that are normatively prohibited from being displayed in front stage areas, they go to one another to vent their negative emotions – if for no other reason than having their individual, as well as their collective images, of the situation reaffirmed by a third party. Even though the process may take just a few moments, it is an everyday occurrence and is considered, at least by the paralegals, as essential to their ability to manage their negative feelings and “get back to work.” As suggested by the above, expressing their anger to “similar others” was by far the most common emotion management strategy mentioned among paralegals. However, some did report also turning to attorneys, secretaries, and, in some cases, even administrators. Those who did turn to attorneys tended to do so when angered by a common threat (e.g. a problematic client), just as those who tended to vent to secretaries tended to do so when they shared the same problematic attorney. In this regard, the paralegals may have been choosing similar

CH16-P374095.indd 297

5/26/2008 4:11:58 PM

298

Social Structure and Emotion

others, whose similarity was based on situational factors rather than traditional status markers. As Laura Garrett and Marion Cartwright explain: I was really perturbed … and, of course, [I called my boss] and was like, “That guy’s crazy … How come this couldn’t have happened when you were here?” You know, and he [my boss] was all frustrated on the phone – he was walking around wanting to yell and curse. When he got on the phone with me, he was like, “Just tell him to go to hell …” But technically we can’t do that, because the guy that brought him in, as a client, is a managing partner, so I didn’t want to tick him off too much, because I didn’t want to get the managing partner upset … [But it was good to be able to talk to him because] it was like we were on the same page … it made me feel pretty good that we were, you know, on the same page – that he understood my frustration as much as I understood his (Laura Garrett, white female, age 27, 6 years experience, mid-sized firm). And one of the reasons that I have a good friendship/relationship with my boss’ secretary is we vent to each other about what’s going on. That really helps me a lot. I mean, it seems like she’ll come and in say, “He’s driving me crazy today,” or you know … just things that come up or the personality of the attorney. The attorney I work for is very moody, for example, very moody. One minute he’s all right then the next minute he growls at you … and that can be very difficult to deal with (Marion Cartwright, white female, age 40, 18 years experience, mid-sized firm).

While all of the above paralegals chose individuals who could help them manage their emotions through venting, others sought of people who could actually help them change the situation (see Thoits 1990). One paralegal explains: … if James Henry … makes me angry, I don’t tell him, because … he’s too – anger doesn’t work on him, because he’s such a baby, so I’ll go tell somebody, like I’ll go tell Daniel (a more senior attorney in their work group). Daniel will go then go take care of it.

And although not the original focus, almost 40 percent (N=20) of the paralegals in this study also mentioned turning to significant others outside of the workplace for assistance in handling their anger and frustration with those in the workplace. However, despite that almost half of the sample had tried, it was the general understanding that expressing anger and frustration about work and co-workers to friends and family members outside of the workplace was generally less effective than expressing it to those within. In particular, a number of paralegals felt like their spouses or domestic partners were often uninterested (or didn’t really care), and, in the cases where their anger involved clients, many felt like they were limited in the amount of detail they could actually disclose without risking client confidentiality. SPEAKING TO BOTH THE TARGET AND SOMEONE ELSE

Though paralegals in this study routinely vented negative emotions to others instead of the rightful target (i.e., to others like themselves) without expressing anger to the target of their anger (be it a supervisor, an attorney, or a client), a smaller number of paralegals also reported speaking to both the target and someone else (again, see Lively and Powell 2006). It is important to note, however, that only in a handful of cases did individuals seek out others after they had expressed their unmanaged anger to its target, in part because expressing

CH16-P374095.indd 298

5/26/2008 4:11:58 PM

Status and Emotional Expression

299

unmanaged emotion was typically viewed as a personal failure, if not a breach of professional conduct (Lively 2001). Ann explains: I hung up the client – okay? I was going, “Oh my gosh, no one speaks to me that way,” but I am not one of those people, either that stands up and says, “Knock it off and don’t talk to me that” – especially when he’s the client … I just said, “Ron, I have to go,” and I hung up the phone. I just cut him off in mid-sentence – just chopped him [off]. And I found myself really teary-eyed all that afternoon and hours afterwards (Ann Watson, white female, age 36, 16 years experience, mid-sized firm).

As did Pam: P:

I, probably, am the only one I know in that office that has lost their temper, verbally, or threw a temper tantrum [in public] – ever … K: And you said that that’s [only] happened a couple of times? P: Just a couple of times. K: That’s not bad in sixteen years. P: But it was bad at the time … I was very embarrassed. [I was] extremely embarrassed – almost to the point of quitting because I didn’t think I could face anybody [afterwards]. (Pam Miller, white female, age 48, 16 years experience, small firm). Even though both of these women reported feeling extreme embarrassment and a desire to avoid their colleagues after the fact, both eventually turned to what Pam referred to as their “structural equivalents” in order to feel better about their so-called inappropriate displays. Andreea, however, felt no embarrassment about losing her temper with a client and her son, both of whom she eventually evicted from her office, under threat of removal: A: Oh, I cried that time too, because I was mad, oh yeah, I was very angry. K: Was there anyone at the office, at that time? A: Oh yeah. There was a paralegal, she was … she was right there listening to the whole thing. She said, “Well, you know, if it had gotten real bad, I was going to send somebody,” and she heard the whole thing, and you know, I vented my frustration to her. I had to actually leave though, for a little while and just kind of walk, because I get real nervous. When I get mad, I shake and everything, and I really had to go gather myself once that happened. In other cases, still, the paralegals were spoken to by others (namely office administrators) after they had lost their temper with their attorney, as this paralegal explains: … And I honestly can’t tell you what – to this day – what he said to me that just pushed the wrong button and I said, “Look!” I said, “I told you that I will have that, and I will get it to you when I’m done. Please do not tell me that again!” You know? And I mean I just,

CH16-P374095.indd 299

5/26/2008 4:11:58 PM

300

Social Structure and Emotion

and I mean, that’s not me. I just, I never … I’m usually not a mean person, by any means, and tried to say it as calmly and as nicely as I possibly could. Well, he went to the office manager and kind of tattled on me. That’s what I call it: tattling. He didn’t come to me and say, “Look, you know, we need to talk about this.” He ran to the office manager and said, “She said something to me that I didn’t like and you need to talk to her about it.”

When she actually spoke with the office manager about the attorney’s behavior (and was spoken to in turn), she received advice similar to that provided to others by peers, albeit in slightly harsher words: “You’re just gonna have to suck it up, because that’s the way he is and he’s not going to change, and there’s nothing I can say to him.” While these four paralegals only reluctantly accepted the emotional assistance from co-workers and colleagues after a blow up, others reported using co-workers to blow off steam, first, as a way to be able to express their anger in more managed (or “professional”) ways (see Lively 2001). Speaking of a particularly ugly encounter that she had with an attorney that she labeled the “true test of professionalism,” Patricia remarked that it was all she could do not to respond to his level. When I asked her why she considered that particular story to represent a true test of professionalism, she replied: P:

Because I don’t believe a professional should yell at another person. You should be able to ask someone to do something in a tone of voice that still lets the person there’s an immediate need. And [for] me not to spout off at him and to say what I was truly thinking – it took self-control to do that. And a lot of things that you do – I think, as a professional, you should not show your true feelings. K: Do you think that he was acting in a professional manner? P: Oh, absolutely not. I don’t believe that he should have wagged his finger at me. I think he should have communicated his desire to get [whatever he needed done] in a manner that was not belittling of me. (Patricia Warner, white female, age 40, 8 years experience, large firm) When asked how this situation resolved itself, Patricia relayed that she had first talked to her closest colleague using many of the same strategies outlined above. Once she had her anger at a more manageable level, however, she did go back and voice her displeasure to its rightful target. To her surprise, their relationship actually improved after this incident; in fact, the attorney in question seemed to treat her with more, rather then less, respect. Indeed, Patricia wasn’t the only paralegal in the sample who actually improved her relationship with her attorney by expressing managed anger; a finding consistent with experimental studies that illustrate that people tend to give more, not less, respect to those who engage in displays of anger (Tiedens 2000), as Marion also revealed: “I definitely took it [my anger] someplace else, da, da, da, [but] then probably within the next two weeks, we had talked about it. I talked to him about it … and that was years ago. [We have a great relationship now].”

CH16-P374095.indd 300

5/26/2008 4:11:58 PM

Status and Emotional Expression

301

CONCLUSION As Hochschild noted over 20 years ago now, emotions can be, and often are, subject to acts of management. To date, there has been considerable interest in the ways in which individuals manage their emotions in workplace settings, where they are required to bring their emotions in line with corporately defined feeling and display norms in order to elicit a response from both clients and other higher status workers. Outside of the workplace, scholars have documented actors’ attempts to meet less coercive feeling norms (for themselves), as well their attempts to change others’ emotions for both personal and political gain. Regardless of whether the focus is on emotional labor that occurs at work or emotion work that occurs in more private sectors (such as the family), emotion management has, for the most part, been limited to the cognitive strategies that individuals direct at themselves in order to bring about a desired feeling state in another (see Hochschild 1983). Less attention has been paid to the more direct roles that others play. Unlike previous analyses of emotion management in the workplace that are usually limited to episodic encounters between actors of unequal status or that occur within the eyesight and earshot of customers or bosses, the analysis discussed above examines more reciprocal forms of emotion management that recur among actors of same or similar status in backstage areas of workplaces, or in Hochschild’s “other field of emotion work” (1983: 114). This is a field that has, thus far, remained relatively unexplored by those interested in workplace emotion. Furthermore, unlike previous work that has focused strictly on the negative consequences of emotion management (especially emotional labor) for individuals (i.e., distress, feelings of inauthenticity, decreased mental and physical well being, etc.), my study also illustrates the more positive or pro-social outcomes that these processes may yield. Such an analysis stretches our previous understanding of emotion management from something that is solely self-directed (much like coping) to include processes that are reciprocated between provider and recipient in ways that benefit both parties and, moreover, have the potential to sustain personal as well as professional relationships within hierarchically organized settings. While a recent study of work and family life has suggested that the work site has become an important haven from the demands of home as corporations have become more sophisticated in their application of emotion norms and the creation of emotion cultures (Hochschild 1997), the roles that emotion management and/or emotional expression among co-workers and colleagues play in the analysis is surprisingly small. Although Hochschild (1997) shows that parents and spouses turn to co-workers in order to compensate for their difficulties at home (and to see themselves as competent, appreciated, etc.), she barely mentions the ways in which workers seek out co-workers in order to compensate for the difficulties that they also face at work (but see Gatta 2002; Lively 2000, 2002). This study, in contrast, shows that individuals actively choose others at work from

CH16-P374095.indd 301

5/26/2008 4:11:59 PM

302

Social Structure and Emotion

whom they solicit (and/or to whom they provide) emotion management assistance, often in the form of venting, damage control, or the joint reidentification of a problematic actor – typically either someone of higher status or representing a common threat. Others, still, chose to seek out more drastic forms of intervention, managing their emotions, instead, by confiding in someone who could change the situation for them literally, as opposed to cognitively, for example, a higher ranking attorney or an administrator. The observation that individuals get together, gripe, moan, and exchange emotional and other forms of social support in their attempts to manage their reactions to work at work is not new. What is new, however, is the extent to which others – loosely defined – influence individuals’ attempts to manage and express emotion in workplace settings (Lively and Powell 2006; Sloan 2004; Stets and Tsushima 2001). As noted above, expressing anger (and other closely related negative emotions) to individuals’ in addition to and/or in place of the rightful targets of those emotions seems to be the most common strategy among individuals angered in the workplace – particularly those angered by someone of higher status (e.g. a boss, a supervisor, or a client; see Lively and Powell 2006). The primary purpose of this paper is to elucidate a series of survey results reported elsewhere (Lively and Powell 2006; Sloan 2004; Stets and Tsushima 2001) by offering insight into who these “others” may be and what influence they may have on both emotional expression and emotion management. The secondary purpose is to address the need for a more complementary methodological approach to the study of emotions. Over the last 20 years, the sociology of emotion has expanded its methodology to include such diverse approaches as experiments, fieldwork, in-depth interviewing, computer simulations, conversational analysis, and, most recently with the creation of the GSS emotion module, large scale survey research (also see Erickson and Ritter 2001). To date, however, few studies of emotional outcomes, or processes, attempt to utilize more than one approach to any given question. As I have argued elsewhere (Lively 2006; Lively and Powell 2006), the sociology of emotion, like other emerging subfields, would continue to benefit from a multi-methodological orientation in which surveys complement and are complemented by these other methods (also see Erickson and Ritter 2001; Simon 1995; Wharton 1993, 1999). As I hope this chapter illustrates, emotions scholars need to incorporate both qualitative and quantitative methods to develop both a depth and breadth of understanding that neither can achieve alone. Having said that, taking the insights from the qualitative findings presented here, it would be interesting to see whether these findings, too, are generalizable to a broader nationally representative sample. In order to test these assertions in a more random sample, we would need, not surprisingly, additional survey data. Although the GSS (1996) emotion module has taken us a long way in determining the way emotion, emotion management, and emotional expression are distributed within a nationally representative sample, there are still many questions that remain unanswered, or rather, unasked. As illustrated in this paper,

CH16-P374095.indd 302

5/26/2008 4:11:59 PM

Status and Emotional Expression

303

others play an important role in emotional expression and or emotion management. In order to test whether these observations from a very particular status hierarchy, governed by its own brand of behavioral and feeling norms, hold in a less specialized population, we need a better understanding not only of who these others are and what types of assistance they provide, but also about the triggering events themselves. For example, was someone angered because they were not given enough information or attention to complete a particular task, or were they angered by a problematic interpersonal counter, or perhaps both? Do the events and the types of stressors – all of which may result in feelings of anger, irritation, or frustration – influence how individuals respond to it? Do they influence the people from whom individuals choose to seek emotion management assistance? A related issue is that we lack a truly systematic understanding of who these others actually are. This work suggests that individuals are more likely to speak to others at work (and is in keeping with Thoits’ (1986) observations regarding similar others being the best source of social support and, subsequently, emotion management). However, paralegals are bound by a code of ethics not to discuss certain aspects of their work outside of the office (NALA 2001); others are not. What events are more likely to motivate seeking emotion management assistance across status lines and, closely related, are some workers more likely to seek assistance across status lines than others? Finally, this chapter also brings our attention back to the fact that emotion management and emotional expression are both interrelated processes. Some of the individuals cited here vented first then confronted later, while others confronted first then had their sense of self put back together by sympathetic others. Future efforts at data collection would profit from including questions that better account for the dynamic and often extended ways in which individuals manage and express their emotions as part in parcel of their daily interactions with both similar and dissimilar others.

ACKNOWLEDGMENT This manuscript was partially funded by NIMH Grant #T32 MH14588 and the Rockefeller Center for Public Policy at Dartmouth College. I would like to thank Peggy Thoits, Robin Simon, Brian Powell, and Laura Bozzuto for reading earlier drafts of this manuscript and for all of their insightful comments.

REFERENCES Clark, Candace. 1990. “Emotions and Micropolitics in Everyday Life: Some Patterns and Paradoxes of “Place”.” In: Theodore Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 305–333. Albany, NY: State University of New York Press.

CH16-P374095.indd 303

5/26/2008 4:11:59 PM

304

Social Structure and Emotion

Collins, Randall. 1990. “Stratification, Emotional Energy, and the Transient Emotions.” In: Theodore Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 27–57. Albany: State University of New York Press. Erickson, Rebecca J. and Christian Ritter. 2001. “Emotional Labor, Burnout, and Inauthenticity: Does Gender Matter.” Social Psychology Quarterly 64:146–163. Francis, Linda. 1997. “Ideology and Interpersonal Emotion Management: Redefining Identity in Two Support Groups.” Social Psychology Quarterly 60(2):153–171. Gatta, Mary L. 2002. Juggling Food and Feelings. Rowman & Littlefield Publishing. Goffman, Erving. 1959. The Presentation of Self in Everyday Life. New York: Anchor. Hegtvedt, Karen A. 1990. “The Effects of Relationship Structure on Emotional Responses to Inequity.” Social Psychology Quarterly 53:214–228. Hochschild, Arlie R. 1979. “Emotion Work, Feeling Rules, and Social Structure.” American Journal of Sociology 85:551–575. Hochschild, Arlie R. 1983. The Managed Heart: The Commercialization of Human Feeling. Berkeley: University of California Press. Hochschild, Arlie R. 1997. The Time Bind: When Work Becomes Home and Home Becomes Work. New York: Metropolitan Books. Johnstone, Quentin and Morris Wenglinski. 1985. Paralegals: Progress and Prospects of a Satellite Occupation. Westport, Connecticut: Greenwood Press. Kemper, Theodore. 1978. A Social Interaction Theory of Emotions. New York: Wiley. Kemper, Theodore and Randall Collins. 1990. “Dimensions of Microinteraction.” American Journal of Sociology 96:32–68. Lawler, Edward J. and Shane R. Thye. 1999. “Bringing Emotions into Social Exchange Theory.” Annual Review of Sociology 25:217–244. Lazarus, Richard S. and Susan Folkman. 1984. Stress, Appraisal and Coping. New York: Springer. Lively, Kathryn J. 1999. Reciprocal and Joint Emotion Management: New Directions in Emotion Management and Social Support. Ann Arbor, MI: UMI Dissertation Services. Lively, Kathryn J. 2000. “Reciprocal Emotion Management: Working Together to Maintain Stratification in Private Law Firms.” Work and Occupations 27:32–63. Lively, Kathryn J. 2001. “Occupational Claims to Professionalism: The Case of Paralegals.” Symbolic Interaction 24:343–366. Lively, Kathryn J. 2002. “Upsetting the Balance and Evening the Field: The Effects of Client Contact on Emotional Labor.” Work and Occupations 29:198–225. Lively, Kathryn J. 2006. “Emotion in the Workplace.” In: Jan E. Stets and Jonathan H. Turner (eds), Handbook of the Sociology of Emotions, pp. 569–590. New York: Springer. Lively, Kathryn J. Forthcoming. “Emotional Segueing and the Management of Emotion by Women and Men.” Social Forces. Lively, Kathryn J. and David R. Heise. 2004. “Sociological Realms of Emotional Experience.” American Journal of Sociology 109:1109–1136. Lively, Kathryn J. and Brian Powell. 2006. “Emotional Expression at Work and at Home: Domain, Status, or Individual Characteristics?” Social Psychology Quarterly. Lively, Kathryn J., Lala Carr Steelman, and Brian Powell. 2004. “Emotion, Housework, and Equity.” Paper presented at the Annual Meetings of the American Sociological Association. San Francisco, CA. Lovaglia, Michael J. and Jeffrey A. Houser. 1996. “Emotional Reactions to Status in Groups.” American Sociological Review 61:867–883. Mabry, J. Beth and K. Jill Kiecolt. 2005. “Anger in Black and White: Race, Alienation, and Anger.” Journal of Health and Social Behavior 46:85–101. National Association of Legal Assistants 2001. Annual Report. National Association of Legal Assistants, Tulsa, Oklahoma. Pierce, Jennifer. 1995. Gender Trials: Emotional Lives in Contemporary Law Firms. Berkeley: University of Berkeley Press.

CH16-P374095.indd 304

5/26/2008 4:11:59 PM

Status and Emotional Expression

305

Ridgeway, Cecilia L. and Cathryn Johnson. 1990. “What Is the Relationship Between Socioemotional Behavior and Status in Task Groups?” American Journal of Sociology 95:1189–1212. Ridgeway, Cecilia L. and Lynn Smith-Lovin. 1999. “The Gender System and Interaction.” Annual Review of Sociology 25:191–216. Schieman, Scott. 2000. “Education and the Activation, Course, and Management of Anger.” Journal of Health and Social Behavior 41:20–39. Simon, Robin W. and Leda E. K. Nath. 2004. “Gender and Emotion in the United States: Do Men and Women Differ in Self-Reports of Feelings and Expressive Behavior?” American Journal of Sociology 109:1137–1176. Sloan, Melissa. 2004. “The Effects of Occupational Characteristics on the Experience and Expression of Anger in the Workplace.” Work and Occupations 31:38–72. Staske, Shirley A. 1996. “Talking Feelings: The Collaborative Construction of Emotion in Talk between Close Relational Partners.” Symbolic Interaction 19:111–142. Staske, Shirley A. 1998. “The Normalization of Problematic Emotion in Conversations Between Close Relational Partners: Interpersonal Emotion Work.” Symbolic Interaction 21:59–92. Staske, Shirley A. 1999. “Creating Relational Ties in Talk: The Collaborative Construction of Relational Jealousy.” Symbolic Interaction 22:213–237. Steinberg, Ronnie and Deb Figart. 1999. Emotional Labor in the Service Economy: Annals of the American Academy of Political and Social Science. Thousand Oaks, CA: Sage Periodicals Press. Stets, Jan and Teresa M. Tsushima. 2001. “Negative Emotion and Coping Responses Within Identity Control.” Social Psychology Quarterly 64:283–295. Sutton, Robert. 1991. “Maintaining Norms about Expressed Emotion: The Case of Bill Collectors.” Administrative Science Quarterly 36:245–268. Thoits, Peggy A. 1986. “Social Support as Coping Assistance.” Journal of Counseling and Clinical Psychology 54:416–423. Thoits, Peggy A. 1990. “Research Agendas in Emotional Deviance.” In: Theodore D. Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 180–206. Albany, NY: State University of New York Press. Thoits, Peggy A. 1995. “Stress, Coping, and Social Support Processes: Where Are We? What Next?” Journal of Health and Social Behavior (Extra Issue):53–79. Thoits, Peggy A. 1996. “Managing the Emotions of Others.” Symbolic Interaction 19(2):85–109. Tiedens, Larissa Z. 2000. “Powerful Emotions: The Vicious Cycle of Social Status Positions and Emotions.” In: Neal M. Ashkanasy, Charmine E. J. Hartel, and Wilfred J. Zerbe (eds), Emotions in the Workplace: Research, Theory, and Practice, pp. 71–81. Westport, CT: Quorum. Weiss, Robert. 1994. Learning from Strangers: The Art and Method of Qualitative Interview Studies. New York: The Free Press. Wharton, Amy S. 1993. “The Affective Consequences of Service Work.” Work and Occupations 20:205–232. Wharton, Amy S. 1999. “Feeling it and Faking it: Understanding the Consequences of Emotional Labor.” In: R. Steinberg and D. Figart (eds), Emotional Labor in the Service Economy. Annals of the American Academy of Political and Social Sciences, pp. 158–176. Thousand Oaks, CA: Sage.

CH16-P374095.indd 305

5/26/2008 4:11:59 PM

PA R T

VI Social Change

CH17-P374095.indd 307

5/27/2008 7:26:14 PM

17 Commentary Ellen M. Granberg Clemson University, Clemson, SC, USA

Serious consideration of the role of emotion in social processes is a relatively recent addition to much of sociology. Gradually, however, concern with feelings, sentiments, moods, and passions has begun to assume a more central place in the discipline. One arena in which the inclusion of emotion has been particularly fruitful is in work on social change and social movements. In this essay, I trace the rise, fall, and reemergence of emotions among movement scholars and discuss two papers that both validate and challenge current thinking on emotions and social change.

EMOTION AND SOCIAL MOVEMENTS: AN INTELLECTUAL BIOGRAPHY In the decades before the 1960s, emotions played a relatively prominent role in scholarship on social change and social movements. Collective action (especially in the form of protests, riots, and revolutions) was typically conceptualized as a form of mass insanity and participants were seen as ruled by powerful, irrational emotional impulses (Aminzade and McAdam 2002). The social movements under study at that time, particularly those involving Marxist insurgencies, were themselves widely viewed as a threat to social stability and this perspective reinforced a view of activists as irrational actors (Goodwin, Jasper, and Polletta 2000). Thus, while early social movement and collective behavior researchers considered emotion to be sociologically relevant, they also conceptualized it chiefly as a threat to existing social structures and as fundamentally incompatible with rational motivations (Williams 2000). Social Structure and Emotion

CH17-P374095.indd 309

309

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/27/2008 7:26:14 PM

310

Social Structure and Emotion

By the 1960s, movement researchers had turned away from the view that advocates of social change were motivated by emotional pathology and focused instead on the rational bases of collective action (Hercus 1999). A number of factors contributed to this transition; at base, the types of movements available for study had changed, and increasing numbers of academics were personally sympathetic to causes such as civil rights and the anti-war movement (Goodwin, Jasper, and Polletta 2000). In many cases, scholars became personally involved in these movements and saw that rather than being driven by maladjustment or motivated by an antisocial temperament, movement participants generally expressed highly rational motivations and engaged in strategic and purposive action to achieve their goals. At the same time, sociologists changed their focus from explaining why collective action arose to explaining how it arose – this move also involved a switch to theories of movement development that emphasized activists as rational actors. The reframing of social change processes as steeped in logic, strategy, and rationality set the stage for important advances in social movement scholarship and led to the generation of theories that form the backbone of current work in the area. Emotions, however, had little place in these developments. In cases where emotion was clearly implicated, such as its role in motivations to join protest movements, it was always routed through a cognitive or rational frame. In this way, the social movement theories dominant from the 1960s through the 1980s maintained the strict separation between rationality and emotion that characterized sociological thought more generally (Massey 2002). At the end of the 1980s, another confluence of theoretical innovation and personal experience sparked a reconsideration of emotion and social change (Groves 1995). By this time, the sociology of emotion had been established as a unique interest area within sociology. Much of the earliest work in this area focused on the existence of “feeling rules” which established expectations for both the feelings one displays and those one experiences (Hochschild 1983). The recognition that feeling rules can be specific to cultures, sub-cultures, and social groups re-opened the door to consider the influence of emotion on social activism and activist organizations (Groves 1995). At the same time that emotions were being introduced to sociology more broadly, increasing numbers of women were working as active scholars, changing the discipline’s sensibilities in ways that were conducive to more robust treatments of emotion. For example, feminist scholars began to challenge overtly the assumptions that kept rationality and emotion separated (Hercus 1999). In particular, they problematized “reason–unreason dichotomies” (Goodwin, Jasper, and Polletta 2000) as indicative of a larger structure of gender inequality in which womanhood and femininity were equated with emotion and irrationality, justifying and reproducing the subordinate status of women. This facilitated a rethinking of the relationship between emotion and reason – one in which the two could be treated as interdependent, mutually constraining, and mutually reinforcing rather than as an oppositional dichotomy (Williams 2000).

CH17-P374095.indd 310

5/27/2008 7:26:14 PM

Commentary

311

Goodwin and colleagues have argued that these events, along with a renewed interested in the social psychological dynamics driving social action have brought about a resurgence of interest in the relationship between emotion and social change (Goodwin, Jasper, and Polletta 2000). Movement scholars have begun to consider again the implications of emotion for collective action, generating an upsurge of work in this area. In remaking a place for emotion, researchers have returned to some of the original ideas found in work on collective behavior (Beyerlein and Young 2005; Klatch 2004) but have also incorporated concepts from the sociology of emotion, social psychology, and cultural studies. A full description of this body of research is too vast to describe adequately in the space available. Instead, I will only review a few of the major ideas emerging from this research and, in that process, describe how the two papers included in this section both augment and challenge it.

THE RETURN OF THE REPRESSED – EMOTION IN SOCIAL MOVEMENTS AND COLLECTIVE ACTION At the conclusion of his 1997 article on emotions and opportunity structures, Jeff Goodwin advocated greater attention to the role of emotion in social change research by calling for a “return of the repressed” – a reconsideration of emotion as a significant explanatory mechanism in social movement and collective action research. In the decade that has followed, much of the interest in emotions and social change has been rooted in the ability of emotion to explain links between ideas and action (Calhoun 2001). Emotion is often described as providing the “muscle” that makes mobilization possible (Reed 2004) in that it demonstrates how “moral norms and injunctions come to have force” (Calhoun 2001:50). Early work incorporating emotion back into collective action research took two primary approaches. In many instances, concepts derived from the sociology of emotion such as emotion work, feeling rules and emotional deviance have been applied to social movement studies as a way of capturing affective dynamics operating during collective action. This approach was used by Perry (2002) to examine strategies employed during the Communist Revolution in China. She found that the Maoist forces made use of emotion work (Hochschild 1983) strategies to incite and sustain peasant rebellions against aristocratic landowners. Similarly, Verta Taylor (1995) applied the notion of emotional deviance (Thoits 1985) to analyze how participants in post-partum depression support groups reframed traditional gender ideology about mothers’ “natural” love for their children, arguing it was not as natural or simple as most people believe. A second approach taken by movement scholars has been to reconsider core concepts long dominant in social movement research, assessing how attention to emotion might alter or augment them. For example, collective action researchers have long understood the importance of network structures in recruiting and

CH17-P374095.indd 311

5/27/2008 7:26:14 PM

312

Social Structure and Emotion

sustaining movement activists (Snow, Zurcher, and Ekland-Olson 1980). More recent work has argued that such ties have an explicitly emotional component with significant power to advance or disrupt a movement’s life cycle (Berezin 2002; Goodwin 1997). Strong emotional ties to those outside a movement may limit the emotional energy persons have to devote to furthering social change. Equally, however, emotional bonds with others inside a movement may be an added source of strength, serving as an “opportunity structure” through which movements may gather the strength it needs to succeed (McAdam 1982). Frame alignment processes, long a tool of movement activists, represent an arena where the consideration of emotion has been particularly fruitful. The ability to align movement goals with the ideological predispositions of potential activists has long been recognized as a key part of movement mobilization (Groves 1995). However, in the past framing has been understood in as a predominantly cognitive process where actors come to see that their material interests or cultural values are best supported by involvement in social change. More recently, theorists have recognized that emotional resonance is also important for effective framing, particularly in helping to explain when frame alignment leads to social action and when it does not (Jasper 1998; Robnett 2004). For example, Berbrier (1998) demonstrated how white supremacy leaders used a frame where the dominant emotions were “love” and “pride” rather than “hate” in order to extend the appeal of their message to middle class whites. By tapping “fundamental sentiments” (Heise and Smith-Lovin 1981) viewed as appropriate to members of racial and ethnic groups, white supremacists were able to provide their movement with a degree of legitimacy that was not available through more traditional framing approaches that used cultural ideology or values (Berbrier 1998). Among the most active areas of research in social change and emotion has been emotion as a motivator of collective action and researchers have linked both positive and negative emotions to this effect (Goodwin, Jasper, and Polletta 2001; Reed 2004). Positive feelings such as pride in one’s racial, ethnic, or nationality group, for example, can lead to protest movement participation (Kim 2002). Also, movement involvement itself may produce positive feelings such as satisfaction or triumph that are motivating (Hercus 1999) and help to sustain social movements over time (Kemper 2001). When positive feelings are shared among movement participants, it produces bonds that improve group solidarity and can increase individuals’ commitments to activism and social change (Reed 2004). Negative emotions can also motivate social change; anger, moral outrage, fear, shame, and guilt are thought to be particularly potent sources of activist energy and consideration of these factors can add much to understandings of the motivations behind social change (Goodwin, Jasper, and Polletta 2001; Hercus 1999; Reed 2004). Barbalet (2002), for example, reexamined the motivations leading to the widespread adoption of secret ballots in the 19th century. Rationalist views of this innovation stressed its roots in widespread electoral corruption. Barbalet, in contrast, argued it was largely driven by aristocratic fears of the growing power of the working class. As the franchise was extended to workers, members of the ruling

CH17-P374095.indd 312

5/27/2008 7:26:14 PM

Commentary

313

class felt threatened by the group solidarity and social support that developed during open voting sessions. A secret ballot, Barbalet argued, ensured that each voter cast his ballot alone, minimizing the possibility that emotions raised during often boisterous open voting sessions would spill over into organized social action. Involvement in social movements or other forms of collective action also has the ability to transmute negative emotions, channeling them in ways that reduce unpleasant feelings or make them easier to cope with (Reid, Roberts, and Hilliard 1998). For example, feelings that persons may label as deviant or wrong can be transformed by activism into understandable or even normative responses to institutionalized arrangements (Thoits 1985, 1990). Hercus (1999) for example demonstrated how emotional framing by feminist social activists transmuted feelings of shame and fear into anger. Similarly, mothers participating in post-partum depression support groups altered their guilt, pain, and alienation by questioning dominant cultural views of motherhood and femininity (Taylor 1995). Such experiences not only may resolve emotional deviance but are likely to fuel dedication to on-going movement participation. Although the vast majority of literature on emotions and social change treats emotion as a motivating or “accelerating” force (Reed 2004), emotions also have the ability to derail social action. One way in which this can occur is through the pull of external ties. Network ties that link movement participants to outside obligations of any type can be destabilizing for activist organizations (Snow, Zurcher, and Ekland-Olson 1980). When these ties also involve strong emotional bonds (e.g., an intimate relationship rather than a demanding profession), the risk of movement disaffiliation may be particularly strong (Goodwin 1997). Of course, the affective tenor of bonds inside movement organizations also has an effect. For example, rapid membership increases may produce internal strains in an organization that create a negative emotional climate (Klatch 2004). This can reduce the degree of satisfaction activists derive from participation, potentially disrupting the movement itself. A related but less developed aspect of emotion in social change is the role it plays in movement nonparticipation. Involvement in social change exacts emotional costs that many potential movement actors may be unwilling to pay. In part, this is a function of anticipated drains on existing relationships. Participation in social change requires time and energy and may take participants away from valued relationships and not everyone is willing or able to make these sacrifices. A less studied but potentially important aspect of movement nonparticipation is the often unacknowledged emotional investments people have in maintaining the status quo (Calhoun 2001). The decision to become involved in a social movement often requires one to accept that something important has gone seriously awry. This awareness can produce powerful negative emotions and the desire to avoid such feelings may result in groups or individuals distancing themselves from a potentially troublesome social issue, engaging in “socially organized denial” (Norgaard 2006). In her study of responses to global warming by residents of a Norwegian town, Norgaard found “community members had sufficient information about the issue but avoided thinking about global warming

CH17-P374095.indd 313

5/27/2008 7:26:15 PM

314

Social Structure and Emotion

at least in part because doing so raised fears of ontological insecurity, emotions of helplessness and guilt, and was a threat to individual and collective senses of identity” (Norgaard 2006:391). Lack of interest or involvement in social issues is generally explained in purely rational terms, especially as a lack of information. By treating the issue from the perspective of the social management of emotions, movement scholars gain a useful tool for understanding the dynamics of both motivation for and resistance to social activism.

CHALLENGES AND NEW DIRECTIONS Now that the return of emotion to social change research is entering its second decade, leading scholars in the area have begun to call for more integrated theoretical treatment of the role of emotion in social change (Goodwin, Jasper, and Polletta 2001). This has highlighted several major questions in the field. For example, it is clear that emotion is implicated in both action and inaction with respect to social change; but what factors differentiate emotion that evokes social action from that which suppresses it? Under what conditions is emotion transmuted and redirected towards social rather than individual transformation? How does the context in which emotions are experienced influence whether or not they lead to organized social change? The paper by Linda Francis and Candace Berger, included in this section, considers some of these questions in the context of health disparities. They analyze the expression and transformation of powerful negative emotions among low income expectant mothers. Their analysis does not address social activism, per se, but rather its precursor in group discussions and dramatic story-telling. Francis and Berger portray women sharing with one another their experiences of inadequate pre-natal medical care and their sometimes explosive emotional reactions to this treatment. The mothers describe episode after episode where their own expectations conflicted with the approach of their caregivers. The women’s belief that aggressive, sometimes even invasive, procedures represented good pre-natal care contrasted sharply with the views of their physicians and nurses. Among medical staff, norms for care stress the importance of not alarming patients without cause and avoiding unnecessary treatment – approaches that left some women feeling ignored and discarded. This was exacerbated by the belief of many of Francis and Berger’s respondents that middle class women were routinely receiving the kinds of care that they had been denied. The situation described in Francis and Berger’s paper is even more fraught because rates of infant mortality and stillbirth are unusually high among the population involved in their study, particularly when compared with middle class mothers in the same geographical area. This suggests that social structural inequalities of some form are at work, exacting real world consequences in the form of infant health. Thus, the ethos developed by these women can be thought of as an effort to address very real disparities in health outcomes. At the time of their

CH17-P374095.indd 314

5/27/2008 7:26:15 PM

Commentary

315

interviews, none of the women were involved in movement activism directed towards changing the way they received care. However, their story-telling and shared outrage highlight emotion conditions that other scholars have noted can precede organized efforts at social change. Two other important questions related to theoretical development in emotions and social change are addressed in the paper by Jonathan Turner. One of the great challenges in generating a general theory of emotion in collective action or social movements is the wide variety of emotions that may be evoked at different points in a social change trajectory, in different contexts, and with different results. Turner’s paper, which is an effort to provide a general theory of social structure and emotions, attempts to do just that – to theorize how specific emotions may result from specific arrangements of social structure. Turner’s model also has the potential to contribute to theorizing in emotion and social movements because he differentiates between primary emotions (biologically and neurologically “hardwired” affective states such as happiness, fear, anger, and sorrow) and first, second, and third order emotions (ranked combinations of primary emotions; for example anger and sadness can combine to produce the feeling of betrayal). Such differentiations may have particular relevance for the role of emotion in social change because, unlike primary emotions, they are “historically and culturally variable” (Goodwin, Jasper, and Polletta 2001). Thus, while primary emotions are clearly implicated in the life cycle of social movements, first, second, and third order emotions allow for a more nuanced application of emotion as intertwined with culture, cognition, and ideology. Finally, a special challenge for theorizing in the area of emotions and social change is to grapple with the psychodynamic aspects of emotions, particularly as they relate to feelings about the self (Calhoun 2001). The Turner paper offers one approach by considering the role of emotional repression in the relationship between emotion, social structure, and social action. Turner argues that repression is important to a sociological treatment of emotion and social change because repressed emotion alters commitment to larger social structures. In part, he argues that this is because fundamental processes of emotional attribution tend to direct blame for negative emotion away from the self and toward mesoand macro-levels of social structure – offering a potential that emotion can be transmuted into action aimed at altering these structures. Thus, Turner offers a starting point that can be extended to encompass the questions that are of particular interest to theories of emotion and social change. The relationship between emotion and social change has a long and varied history. Scholars working in the area have moved through periods where they focused on emotion to the exclusion of reason and reason to the exclusion of emotion. Today, emotion is widely recognized as theoretically compatible with more rational representations of human behavior. It is also increasingly seen as an important mechanism for explaining the life cycle of social activism. Research in this area is developing rapidly and promises to add substantially to sociological knowledge concerning the relationship between emotion and social structure.

CH17-P374095.indd 315

5/27/2008 7:26:15 PM

316

Social Structure and Emotion

REFERENCES Aminzade, Ron and Doug McAdam. 2002. “Emotions and Contentious Politics.” Mobilization: An International Journal 7:107–109. Barbalet, J. M. 2002. “Secret Voting and Political Emotions.” Mobilization: An International Journal 7:129–140. Berbrier, Mitch. 1998. “Half the Battle: Cultural Resonance, Framing Processes, and Ethnic Affectations in Contemporary White Separatist Rhetoric.” Social Problems 45:431–450. Berezin, Mabel. 2002. “Secure States: Towards a Political Sociology of Emotion.” In: J. Barbalet (ed.), Emotions and Sociology, pp. 33–52. Oxford, UK: Blackwell. Beyerlein, Kraig and Michael P. Young. 2005. “Collective Emotions and Mobilizing CongregationBased Protest.” Presented at the Annual Meetings of the American Sociological Association. Philadelphia PA. Calhoun, Craig. 2001. “Putting Emotions in Their Place.” In: J. Goodwin, J. M. Jasper, and F. Polletta (eds), Passionate Politics: Emotions and Social Movements, pp. 45–57. Chicago: University of Chicago Press. Goodwin, Jeff. 1997. “The Libidinal Constitution of a High-Risk Social Movement: Affectual Ties and Solidarity in the Huk Rebellion, 1946–1954.” American Sociological Review 62:53–69. Goodwin, Jeff, James M. Jasper, and Francisca Polletta. 2000. “The Return of the Repressed: The Fall and Rise of Emotions in Social Movement Theory.” Mobilizations 5:65–83. Goodwin, Jeff, James M. Jasper and Francisca Polletta. 2001. Why Emotions Matter, In: Passionate Politics. Groves, Julian McAllister. 1995. “Learning to Feel: the Neglected Sociology of Social Movements.” The Sociological Review 43:435–461. Heise, David and Lynn Smith-Lovin. 1981. “Impressions of Goodness, Powerfulness, and Liveliness from Discerned Social Events.” Social Psychology Quarterly 44:93–106. Hercus, Cheryl. 1999. “Identity, Emotion, and Feminist Collective Action.” Gender and Society 13:34–55. Hochschild, Arlie Russell. 1983. The Managed Heart: Commercialization of Human Feeling. Berkeley: University of California Press. Jasper, James M. 1998. “The Emotions of Protest: Affective and Reactive Emotions In and Around Social Movements.” Sociological Forum 13:397–424. Kemper, Theodore. 2001. “A Structural Approach to Social Movement Emotions.” In: J. Goodwin, J. M. Jasper, and F. Polletta (eds), Passionate Politics: Emotions and Social Movements, pp. 58–73. Chicago: University of Chicago Press. Kim, Hyojoung. 2002. “Shame, Anger, and Love in Collective Action: Emotional Consequences of Suicide Protest in South Korea, 1991.” Mobilization: An International Journal 7:159–176. Klatch, Rebecca E. 2004. “The Underside of Social Movements: The Effects of Destructive Affective Ties.” Qualitative Sociology 27:487–509. Massey, Douglas S. 2002. “A Brief History of Human Society: The Origin and Role of Emotion in Social Life: 2001 Presidential Address.” American Sociological Review 67:1–29. McAdam, Doug. 1982. Political Process and the Development of Black Insurgency. Chicago: University of Chicago Press. Norgaard, Kari Marie. 2006. “People Want to Protect Themselves a Little Bit: Emotions, Denial, and Social Movement Nonparticipation.” Sociological Inquiry 76:372–396. Perry, Elizabeth J. 2002. “Moving the Masses: Emotion Work in the Chinese Revolution.” Mobilization: An International Journal 7:111–128. Jean-Pierre Reed. 2004. “Emotions in Context: Revolutionary Accelerators, Hope, Moral Outrage, and Other Emotions in the Making of Nicaragua’s Revolution.” Theory and Society 33:653–703. Reid, Lesley Williams, J. Timmons Roberts, and Heather Monro Hilliard. 1998. “Fear of Crime and Collective Action: An Analysis of Coping Strategies.” Sociological Inquiry 68:312–328. Robnett, Belinda. 2004. “Emotional Resonance, Social Location, and Strategic Framing.” Sociological Focus 37:195–212.

CH17-P374095.indd 316

5/27/2008 7:26:15 PM

Commentary

317

Snow, David A., Louis A. Zurcher, and Sheldon Ekland-Olson. 1980. “Social Networks and Social Movements: A Microstructural Approach to Differential Recruitment.” American Sociological Review 45:787–801. Taylor, Verta. 1995. “Self-Labeling and Women’s Mental Health: Postpartum Illness and the Reconstruction of Motherhood.” Sociological Focus 28:23–47. Thoits, Peggy. 1985. “Self Labeling and Mental Illness: The Role of Emotional Deviance.” Thoits, Peggy A. 1990. “Emotional Deviance: Research Agendas.” In: T. Kemper (ed.), Research Agendas in the Sociology of Emotions, SUNY Series in the Sociology of Emotions. Albany, NY: State University of New York Press. Williams, Simon J. 2000. “Is Rational Choice Theory Uunreasonable?” In: M. Archer and J. Tritter (eds), Rational Choice Theory: Resisting Colonization, pp. 57–72. New York: Routledge.

CH17-P374095.indd 317

5/27/2008 7:26:15 PM

18 Emotions and Social Structure: Toward A General Sociological Theory Jonathan H. Turner University of California, Riverside, CA, USA

ABSTRACT Social structures and their attendant cultures are sustained or changed by people’s emotional reactions to their circumstances. Indeed, it is not too much to argue that emotions are the “glue” that holds societies together, or the dynamite that blows them apart – to phrase the issue metaphorically. Surprisingly, during sociology’s first 150 years, comparatively little was said about human emotions and social structure. Since the 1970s, however, new theories and research programs have emerged; and some of these programs have addressed the linkage between social structure and emotions but typically within a rather narrow theoretical tradition (see Kemper 1990; Stets and Turner 2006; Turner and Stets 2005; for reviews). In this paper, my goal is to present a more general theory of emotions and social structure which, I hope, incorporates the useful elements from existing theories and cumulative research findings. CONCEPTUALIZING SOCIAL STRUCTURE AND EMOTIONS A SIMPLE CONCEPTUAL SCHEME ON LEVELS OF SOCIAL REALITY

In general terms, social structures are patterns of social relations among individual and collective actors that endure for a time. In recent years, I have argued Social Structure and Emotion

CH18-P374095.indd 319

319

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 4:56:16 PM

320

Social Structure and Emotion

for a view of social structure as naturally unfolding at micro, meso, and macro levels (Turner 2000, 2002, 2003, 2005a,b, 2006, 2007a). For each level of social reality, there are generic types of structures unique to this level; and moreover, there are fundamental forces driving the formation and change of these structures, either directly or indirectly. I visualize forces as dynamic properties of the social universe that push individual and collective actors to act in ways that produce, reproduce, or change social structures. At the macro level, forces drive the formation of institutional domains (e.g., economy, politic, religion, kinship, etc.) and stratification systems (ranked classes and strata), whole societies, and systems of societies; at the micro level, forces create focused and unfocused encounters or episodes of mutual awareness and, in the case of focused encounters, face-to-face interaction (Goffman 1961, 1963). In response to these macro and macro forces, meso-level structures are generated by actors. These mesolevel structures are of two basic types: corporate units revealing a division of labor in pursuit of goals (e.g., groups, organizations, and communities) and categoric units (or diffuse status characteristics) consisting of distinctions that classify individuals (e.g., age, sex, ethnicity, class). Figure 18.1 sketches this view of social structure. Generic types of structures at each level of reality are driven by the micro and macro forces enumerated on the left of the figure, with these two sets of forces defined in Table 18.1. Levels of social reality are embedded within each other, from both a top down and bottom up direction. From a top down perspective, macro-level structures constrain the formation of corporate and categoric units which, in turn, impose parameters upon interaction in encounters. From a bottom up view, meso-level structures are constructed from chains of interaction in encounters, while institutional domains and stratification systems are built, respectively, from corporate units and categoric units (or diffuse status characteristics). In turn, societies and systems of societies are built from institutional and stratification systems. Since emotions are a micro-level force, the task of building a theory is to see how meso and macro structures affect the arousal of emotions at the level of the encounter, and reciprocally, how such arousal has effects on corporate and categoric units and, by extension, macro-level social structures. Emotions are aroused at the micro level of reality in encounters, but these encounters are almost always embedded in corporate and categoric units that, in turn, are lodged, respectively, in institutional domains and stratification systems (and, by extension, whole societies and, potentially, systems of societies). Emotions flow from micro-level face-to-face encounters to meso and macro levels of reality; and in so doing, they can create new patterns of social organization, sustain existing patterns, or provide energy for structural transformation. Conversely, emotions are aroused under varying types of social structural arrangements, particularly those at the meso level of corporate and categoric units that provide the immediate structural context for virtually all encounters.

CH18-P374095.indd 320

5/26/2008 4:56:16 PM

321

Emotions and Social Structure: Sociological Theory Inter-societal system

Society

Macro-level structure Population Production Distribution Regulation Reproduction

Meso-level structure

Institutional domains

Stratification system

Corporate units

Categoric units

Micro-level structure Emotions Transactional needs Symbols Roles Status Demography Ecology

Encounters of face-to-face interaction

Individual behaviors, cognitions, and emotions

FIGURE 18.1

Levels of social reality.

A CONCEPTUALIZATION OF VARIATIONS IN HUMAN EMOTIONS

There are several very useful classifications of emotions (e.g., Kemper 1987; Plutchik 1980; Thamm 2006). Table 18.2 presents my views on variations in human emotions (Turner 1999, 2000, 2002, 2007a). As is evident, I see four emotions as primary – that is, emotions that are hard-wired in human neuroanatomy – with their

CH18-P374095.indd 321

5/26/2008 4:56:16 PM

322 TABLE 18.1

Social Structure and Emotion Forces of the Social Universe

Macro-level forces 1

Population

The absolute number, rate of growth, composition, and distribution of people.

2 Production

The gathering of resources from the environment, the conversion of resources into commodities, and the creation of services to facilitate gathering and conversion.

3

Distribution

The construction of infrastructures to move resources, information, and people in space as well as the use of exchange systems to distribute resources, information, and people.

4

Regulation

The consolidation and centralization of power along its four bases (coercion, administrative structures, manipulation of material incentives, and symbols) in order to control and coordinate members of a population.

5

Reproduction

The procreation of new members of a population and the transmission of culture to these members as well as the creation and maintenance of socio-cultural systems that sustain life and social order.

Micro-level forces 1

Emotions

The arousal of variants and combinations of fear, anger, sadness, and happiness.

2

Transactional needs

The activation of needs for confirmation of self, positive exchange payoffs, trust and predictability, facticity or the sense that things are as they appear, and group inclusion.

3

Symbols

The production of expectations (normatization) with respect to categories of people present, nature of the situation, forms of communication, frames of what is included and excluded, rituals, and feelings.

4

Roles

The presentation of sequences of gestures to mark a predictable course of action (role making) and the reading of gestures to understand the course of action of others (role taking).

5

Status

The placement and evaluation of individuals in positions vis-à-vis other | positions and creation of expectation states for how individuals in diverse and differentially evaluated positions should behave.

6

Demographic

The number of people co-present, their density, and their movements, as well as the meanings assigned to number, density, and movements of individuals.

7

Ecological

The boundaries, partitions, and props of space as well as the associated meeting of boundaries, partitions, and props.

elaboration into low-, medium-, and high-intensity states. From these primary emotions are built what I term first-order elaborations consisting of a dominant amount of one primary emotion and a lesser amount of another primary emotion. A secondorder emotion is a combination of three primary emotions rank ordered by their relative dominance. I suspect that some first- and second-order emotions are also hard-wired, but I will leave this question open since it falls outside of specific questions that I am addressing in this paper (see Turner 2000, for my analysis of the neurological bases for human emotions). I should also emphasize that the notion

CH18-P374095.indd 322

5/26/2008 4:56:16 PM

Emotions and Social Structure: Sociological Theory

TABLE 18.2

323

Variant and Elaborations of Primary Emotions Variations Low intensity

Moderate intensity

High intensity

Satisfaction– happiness

Content Sanguine Serenity Gratified

Cheerful Buoyant Friendly Amiable Enjoyment

Joy Bliss Rapture Jubilant Gaiety Elation Delight Thrilled Exhilarated

⫹Fear: wonder, hopeful, gratitude, prided ⫹Anger: vengeance, appeased, calmed soothed, relish, triumphant, bemused ⫹Sadness: nostalgia, yearning, hopefulness

Aversion–fear

Concern Hesitant Reluctance Shyness

Misgivings Trepidation Anxiety Scared Alarmed Unnerved Panic

Terror Horror High anxiety

⫹Happiness: awe, reverence, veneration ⫹Anger: revulsed, repulsed, dislike, envy, antagonism ⫹Sadness: dread, wariness

Assertion–anger

Annoyed Agitated Irritated Vexed Perturbed Nettled Rankled Piqued

Displeased Frustrated Belligerent Contentious Hostility Ire Animosity Offended Consternation

Dislike Loathing Disgust Hate Despise Detest Hatred Seething Wrath Furious Inflamed Incensed Outrage

⫹Happiness: snubbing, mollified, rudeness, placated, righteousness ⫹Fear: abhorrence, jealousy, suspicion ⫹Sadness: Bitterness, depression, betrayed

Sorrow Heartsick Despondent Anguished Crestfallen

⫹Happiness: acceptance, ⫹Fear, anger: moroseness, solace, guilt melancholy ⫹Fear: Forlornness, ⫹Anger, fear: remorsefulness, shame misery ⫹Anger, fear: ⫹Anger: aggrieved, alienation discontent, dissatisfied, unfulfilled, boredom, grief, envy, sullenness

Emotion

Disappointment– Discouraged Dismayed sadness Downcast Disheartened Dispirited Glum Resigned Gloomy Woeful Pained Dejected

CH18-P374095.indd 323

First-order emotions

Second-order emotions

5/26/2008 4:56:16 PM

324

Social Structure and Emotion

TABLE 18.3

The Structure of Second-order Emotions: Shame, Guilt, and Alienation Rank ordering of constituent primary emotions

Emotion

1

2

3

Shame

Disappointment–sadness (at self)

Assertion–anger (at self)

Aversion–fear (at consequences for self)

Guilt

Disappointment–sadness (at self)

Aversion–fear (at consequences for self)

Assertion–anger (at self)

Alienation

Disappointment–sadness Assertion–anger (at others, (at self, others, situation) situation)

Aversion–fear (at consequences for self)

of elaborations or “mixes” of primary emotions is metaphorical; at this point, it is not clear just how these more complex emotions are generated, although data from imaging studies on the areas of the brain responsible for these first- and secondorder emotions are increasingly available. Thus, primary emotions (column 1 in Table 18.2) vary along a continuum ranging from low-intensity through medium-intensity to high-intensity states (columns 2, 3, and 4 of Table 18.2). First- and second-order emotions can also vary in intensity, depending upon the level of intensity of the primary emotions from which they are built. First-order elaborations of emotions are listed in column 5 in Table 18.2. As noted above, these emotions are built from a combination, if this is the right metaphor, of a greater amount of happiness, fear, anger, or sadness and a lesser amount of one other primary emotion. And the only second-order emotions listed – shame, guilt, and alienation – are built from sadness, as the dominant emotion, combined with lesser amounts of anger and fear (with the rank ordering, relative intensity of fear and anger, and the targets of anger fear, and sadness determining whether an individual experiences shame, guilt, or alienation). Table 18.3 outlines the structure of shame, guilt, and alienation. Once emotions are aroused in encounters, they are directed at a variety of potential objects. At the level of persons, emotions can be directed at two targets: self and other(s). At the level of social structures, there are only seven potential targets: encounters, categoric units, corporate units, institutional domains, stratification systems, whole societies, and inter-societal systems. Thus, a sociological theory of emotions must explain the dynamics of emotional arousal and of targeting of emotional energies on self, others, or social structures.

EMOTIONAL DYNAMICS THE PRINCIPLES OF POSITIVE AND NEGATIVE EMOTIONAL AROUSAL

Two basic conditions cause emotional arousal: (1) expectations and (2) sanctions. In virtually all situations, individuals have expectations about what should

CH18-P374095.indd 324

5/26/2008 4:56:16 PM

Emotions and Social Structure: Sociological Theory

TABLE 18.4 1

325

Emotional Arousal, Defense Mechanisms, and Attributions

When individuals receive positive sanctions and/or meet expectations for what will transpire in an encounter, especially expectations generated by the operation of microdynamic forces (see Table 18.1), they will experience positive emotional arousal.

2 When individual receive negative sanctions and/or fail to meet expectations for what would transpire in an encounter, especially expectations generated by the operation of microdynamic forces (see Table 18.1), they will experience negative emotional arousal. 3 When individuals experience negative emotional arousal, they become more likely to protect self through the activation of defensive mechanisms that push the negative emotions below the level of consciousness; and the more intense are the negative emotions repressed and the more salient is self in a situation where negative emotions have been aroused, the more likely will repression lead to the intensification of the repressed emotions and to their transmutation into new variants of negative emotional arousal. 4

When individuals experience either positive or negative emotional arousal, they will make attributions about the causes of their emotional experiences to one or more of the following objects: self, other(s), structure of the encounter, structure and culture of corporate unit, members of categoric units, institutional domain, stratification system, whole society, or system of societies.

and will occur. When expectations are realized or exceeded and when positive sanctions are experienced, individuals will experience a range of positive emotions along the satisfaction–happiness continuum. Conversely, when expectations are not met and when negative sanctions are received, individuals will feel negative emotions which can range along (a) the assertion–anger, aversion–fear, disappointment–sadness continua for primary emotions, (b) first-order combinations of these three negative primary emotions, or (c) second-order combinations of anger, fear, and sadness in the form of shame, guilt, or alienation. Propositions 1 and 2 in Table 18.4 summarize in simple form these two basic laws. Though these laws seem obvious, they are fundamental to understanding the relationship between emotions and social structure. Structures and their cultures set up expectations and provide the ideological and normative yardsticks by which performances are evaluated and sanctions meted out; and the emotions aroused from meeting or failing to meet expectations and from sanctions, whether positive or negative, will affect not only people’s behaviors but also the emotions that they direct at social structures. These dynamics are, however, complicated by other psychodynamic processes, particularly the activation of defense mechanisms to protect self from the negative emotions aroused when individuals perceive that they have failed to meet expectations or that they have received negative sanctions from others. THE ACTIVATION OF DEFENSE MECHANISMS

With but a few exceptions (e.g., Scheff 1988), sociologists have not adequately conceptualized an important dynamic: repression of negative emotions.

CH18-P374095.indd 325

5/26/2008 4:56:16 PM

326

Social Structure and Emotion

When individuals fail to meet expectations – whether their own or those of others – and when they experience negative sanctions, self becomes more salient. Individuals are suddenly more self-aware; and once self is on the line, individuals will often employ a variety of defensive strategies and defense mechanisms to protect self, particularly from a second-order negative emotion like shame that makes self feel small and inadequate (Tangney and Dearing 2004). Repression (and denial) are, I believe, master defense mechanism because they push negative emotions about self below the level of consciousness; and as they do so, other defense mechanisms – such as displacement, projection, reaction formation, sublimation, and most importantly, attribution – can be activated (see Turner 2007a, for a more complete analysis of defense mechanisms; for the present, I will focus on repression and attribution which I see as the most important defense mechanism from a sociological perspective). Once emotions are repressed, they are transmuted into new kinds of emotions, in several ways. First, repressed emotions will generally increase in intensity, thus requiring ever-more cortical control and repression in a cycle that can lead to severe behavioral pathology. Repressed anger, for example, will often move to high-intensity forms of anger that periodically break through the cortical sensors as rage and hatred which, in turn, may invoke feelings of shame and guilt that also must be repressed, only to emerge periodically as emotional spikes of intense shame and guilt. Repressed sadness will often lead to severe depression which individuals may repress through displays of happiness, spiked by sudden arousal of intense ennui and despair. Repressed fear will often reemerge in anxiety attacks and even more intense fear responses such as terror. Second, repressed emotions will often be transmuted into new kinds of emotions. Such is particularly likely to be the case when shame is repressed; it will be transmuted into forms of anger that will periodically break through the defensive regime of persons, causing them to experience more shame and perhaps guilt in a cycle that escalates the intensity of the repressed shame. Alienation may be transmuted into any one of its component emotions – sadness, anger, and fear. Thus, almost any negative emotion which is repressed will also be transmuted into a new emotion. For example, as Freud revealed in his studies of hysteria, repressed guilt can become intense fear and anxiety (often expressed behaviorally as phobias); anger at father can be transmuted into guilt and even shame. Repressed fear is often converted into anger at objects perceived to cause fear or displaced upon safer objects that cannot fight back. Repressed anger often is transmuted into sadness, or more complex emotions like shame in which sadness is the dominant emotion. These kinds of dynamics are summarized, at an abstract level, in Proposition (3) in Table 18.4. ATTRIBUTION AS A KEY DYNAMIC

Edward Lawler (2001) has argued that negative emotions have a “distal bias” in that individuals will generally make external attributions to others and social structures as the cause of negative emotional arousal. In contrast, positive emotions

CH18-P374095.indd 326

5/26/2008 4:56:16 PM

Emotions and Social Structure: Sociological Theory

327

have a “proximal bias,” with individuals making self-attributions for positive emotional arousal. This insight suggests that attribution is a key defense mechanism. Attribution is, of course, basic to cognitive functioning because individuals are always assessing the causes of their own and the actions of others. But, attribution also operates as a defense mechanism because, as people repress negative emotions, they almost always begin to make assessments of the cause for the intensified and transmuted emotions that they do experience consciously. Both Lawler and I use the concept of attribution in a slightly different way than its original Gestalt formulation (see Weiner 2006, for a more recent application in psychology). For us, internal attribution refers not so much to the internal states of others but to self-attributions in which a person sees his or her behavior as the cause of some outcome. Individuals will typically see self as responsible for meeting expectations and for receipt of positive sanctions – as explicated in Lawler’s notion of a proximal bias. Self-attributions for positive emotions also set into motion the flow of positive emotions in encounters. Individuals are likely to give off positive sanctions to others when they experience positive emotion, causing others to give back positive sanctions to others that enhance the flow of positive sentiments in encounters. Positive emotions arising from self-attributions can thus begin to circulate in local networks. In contrast, negative emotions are much more likely to be pushed outward, away from the local encounter as individuals repress negative feelings toward self and make external attributions for their intensified and transmuted feelings. External attribution operates like any other defense mechanism, with individuals “blaming” others, categories of others, the structure and culture of corporate units, institutional system, stratification system, whole society, or system of societies for their negative emotional feelings. This elementary principle on these attribution dynamics is summarized in Proposition (4) in Table 18.4. As will become evident, I will elaborate on this principle of attribution because I see it as a critical link between emotions and social structure.

PSYCHODYNAMICS I am using the term, psychodynamics, to denote the processes that connect emotions to social structures. These processes revolve around (1) emotional arousal within encounters embedded in meso-level structures (and, by extension, macro-level structures), (2) attributions assigning causality about the sources of emotional arousal, (3) activation of defense mechanisms which inevitably alter the valences of emotional arousal and the targets of attribution dynamics. POSITIVE EMOTIONS AND SOCIAL STRUCTURE

When individuals meet expectations and receive positive sanctions from others, they will experience variants of satisfaction–happiness, and the more self-conceptions are on the line, the more intense will this happiness be.

CH18-P374095.indd 327

5/26/2008 4:56:17 PM

328

Social Structure and Emotion

And, if they had some fear about their ability to meet expectations or receive positive sanctions, this fear mixed with happiness generates pride (see proposition (5) in Table 18.5). Because positive emotions have a proximal bias, individuals are likely to make self-attributions, seeing their own actions as the cause of positive emotional arousal. And, the higher the self-esteem of a person, the more likely is this person to make self-attributions for positive feelings (proposition (6) in Table 18.5). Moreover, they are less likely to activate defense mechanisms because there is no reason to protect self, although very low self-esteem persons might have a tendency to avoid self-attributions as they try to protect a negative view of self from disconfirming evidence. If individuals make external attributions under conditions of positive emotional arousal, it will generally be to others who are perceived to have helped them meet expectations or to receive positive sanctions, and under these conditions, these individuals will experience and express gratitude to these others; and the more they had some fear about meeting expectations or receiving positive sanctions, they will experience more intense forms of gratitude and perhaps relief (proposition (7)). Positive emotional arousal, as scholars from Durkheim (1912) to Collins (2004) have emphasized, generates positive emotional energy that has an “effervescent” character; those receiving positive sanctions or meeting expectations are likely to offer positive sanctions to others who will reciprocate with positive sanctions of their own. And, out of this ratcheting up of the positive emotional flow, individuals will become emotionally entrained and will develop commitments to each other and the local encounter (Collins 2004), and potentially, to the meso structure in which the encounter is embedded [proposition (8)]. They may also make external attributions to the encounter as structured in ways that facilitated their success, and as a result, they will develop commitments to the encounter and those who participate in this encounter, now and into the future (proposition (9)). Indeed, they may develop particularized cultural capital and mark the encounter with symbols or totems that have the power to activate positive emotions (Collins 2004). These positive emotional dynamics revolving around attributions to others and the encounter are most likely to occur under particular structural conditions (listed in proposition (9-A to 9-E) in Table 18.6): the density of networks among participants in the encounter is high, the need for close coordination of activities in the encounter is high (or, productive exchange in Lawler’s, 2001, terms), the participants stand at equivalent places vis-à-vis the authority structure of the corporate unit in which the encounter is embedded, the participants to the encounter are members of the same categoric unit, and the positive symbols arising from emotional entrainment are juxtaposed against negative symbols denoting outsiders. There is a tendency for positive emotional energy to stay local – circulating among individuals in encounters. This fact presents significant problems for meso and macro structures since their viability depends upon development of emotional attachments and assignment of legitimacy to the structure and

CH18-P374095.indd 328

5/26/2008 4:56:17 PM

329

Emotions and Social Structure: Sociological Theory

TABLE 18.5

Positive Emotional Arousal and Social Structure

5

The more self is salient in an encounter, the more intense will be the variants of satisfaction–happiness when receiving positive sanctions from others and/or meeting expectations; and if a person had some fear about receiving positive sanctions and/or meeting expectations when self was salient, this person will experience pride in having done so.

6

The more positive emotions are aroused in encounters, the more likely are individuals to make self-attributions for their success, and particularly so when this person possesses high levels of self-esteem.

7

If others are seen by an individual to have facilitated the receipt of positive sanctions and/or realization of expectations, this person will also experience and express gratitude to these others; and if an individual had fear about receiving positive sanctions and/or meeting expectations, the person will experience more intense forms of gratitude and relief, while giving off more effusive positive emotions to these others.

8

The more individuals experience positive emotional arousal in a situation, even under conditions where self-attributions are made for this arousal, the more likely are these persons to give off positive sanctions to others which are more likely to be reciprocated in a cycle of emotional entrainment that increases commitments to others, the structure of the encounter, and potentially, the meso and macro units in which the encounter is embedded.

9

If the structure and culture of the encounter are seen by an individual to have facilitated receipt of positive sanctions and/or realization of expectations, this person will give off positive emotions to others in the encounter and develop stronger commitments to the encounter, now and in the future; and the more a person is committed to an encounter and the more this encounter is iterated, the more likely is particularized cultural capital to emerge among its members, and the more likely is the encounter to develop its own symbols and totems. These outcomes are more likely when: A The density of network ties of participants is high. B The participants stand in the same or structural equivalent positions vis-à-vis the authority structure of the corporate unit in which the encounter is embedded. C The encounter involves close coordination of activities among its participants. D The participants to the encounter are members of the same categoric unit. E The symbols and totems of the encounter that arise from the emotional entrainment of its participants are juxtaposed against negative symbols denoting outsiders and, in extreme cases, enemies.

10

Positive emotions generated in encounters will tend to stay local and circulate among members of iterated encounters under the conditions listed above, but this circulation of positive emotions and symbols leads to external attributions to meso-level structures when: A The corporate unit within which encounters are embedded is: – small, allowing for high rates of interaction among its members, – perceived as facilitating members’ activities, – structured horizontally with low levels of inequality and hierarchy. B The categoric units in which encounters are embedded are positively evaluated.

11

Positive emotions generated in encounters will be transformed positive sentiments giving both legitimacy and commitments to macro-level structures when individuals have consistently been able to receive positive sanctions and meet expectations across a wide range of corporate units within institutional domains and across diverse categoric units within the stratification system. (Continues)

CH18-P374095.indd 329

5/26/2008 4:56:17 PM

330 TABLE 18.5

Social Structure and Emotion (Continued)

12 If individuals make self-attributions for the arousal of negative emotionsas a result of sanctions or failure to meet expectations, they are likely to feel variants of disappointment–sadness; and, if failure to receive positive sanctions or meet expectations is evaluated in moral terms, they will also experience guilt. And the more individuals experience guilt, the more likely are they to engage in repair rituals that restore social bonds leading to positive sanctioning from others which, in turn, increases the flow of positive emotions and the commitment to others, the encounter, and potentially, to meso and macro structures in which the encounter is embedded. 13

Positive emotional arousal in encounters will create baseline expectations for subsequent interactions that are often converted into rising expectations above this baseline. If expectations are realized in subsequent interactions of an encounter, the flow of positive emotional energy will be sustained, and individuals will maintain commitments to the encounter and the meso structures in which this iterated encounter is embedded, while at the same time becoming more likely to bestow legitimacy on meso and macro structures.

TABLE 18.6

Negative Emotional Arousal and Social Structure

14

The more baseline expectations are converted into rising expectations, the more likely will expectations go unrealized in subsequent iterations of an encounter, with the result that the level of positive emotional energy will decline and be supplemented with variants of disappointment–sadness and assertion–anger; and if these negative emotions do not lead to a downward adjustment of expectations and, thereby, persist across iterated encounters, commitments to meso and macro structures in which an encounter is embedded will decline.

15

If defense mechanisms are not activated as negative emotions are aroused in encounters, and individuals make accurate internal or self-attributions for their failure to meet expectations and/or receipt of negative sanctions, the negative emotions aroused will generally lower commitments to social structures through a variety of routes: A The more variants of disappointment–sadness are aroused with self-attributions, the lower will be the modal level of emotional energy of individuals and the less likely are they to develop strong commitments to others in encounters or the structures in which the encounter is embedded. B The more salient is self under conditions of self-attributions, the more likely are individuals to experience shame, but even if this shame does not lead to repression, it will still generate the three negative emotions that will likely activate alienation and, thereby, lower commitments to others and social structures. C The more salient are moral codes under conditions of self-attributions, the more likely are individuals to experience guilt which will increase corrective behaviors (see proposition (14)) but which, if unsuccessful, will generate the three negative emotions that will likely activate alienation and, thereby, lower commitments to others and social structures.

16 If defense mechanisms are not activated as negative emotions are aroused in encounters, and if individuals make accurate external attributions for their failure to meet expectations and/or receipt of negative sanctions, these individuals are more likely to experience and, if possible, express anger toward, and/or fear of, others; and if these emotions consistently breach the flow of interaction in encounters, individuals’ commitments to the encounter and structures within which the encounter is embedded will decline, and especially so if sadness, anger, and fear combine to generate alienation from others and social structures. (Continues)

CH18-P374095.indd 330

5/26/2008 4:56:17 PM

Emotions and Social Structure: Sociological Theory

TABLE 18.6

331

(Continued)

17 If defense mechanisms are activated as negative emotions are aroused in encounters, these emotions will increase in intensity and, if not transmuted, will periodically erupt in episodes of intense fear (high anxiety, terror), anger (rage), or sadness (enuii, depression) that disrupt and breach encounters and activate shame; and if shame is intense (e.g., humiliation) and is repressed, it will be transmuted into variants of anger and first-order combinations of anger and happiness (righteousness and revenge) directed at others and social structures. 18

If shame is repressed over a long period of time and habitually, it will not only surface in episodes of intense anger at external targets, but individuals will also experience episodes of the two other emotions that create shame, sadness and fear, that will alienate them from others and social structures. A If corporate units are the targets of repressed shame, individuals will become angry at, and alienated from, these structures. B If categoric units are the targets of repressed shame, individuals will experience diffuse anger at, develop prejudices about, and commit discriminatory acts against members of these units. C If intense episodes of anger arising from transmuted shame breach encounters where anger is expressed, individuals are more likely to become trapped in shame–anger–shame cycles that ratchet up the intensity of negative emotions and the anger directed at ever-more mesoand macro-level targets.

19 The more biography of individuals has involved the repression shame in corporate units within those institutional domains distributing valued resources (e.g., economy, polity, education), the more distal will be the targets of anger and combinations of anger and happiness (righteousness and revenge); and the more likely are individuals to commit aggressive acts against meso and macro structures, especially when: A The targets of these external attributions can be symbolized in negative terms. B The local networks in which anger and first-order combinations of anger at external targets are expressed provide positive sanctions for expressions of anger and its variants. C The local networks in which anger and first-order combinations of anger at external targets provide a sense of realizing goals and expectations through the thinking, planning, and expression of anger, per se. D The positive emotions and symbols generated by interaction rituals in local networks are juxtaposed to the negative symbols of the targets of anger. 20

The external attributions for diffuse and righteous anger will be more likely to be directed at macro structures of a society, a society as a whole, or system of societies when: A The connection between meso structures originally generating shame and the anger that arises from transmuted shame has been severed through activation of defense mechanisms, or has been manipulated by actors with an interest in deflecting anger of individuals outward. B The corporate units that organize local networks have material, organizational, and symbolic resources to channel diffuse and righteous anger at external targets; and the greater the resource base of these corporate units, the more distal will collective external attributions become.

culture of these larger units. There are, I believe, several structural conditions that increase the likelihood that individuals will make external attributions for positive emotions to meso-level structural units (propositions (10-A) and (10-B)): the corporate unit in which an encounter is embedded is small, the hierarchy of authority is limited and the level of inequality is low, the structure of the

CH18-P374095.indd 331

5/26/2008 4:56:17 PM

332

Social Structure and Emotion

corporate unit is seen as enabling as opposed to restrictive, success at the level of the encounter is associated with positively evaluated categoric units tied to the division of labor and authority structure of the corporate unit. If positive sentiments toward corporate and categoric units are intense, then there will be some spill over to the larger institutional domains and stratification systems in which, respectively, corporate and categoric units are embedded. Thus, individuals who are happy with their job in a series of corporate units are likely to feel commitments and give legitimacy to the economy; or persons who have experienced positive emotions in iterated interactions with members of a categoric unit will generally give legitimacy to the place of this categoric unit in the larger stratification system of a society (proposition (11)). Positive emotions will be extended to macro structures under predicable conditions. As noted above, one is consistent positive emotional arousal in encounters within corporate and categoric units. Another important condition is consistently positive emotional arousal across an array of encounters embedded in diverse corporate and categoric units (proposition (11)). Under these conditions, individuals will be likely to give legitimacy to, and develop commitments to, all institutional domains, the larger stratification system, and the society as a whole. The key is for individuals to have the capacity to meet expectations and receive positive sanctions in encounters within corporate units within diverse institutional domains (e.g., family, polity, economy, religion, and education) and categoric units within these domains. Individuals will, under these conditions, give off diffuse legitimacy to macro structures and will continue to do so as long as they perceive that these structures provide them with options to meet expectations and receive positive sanctions. For example, those with patriotic fervor are usually those who have experienced positive emotions across diverse corporate and categoric units, whereas those who have not experienced positive emotions consistently across corporate or categoric units are far less likely to give legitimacy to macro structures because of the negative emotions aroused, the activation of defense mechanisms to protect self, and the negative causal attributions made to both meso and macro structures. Of course, persons do not always employ defense mechanisms, and sometimes they blame self for failures to meet expectations and for receipt of negative sanctions. If such is the case, then it becomes possible for a person to engage in repair rituals that re-establish social bonds which, in turn, can increase their ability to meet expectations and receive positive sanctions (Scheff 1979). Some negative emotions are more likely than others to individuals to initiate repair rituals in breached encounters. Guilt, for example, is an associative motion because individuals typically do not blame their global self-conception for acts that violate moral codes; rather, they see specific behaviors as inappropriate, while generally leaving intact the sanctity of self (Tagney and Dearing 2004). As a result, defense mechanisms are less likely to be invoked, with individuals engaging in corrective behaviors and repair rituals to restore associative relations with others in an encounter. In so doing, commitments to the culture and structure of

CH18-P374095.indd 332

5/26/2008 4:56:17 PM

Emotions and Social Structure: Sociological Theory

333

encounters and the meso structures in which they are embedded is maintained (proposition (12) in Table 18.5. There is also a dialectical element in positive emotional arousal at any level of social structure: rising expectations. When individuals receive positive sanctions and meet expectations at one point in time, there is a tendency of these to become a baseline of expectations for subsequent iterations of the encounter. If subsequent interactions meet these baseline expectations, the flow of positive emotional energy will be sustained, and individuals will maintain commitments to the encounter and, in all likelihood, to the meso and macro structures within which the encounter is embedded (proposition (13)). There is also a tendency for baseline expectations to be transformed into rising expectations, with individuals implicitly making the assumption that one level of positive emotional arousal will be followed in subsequent interactions by even more positive emotional arousal. As is enumerated in proposition (14) in Table 18.6, what makes these dynamics dialectical is that baseline or rising expectations cannot always be realized which arouses negative emotions such as disappointment or anger that subtract from the positive emotions that are aroused. Thus, an encounter that has been particularly gratifying sets up expectations that will be at least as gratifying in the future and, potentially, even more gratifying. When iterations of the encounter do not meet either baseline or rising expectations, negative emotions will intrude into the positive emotional entrainment, although if these negative emotions lead individuals to lower expectations, a realistic equilibrium point in expectations may be realized. The dilemma faced by all meso and macro structures is how to pull positive emotional energy from encounters so that these sentiments are transferred to larger-scale social structures. As emphasized, there is a proximal bias for the flow of positive emotions, and it is difficult to overcome this bias because positive emotions are aroused most readily through actual face-to-face contact. As Durkheim (1912) argued long ago, the proximal bias of positive emotions can only be broken when the larger structures in which encounters are embedded can be symbolized in some manner so that rituals directed at these symbols can redirect the flow of positive emotions from the local encounter to larger-scale structures. But, even if positive emotions are directed at meso and macro structures, it is often a difficult balance to keep this emotional energy so directed, not only because of the inherent proximal bias of positive emotional energy but also because of the distal bias of negative emotions that, when aroused, will be directed outward toward meso and macro structures (see propositions in Table 18.6). And, coupled with the dialectic of rising expectations, it should not be surprising that commitments to meso and macro structures are difficult to sustain. NEGATIVE EMOTIONS AND SOCIAL STRUCTURE

Negative emotional arousal reveals a distal bias, with individuals externalizing their feelings onto others or social structures. The dialectic of rising expectations

CH18-P374095.indd 333

5/26/2008 4:56:17 PM

334

Social Structure and Emotion

assures that, at some point, negative emotions will arise when iterations of an encounter to not allow individuals to meet expectations, an event that often leads to negative sanctioning that further erodes the flow of positive emotions. For a time, it may be possible for individuals to meet rising expectations through the elaboration of new symbols and emission of rituals that keep pace with rising expectations, but these processes only set up a situation where the failure to realize expectations will generate even more intense negative emotions that will often be externalized to meso and macro structures. Thus, for example, those who have been rising up the corporate ladder who suddenly hit a ceiling will feel doubly angry not only at others and the structure of the corporate unit but potentially the economic institutional domain as well. These dialectical processes are particularly likely in corporate units in institutional domains that distribute key resources (e.g., money, power, and prestige) and that generate ideologies emphasizing achievement, progress, and doing well as a basis for establishing expectations for self and for what would constitute positive sanctioning. But, as proposition (14) in Table 18.6 summarizes, at some point most encounters will arouse variants and combinations of disappointment–sadness and perhaps even assertion–anger that will, in turn, subtract from the flow of positive emotional energy. Repeated cycles of rising expectations followed by failures to meet these expectations will, over time, settle around a sub-optimal equilibrium point where expectations and emotional payoffs are kept in balance. However, if this downward adjustment of expectations does not occur, then persistent arousal of negative emotions will erode commitments to meso and macro structures. If the arousal of negative emotions does not lead to repression, transmutation, and external attribution, the intensity of the emotions will not be as high as is the case when repression sets into motion the dynamics summarized in proposition (17) in Table 18.6. If individuals can avoid activation of defense mechanisms and can make accurate external attributions, the emotions will still lower commitments to social structures, but not to the degree evident when repression occurs. Repression, intensification, and transmutation distort the person’s feelings and the external attributions, but even when this distortion does not occur, the distal bias of negative emotional arousal disrupts social bonds with others and lowers commitments to social structures. This lowering of commitments can come through several routes. One route is through self-attributions (proposition (15-A to 15-C)). When individuals blame self for their failings to meet expectations and/or receipt of negative sanctions, they will often experience variants and combinations of disappointment–sadness. They may also experience shame, if self was highly salient; and they will experience guilt if moral codes are highly salient. Since sadness is the dominant emotion in alienation and since shame and guilt activate the other two emotions – anger and fear – contained in second-order emotions like alienation (see Table 18.3), self-attributions can lower emotional energy and cause individuals to feel alienated from others, the encounter, or meso units in which the encounter is embedded. Of course, if self-attributions are accurate and

CH18-P374095.indd 334

5/26/2008 4:56:17 PM

Emotions and Social Structure: Sociological Theory

335

avoid activation of shame (which is the emotion most likely to activate defense mechanisms) individuals may be able to take corrective actions so as to meet expectations and/or receive positive sanctions. But if these efforts fail, then sadness becomes more chronic, causing individuals to lower commitments to others and social structures; and when guilt and shame activate the other negative emotions involved in alienation – that is, anger and fear – then commitments to social structures will be that much lower (propositions (15-B) and (15-C)). Another route is through external attributions (proposition (16)). When individuals accurately perceive that others are the cause of their failure to meet expectations or that others have unjustly sanctioned them negatively, they will experience anger at these others and, potentially, fear of them as well. If anger is expressed, it will generally breach the interaction through mutual negative sanctioning, thereby lowering commitments to others and the encounter. If fear is expressed in the form of emotions such as nervousness, anxiety, and stress, it will also breach the interaction, but in a less direct manner. Fear will make the interaction awkward and difficult, lowering commitments to others and the encounter. If individuals also experience sadness along with anger toward and fear of others, then alienation will typically be generated, thereby lowering commitments to others and social structures. These accurate external attribution dynamics will remain comparatively local, focusing primarily on others and the encounter. It is only when alienation is aroused across iterated encounters for longer periods of time that individuals begin to make more distal external attributions to meso and, potentially, macro structures. Tense interactions with others in local encounters are costly, and it is often easier to begin to blame the meso units within which encounters are embedded for failures to meet expectations and/or receive negative sanctions. Others become symbolized as representatives of the structure and culture of the more inclusive meso unit, extending alienation from specific individuals in the local encounter to the meso units in which the encounter is embedded. And, if repression is involved in this process of externalization, the alienation from the meso structure will be that much greater and will often be punctuated by episodes of anger directed at, or anxiety about, this structure. These more distal attributions become more likely to the extent that individuals experience alienation across not only iterated encounters within meso structures but also encounters in a variety of corporate units within an institutional domain or categoric units within the larger stratification system. When defense mechanisms are activated in the face of negative emotional arousal stemming from the failure to meet expectations and/or receipt of negative sanctions, the emotional dynamics are dramatically changed. External attributions become much more likely and, not only likely, but also distal as well. Once emotions are pushed from consciousness, they will often intensify in the form of sudden spikes of intense anger (e.g., rage), fear (high anxiety), or sadness (ennui, despair), all of which disrupt social bonds, breach encounters, and disrupt the meso structures within which the encounter is embedded. And when this

CH18-P374095.indd 335

5/26/2008 4:56:17 PM

336

Social Structure and Emotion

disruption leads individuals to repress shame, transmutation and intensification of this shame will lead individuals to experience intense anger and first-order combinations of anger and happiness, such as righteousness and desires for revenge (proposition (19)). Moreover, even without shame, fear and anger are likely to be transmuted into sadness and depression which will, in turn, cause fear, anger, and sadness to be mixed to generate alienation from, and reduce commitments to social structures (proposition (18)). Shame is an emotion that attacks the integrity of self, and hence, is more likely to be repressed. Often the source of shame is repressed and transmuted into anger or perhaps fear that, in turn, disrupts social relations, causing individuals to experience more shame which, in turn, may be repressed in a vicious cycle of anger/fear–shame–anger/fear-more shame. Once shame is repressed, shame is almost always transmuted to anger in ever-more intense forms, and when this anger surfaces, it disrupts social bonds (Scheff 1979; Scheff and Retzinger 1991) in ways that lead to more shame which must be repressed – thereby placing individuals in a shame–anger–shame cycle. Shame makes individuals “feel small,” and so it is not surprising that it is often repressed in some manner, but if repressed, it comes to the surface in more intense forms of anger that are directly outward toward external targets (propositions (17) and (18) in Table 18.6). If, however, shame can be acknowledged, it can motivate individuals to repair social bonds (Scheff 1979), as is outlined in proposition (12) in Table 18.5. The key is for a person to recognize the causes of their shame accurately – whether from their own actions, the actions of others, or the structure of the situation – and then to engage in behaviors that eliminate the source of shame. These kinds of repair rituals are difficult for several reasons. First, shame is painful, and individuals will often have difficulty in recognizing it for what it is and then making accurate attributions as to its causes. Another reason is that if others are the cause of shame, and if a person’s attributions are accurate on this score, it is often difficult to change another’s behavior (which may be fueled by his or her own shame–anger–shame cycles). As a result, even if an individual initially seeks to deal with the sources of shame, it may not be possible to eliminate the conditions causing this shame. Eventually, under these bleak conditions, persons will sometimes repress their shame, causing the arousal in various guises of the emotions from which shame is constructed – sadness, fear, and anger. As these negative emotions are experienced, they set up interaction rituals that can generate negative emotional energy that alienates individuals from, makes them fearful of, or causes them to be angry at the symbols and structure of the encounters and the meso units in which these encounters are embedded (proposition (18)). As a result, social structures remain viable only by the extent to which persons experiencing shame are dependent upon the resources circulating in these shame-producing encounters and/or the power structure of the micro or meso units is sufficiently great as to impose a regime of shame on its members. These structures will, however, be susceptible to change if conditions allow its

CH18-P374095.indd 336

5/26/2008 4:56:17 PM

Emotions and Social Structure: Sociological Theory

337

shamed members to organize and pursue their interest in eliminating the causes of their negative emotions. Once shame is aroused and repressed, it is typically transmuted into one of its constituent emotions but, as emphasized above, most typically anger that is then directed outward through attribution processes. These external attribution dynamics often focus on another person, usually one who is not powerful and cannot fight back (such as a spouse in an abusive relationship with a male trapped in a shame–anger–shame cycle). If others are not available as targets or are too powerful to confront, then the shame that fuels anger will generally be directed at meso structures, either the structure and culture of a corporate unit or the members of a categoric unit. In the case of corporate units, the individual will experience not only anger at this unit but also alienation (anger mixed with sadness and fear) from the unit, and if sufficient numbers of persons have the same emotional reaction, the corporate unit becomes less functional (proposition (18-A)). And, if sufficient numbers of individuals across a range of corporate units reveal these emotional responses, the viability of the macro-level institutional domain built from corporate units may decrease. When members of a categoric unit are the target of external attributions (proposition (18-B)), individuals will develop prejudicial beliefs against members of these units and, if possible, engage in acts of discrimination which, in turn, may sustain the visibility and salience of categoric unit boundaries (of both discriminators and their victims). As categoric unit boundaries solidify, especially if rates of interaction among members of different categoric units are limited by discrimination and if inequalities in the distribution of resources are correlated with categoric unit membership, the macro-level stratification system built from categoric units will be reproduced. Shame is often the result of failing to meet expectations or receiving negative sanctions in embedded encounters. When corporate units are hierarchical and punitive, they are likely to generate daily shame that cannot be expressed, with the result that it is repressed. If the shame is transmuted to anger but this anger cannot be expressed in local encounters, the meso and, potentially, macro units in which these encounters are embedded will become the target of this anger. For example, the corporate unit as a whole, the members of low ranking categoric units, the larger stratification system formed from categoric units, or even the institutional domain may all be potential targets of the anger arising from repressed shame (proposition (18-C)). Repressed shame that is transmuted into anger, which in turn must also be repressed, often pulls out the dominant emotion of shame – sadness – with individuals often feeling sad about, and alienated from, meso and macro structures. Since sadness is often an inappropriate emotion in situations, sad individuals may be sanctioned negatively or they may not have sufficient energy to realize expectations, with the consequence that they experience another level of shame. The breaking out of the constituent emotions from which shame is built – that is, sadness, fear, and anger – opens many potential emotion cycles. For example,

CH18-P374095.indd 337

5/26/2008 4:56:17 PM

338

Social Structure and Emotion

one cycle may revolve around shame–anger–shame, another around fear–anger– shame–fear–alienation–shame, still another sadness–shame–anger–shame, and perhaps other cycles that can be generated by combinations of the three negative emotions from which shame and alienation are built. Thus, the more individuals repress shame across a variety of corporate units and the longer the period of time that individuals have had to repress shame, the more likely are attributions to target macro-level structures (see propositions (19) and (20)). For example, it has been argued that terrorists are often recruited from individuals who have not realized expectations and been negatively sanctioned in meso structures in key institutional domains (economic, educational, or political), causing them to repress their shame (Scheff and Retzinger 1991; Turner 2007b, 2007c; Volkan 2004), and direct it outward to macro-level structures (with help from terrorist leaders and media portrayals of external enemies). Terrorism also depends upon the tendency of positive emotional arousal to stay local, with individuals in terrorist groups generating positive emotions in interaction rituals that generate totems or symbols personifying the group, often in juxtaposition to negative symbols of “the enemy.” From this mix of symbols, individuals will experience a sense of pride and happiness in local encounters with fellow terrorists and intense hatred directed outward against an external enemy (propositions (19-A to 19-D) and 20-A, 20-B)). These dynamics do not need to be as dramatic as they are in terrorism. They generally operate in less intense forms in corporate units or communities where individuals (at one level of a corporate unit or place in the stratification system) experience shame that is repressed and transmuted into anger and prejudicial beliefs about those believed to be responsible for the negative emotional arousal. From interaction rituals, shamed individuals charge up positive emotional energy in support of symbols marking their group boundaries (say, workers on the line in a factory) and those against whom prejudices are leveled (such as white-collar management). As this mix of positive and negative emotional energy is charged up, individuals will display anger at, and alienation from, meso and macro structures. Any time that this set of dynamics is activated – that is, positive emotional arousal at the level of encounters among those in structurally equivalent positions coupled with negative arousal toward meso and macro structures – the viability of these meso and macro structures becomes problematic. Often this combination of positive emotions, coupled with intense negative emotional arousal directed externally, can simply simmer for a long period of time, especially when those experiencing shame have few other options or those who are the object of prejudices have power. At other times, leaders of social movements can mobilize this anger, while providing a corporate unit within which encounters can arouse positive emotions and group symbols. When such mobilization occurs, it can potentially change meso and macro structures, with the success of such movements dependent upon the level of other resources that the movement can mobilize. Thus, change-oriented social movements are almost always fueled by anger at existing conditions. This anger may be direct and not intensified

CH18-P374095.indd 338

5/26/2008 4:56:17 PM

Emotions and Social Structure: Sociological Theory

339

through repression and transmutation, but if the anger comes from repressed and transmuted shame and humiliation arising from people’s inability to meet expectations and receive positive sanctions for their acts in corporate units within key institutional domains and/or their inability to be mobile within the larger stratification system, then the emotional fuel of the social movement will be that much more intense. If other resources can also be mobilized, such as recruitment of allies, money, media exposure, symbols, and organizational resources, then this underlying emotional fuel can change social structures in dramatic ways. In sum, then, negative emotional energy not only breaches the local encounter, it can be directed further outward through repression, transmutation, and attribution dynamics. When meso units are targeted by large numbers of individuals, these units become less viable, as do the macro structures built from these mesolevel units. The effects of emotional arousal on meso and macro structures will vary, depending upon which combination of negative emotions is aroused at what level of intensity with what degree of repression and transmutation. If sadness, anger, and fear aroused in encounters are acknowledged and motivate individuals to work at reducing the causes of anger and fear, the result is for encounters to develop more solidarity that will increase the viability of the corporate unit within which they are embedded. When any of these emotions is repressed, however, the emotional dynamics unleashed decrease the viability of meso and macro structures. And such is particularly likely to be the case when all three negative emotions are aroused to produce repressed shame. The more intense emotions that ensue can fuel actions that disrupt the structure and culture of meso and macro units, if individuals experience these emotions on a consistent basis. As attribution processes target structures, the connection between the locale where negative emotions have been initially aroused and the target of these negative emotions may be severed (proposition (20-A)). Repression distorts the feedback system behavior outputs and informational inputs such that persons may target members of a social category having little to do with their emotional experiences (say, Jews in the shame culture of pre-Nazi Germany) or corporate in which they have never participated (e.g., corporate units in western societies by terrorists). Repression and transmutation can thus cause individuals to target remote social structures which had very little to do with their original negative emotional arousal. When elites and media in the Middle East, for example, attack the institutional systems of the west on a daily basis (thus diverting people’s anger at their failings within corporate units of their own societies), the process of repression, transmutation, and external attribution will target social structures and cultures far removed their actual origins. The distal bias of negative emotional arousal will become ever-more distal to the degree that the causal connection between the origin of emotional arousal and social structures is severed through repression and external attribution. The negative emotional energy becomes diffuse and subject to manipulation by actors with vested interests in deflecting negative emotions from their actual structural origins (proposition (20-A, 20-B)), the targets of attribution can be symbolically demonized, the

CH18-P374095.indd 339

5/26/2008 4:56:17 PM

340

Social Structure and Emotion

local networks in which encounters arouse positive emotional energy can also generate group symbols that portray the cause in positive terms and enemies in negative terms, and the corporate units mobilizing these networks and encounters have sufficient resources to sustain themselves over time and in diverse locales (proposition (19-A to 19-D)).

CONCLUSION The theory that unfolds in these twenty propositions (for a more detailed version the theory, see Turner 2007a). In general terms, positive emotional arousal leads individuals to emit positive sanctions to others and, reciprocally, receive positive emotional responses from others through interaction rituals that charge up group symbols and cause individuals to become committed to each other and the encounter. Positive emotions have a proximal bias, causing individuals to make self-attributions and to see the immediate situation as responsible for their success in meeting transactional needs. But, as the propositions specify, there are structural conditions that increase the likelihood that positive emotions will be extended to meso structures and the macro-level structures in which encounters are embedded. Failure to meet expectations and negative sanctions have the opposite effect; there is a distal bias to negative emotional arousal as individuals attempt to protect self; and if defense mechanisms are activated, the emotional dynamics change in significant ways. One way is that repressed emotions become more intense, often erupting in spikes of anger, fear, or sadness that disrupt encounters and the meso structures in which they are embedded. Another way is through transmutation of emotions into not only more intense but also different affective states. When all three negative emotions are combined to produce shame, repression becomes more likely; and repressed shame often surfaces as anger directed outward at meso and macro structures, particularly if individuals have experienced consistent failure to realize expectations in the meso structures of key institutional spheres and in their efforts to be mobile within the stratification system. Shame and, at times, guilt make it possible for individuals to feel alienated from others, encounters, meso structures, and macro structures because; for, once the three negative emotions in shame and guilt are activated simultaneously (see Table 18.3), the neurological conditions for alienation are present. Thus, when either shame, guilt, or alienation is activated, the chances of the other two second-order emotions being aroused increase. Just how distal the targets of negative emotions become depends upon the specific emotions repressed, the operation of attribution processes, and the flow of positive emotions in local encounters. If shame is repressed, it will generally be transmuted into anger directed at meso and macro structures. If positive emotions are aroused and symbolized in local encounters, this combination of positive emotions that evidence a proximal bias and negative emotions that exhibit

CH18-P374095.indd 340

5/26/2008 4:56:17 PM

Emotions and Social Structure: Sociological Theory

341

a distal bias leads to ever-more distal attributions. And, if positive sanctions in local encounters are meted out for making such distal attributions, external attributions become even more distal. World-level terrorism is probably the extreme case where dense networks of terrorists provide positive sanctions and a sense that expectations are met by planning and executing violence against targets in other nations. The shame experienced by terrorists in key institutional spheres or in the stratification system of their societies of origin is repressed and transmuted into intense anger, hatred, righteous anger, and desires for revenge which, through media manipulation and positive sanctions for such anger within terrorists networks, are directed outward toward enemies far removed from the original source of the shame. Thus, people’s anger at, alienation from, and lack of commitment to social structures almost always involve a mix of repressed shame that is transmuted into anger and alienation (and occasionally fear) directed at meso and, at times, macro structures. These more distal attributions become ever-more likely when individuals are rewarded with positive sanctions in local encounters and dense networks for making external attributions. Whatever the merits of these ideas, sociological theorizing on emotions needs to recognize the importance of repression and attribution as key forces that shift the emotional experiences of individuals and those who must respond to these individuals and that alter the targets of emotional arousal. Sociology can, I feel, gain much from psychoanalytic theorizing because of the distorting effects of repression and attribution; and indeed, it would difficult to explain individuals’ reactions to meso and macro structures without some understanding about the degree to which negative emotions have been repressed, intensified, and transmuted. Conversely, it would not be possible to understand the microdynamics of commitments and attachments to meso and macro structures without appreciating the proximal bias and self-attributions that accompany positive emotional arousal. In fact, I would argue that the sociology of emotions will be one way to fill in the elusive “micro–macro” gap in sociological theorizing.

REFERENCES Burke, Peter J. 1991. “Identity and Social Stress.” American Sociological Review 56:836–849. Collins, Randall. 2004. Interaction Ritual Chains. Princeton, NJ: Princeton University Press. Durkheim, Émile. [1912] 1947. The Elementary Forms of the Religious Life. New York: Free Press. Goffman, Erving. 1961. Encounters: Two Studies in the Sociology of Interaction. Indianapolis, IN: Bobbs-Merrill. Goffman, Erving. 1963. Behavior in Public Places: Notes on the Social Organization of Gatherings. New York: Free Press. Heise, David R. 1979. Understanding Events: Affect and the Construction of Social Action. Cambridge: Cambridge University Press. Kemper, Theodore D. 1987. “How Many Emotions Are There?” American Journal of Sociology 93:262–263. Kemper, Theodore D. (ed.). 1990. Research Agendas in the Sociology of Emotions. Albany, NY: SUNY Press.

CH18-P374095.indd 341

5/26/2008 4:56:17 PM

342

Social Structure and Emotion

Lawler, Edward J. 2001. “An Affect Theory of Exchange.” American Journal of Sociology 107:321–352. Luhmann, Niklas. 1982. The Differentiation of Society. In: S. Holmes and C. Larmore (trans.), New York: Columbia University Press. McCall, George P. and J. L. Simmons. 1978. Identities and Interactions. New York: Basic Books. Plutchik, Robert. 1980. Emotion: A Psychoevolutionary Synthesis. New York: Harper and Row. Scheff, Thomas J. 1979. Catharsis in Healing, Ritual, and Drama. Berkeley, CA: University of California Press. Scheff, Thomas J. 1988. “Shame and Conformity: The Deference-Emotion System.” American Sociological Review 53:395–406. Scheff, Thomas J. and Suzanne M. Retzinger. 1991. Emotions and Violence: Shame and Rage in Destructive Conflicts. Lexington, MA: Lexington Books. Smith-Lovin, Lynn. 1990. “Emotion as the Confirmation and Disconfirmation of Identity: An Affect Control Model.” In: T. D. Kemper (ed.), Research Agendas in the Sociology of Emotions, pp. 238–270. New York: SUNY Press. Stets, Jan E. and Jonathan H. Turner. 2006. Handbook of the Sociology of Emotions. New York: Kluwer. Tangney, June Price and Ronda L. Dearing. 2004. Shame and Guilt. New York: Guildford Press. Thamm, Robert. 2006. “The Classification of Emotions.” In: J. E. Stets and J. H. Turner (eds), Handbook of the Sociology of Emotions, pp. 1–28. New York: Kluwer. Turner, Jonathan H. 1999. “Toward a General Sociological Theory of Emotions.” Journal for the Theory of Social Behaviour 29:109–122. Turner, Jonathan H. 2000. On the Origins of Human Emotions. Stanford, CA: Stanford University Press. Turner, Jonathan H. 2002. Face-to-Face: Toward a Sociological Theory of Interpersonal Behavior. Stanford, CA: Stanford University Press. Turner, Jonathan H. 2003. Human Institutions: A Theory of Societal Evolution. Boulder, CO: Rowan and Littlefield. Turner, Jonathan H. 2005a. “A New Approach for Theoretically Integrating Micro and Macro Analysis.” In: C. Calhoun, C. Rojek, and B. S. Turner (eds), Handbook of Sociology, pp. 230–260. London: Sage. Turner, Jonathan H. 2005b. “The Dynamics of Differentiation and Integration.” In: C. Lahusen and H. Aretz (eds), Die Ordnung der Gesellshaft, pp. 93–120. Berlin: Peter Lang. Turner, Jonathan H. 2007a. Human Emotions: A Sociological Theory. London: Routledge. Turner, Jonathan H. 2007b. “The Social Psychology of Terrorism.” The Part/Whole Approach to Terrorism. Boulder, CO: Paradigm. Turner, Jonathan H. 2007c. “Self, Emotions, and Extreme Violence: Extending Symbolic Interactionist Theorizing.” Symbolic Interaction. In press. Turner, Jonathan H. and Jan E. Stets. 2005. The Sociology of Emotions. New York: Cambridge University Press. Volkan, Vamik. 2004. Blind Trust: Large Groups and Their Leaders in Times of Crisis and Terror. Charlottesville, VA: Pitchstone Publishing. Weiner, Bernard. 2006. Social Motivation, Justice, and Moral Emotions: An Attributional Theory. Mahway, NJ: Lawrence Erlbaum.

CH18-P374095.indd 342

5/26/2008 4:56:17 PM

19 Emotion and Inequality in Maternity Care: Anguish and Anger in Prenatal Services for the Poor Linda E. Francis1, Candyce S. Berger2 and Karina Kim2 1

The University of Akron, Akron, OH, USA Stony Brook University, Stony Brook, NY, USA

2

ABSTRACT For this study, we brought together groups of disadvantaged women of color for focus groups designed to assess a county perinatal services system. Many of the systemic barriers and adverse outcomes identified by quantitative health services research are apparent. However, emergent from the discussions are two additional subtexts on: (a) the emotions and emotion norms in the medical system for the poor, and (b) the manner in which the women perceive inequality as maintained or exacerbated through medical interactions. Drawing on research on inequality in interaction and emotion, the author illustrates how the expectations of these women are both created by, and yet clash with those of their medical providers. For poor African American women in particular, results show their interpretations of their medical experiences often lead them to react with anger and aggression. The emotional mismatch between doctor and patient appears to be an unacknowledged barrier to treatment, increasing the already high risk of negative, even tragic consequences for mother and child. Social Structure and Emotion

CH19-P374095.indd 343

343

Copyright © 2008 by Elsevier Inc. All rights of reproduction in any form reserved.

5/26/2008 4:57:34 PM

344

Social Structure and Emotion

BACKGROUND This study began as an applied needs assessment for a network of local nonprofit social services providers. The impetus for this study was the juxtaposition of two apparently contradictory facts: The county in question is consistently among the wealthiest counties in the country, and yet also has an extremely large disparity in infant mortality rates between whites and people of color, especially African Americans. Working with a county-wide Perinatal Services Network (PSN), a loose voluntary network of service providers for women and children in the county, we designed an exploratory study to try to tap into the prenatal needs of disadvantaged women. While we already knew many of the structural barriers, we sought insight into the experiences of poor pregnant women in a suburban health care system. To do the study, we brought together groups of disadvantaged women of color for focus groups designed to assess perinatal services for the poor in the county. We already knew many of the systemic barriers and adverse outcomes identified by quantitative health services research, but we sought detailed descriptions of the women’s experiences. Of particular interest in our results was that in addition to the overt findings on specific service needs, we uncovered two emergent subtexts regarding emotion and inequality. These subtexts reflected: (a) the emotions and emotion norms of African American women in the medical system for the poor, and (b) the manner in which the women perceive inequality as maintained or exacerbated through interaction more than through concrete obstacles or structural barriers in medical settings. Drawing on research on inequality in interaction and emotion, we illustrate how the expectations of these women are both created by, and yet clash with those of their medical providers. For poor African American women in particular, results show their interpretations of their medical experiences often lead them to react with anger and aggression. The emotional mismatch between doctor and patient appears to be an unacknowledged barrier to treatment, increasing the already high risk of negative, even tragic consequences for mother and child.

DATA AND METHODS Our study design involved six focus groups with consumers of maternal and child health services, of which three are included in this study. These three groups were entirely comprised of African American women, one with pregnant women, one with women who had given birth within the previous twelve months, and one group of women in substance abuse recovery. Among the three groups not included in this study were two groups of Latina women and one of adolescent mothers of diverse ethnicities. Flyers were sent to agencies participating in the network, instructing them to give copies of these flyers to their clients. Potential subjects contacted the County

CH19-P374095.indd 344

5/26/2008 4:57:34 PM

Emotion and Inequality in Maternity Care

345

Perinatal Network to register for the focus groups. Each group had from eight to ten participants who were homogeneous on the criteria for the group session. Each group began with an introduction of the facilitators and the study. The two facilitators were either both the authors or one author and a graduate student. All participants were given a copy of the consent form, which we went over with them, and about which we encouraged them to ask questions. When all had signed, we turned on the tape recorders and began the interview. We ran the focus groups in a semi-structured interview format. Each group had 5 open-ended questions presented to them on a flip chart, with each question on a separate page. The topics of these questions (including prompts) were: ● ●

●

●

●

Medical Care (prenatal and postpartum) Social Services (public assistance, Medicaid, housing, WIC, transportation) Personal Challenges (smoking, alcohol/drug abuse, stress, depression, or psychological issues, teen pregnancy and/or teen parenting) Locating Information/Services (word of mouth, ads, phone book, internet, pamphlets) Changes would make (in system, or in own actions if could do over again).

Once discussion was underway, the facilitators intervened only as necessary to guide, probe, or provide support as needed. The audiotapes of the focus groups were transcribed and the transcripts checked against the tape by the facilitator who ran the group. They were then uploaded into a qualitative text analysis program and coded independently by both authors for content on needs and services. We met regularly to check our agreement on interpretation. Upon completion of this round of coding, the first author returned to the data to code for emotion and emotion norms. This chapter is the result of the second coding.

OVERT RESULTS The meat of the discussions in all groups took place under the topics of medical and social services, particularly the problems the women had faced with them. Unsurprisingly, the main barriers and problems found in survey-based studies also made their appearance in this project. However, what the women really wanted to talk about were their experiences of being treated poorly, both in the sense of poor quality health care and poor interpersonal treatment by staff. This comes out overtly in direct statements about their main complaints and concerns. These are given below in order by the importance given to them by the women in the groups. In terms of medical services, the women felt (1) that providers were uncaring, disrespectful, insensitive, and sometimes even punitive in their interactions with the women under their care. (2) There was widespread agreement among the women that providers made very little effort to talk

CH19-P374095.indd 345

5/26/2008 4:57:34 PM

346

Social Structure and Emotion

to them about their health and that of the baby in terms that they understood. They often did not know if what the provider said meant everything was okay or not. (3) Poor communication lead them to worry about the quality of their care, particularly in comparison to middle class, usually white, women they saw being treated in the same hospital. (4) Related to this was a strong suspicion that they were victims of discrimination, and that white middle class women were getting better care because they had insurance to pay for it. (5) Many women complained about fragmented care, never getting to see the same doctor more than once. (6) Finally, the hassles of waiting for weeks for appointments, waiting for hours in the waiting room, of dirty, unsanitary clinics, and incomprehensible paperwork are widely recognized problems of public health care. In social service settings, the pattern repeated itself. Women were frustrated by the impossibility of dealing with the system. They complained that (1) the requirements for information were disorganized, and they would frequently have to make multiple trips to bring in various documents instead of being told up front what would be needed for the entire process. (2) Often women felt that the case workers were not telling them things that would have been helpful, and so only people who knew what to ask got access to some of the most highly valued services. (3) Widespread disrespect and treating the women as lacking any credibility were among the issues that generated the most anger among women, combined with (4) long waits for services. For a pregnant woman in this county to get Medicaid, for instance, took an average of six months. (5) To these women, confusing rules and inconsistency in their application seemed to be indicators that Social Services did not want to help them. (6) This was especially the case when they found themselves in Catch-22 situations, such as not being able to get housing because they had no income, but being ineligible for financial assistance because they lacked a stable address. The women understandably evinced an enormous level of frustration from being forced to bow down perpetually to a system that they felt was deliberately withholding help from them.

EMERGENT RESULTS In the process of coding the above categories of overt barriers, we were repeatedly struck by the emotional tenor of many of the discussions. Fear and frustration, anger and anguish, sorrow, resignation and depression were common themes. Among the African American women in particular, the stories created emotion-colored images of desperate women trying to survive in a punitive, uncaring, unfair world. Upon completion of the report to the county, we returned to the data to code for the emotion in the stories, not just the problems. We found that what emerged from this process was a deeper story line with two subtexts: that of the women’s emotional experiences of their medical and social services; and that of the emotion norms and emotion management that they generate among themselves to deal with their circumstances. These emotion norms

CH19-P374095.indd 346

5/26/2008 4:57:34 PM

Emotion and Inequality in Maternity Care

347

encourage actions and reactions in medical settings which are understandable in the context of the women’s interpretations, but which often appear to put them in conflict with the providers who are supposed to be helping them. We present these results in a logical order for ease of understanding, but it is important to remember that the linearity is deceptive. The reality is much messier, with substantial variation in how and when each woman experienced these processes. What we try to capture here are the key commonalities in experience, including emotional crisis points and the construction and enacting of intragroup emotion norms. POOR CREDIBILITY

Most of the women in the study begin their emotional odyssey with a trip to social services to get basic assistance, including housing, Medicaid, WIC (nutritional supplements for Women, Infants and Children), and food stamps. What they find consistently with the Department of Social Services (DSS) is that their biggest problem is having to try to prove that they need help. As pointed out before, the system is set up to be demeaning, intrusive, and discouraging to applicants. Unsurprisingly, this results in a great deal of frustration and resentment: A:

K:

Well, I have been trying to get public assistance since my husband went to jail. And I having a really bad time, it’s hard. They want to know all your business and information in the world. [in agreement] You have to tell them when you pee!

Another woman states a common complaint: They asked me for about 13 different things, I brought ‘em in and now they want something else.

A particular frustration was the fact that often, the DSS staff simply did not believe they were telling the truth. You know, they made me do a special investigation, because they didn’t believe I was telling the truth about the place I was living at. Then they turned around and told me that I needed some w2 forms. I’m like, “I never had a job!”

Among most of these women, DSS symbolized the racism and discrimination perpetrated on them by a white-dominated society. They felt that poor, black women were treated badly simply because of who they were. Credibility continued to be a problem when they got to the medical provider’s office. A number of women reported experiences of not being believed, of being told they were imagining things or over-excited, and of being denied timely care. They tell me to wait, and I keep tell them I can’t breathe and it’s in the doctor’s office. She telling me, “oh, well your nails ant blue and your lips ant blue and this and that” and I said “stupid I can’t even breath and I can’t get out of the bed to go into the kitchen” it’s so hard and its really bad and I thought if you tell the doctor something, they should look

CH19-P374095.indd 347

5/26/2008 4:57:34 PM

348

Social Structure and Emotion

into it. It’s not good. When I went home, I was wheezing and stuff. She didn’t do nothing at the doctors office. She tic-tac on my stomach and that’s it.

WORRY AND FEAR OF POOR CARE

For these women, their greatest fear was that the lack of credibility meant that they were not getting adequate care. In other words, they worried that medical services would reflect their experience with social services. Unlike the average middle class woman, they did not necessarily expect their pregnancies to be uneventful and safe. Most of them knew people who had had bad, even catastrophic health events partly as a result of inadequate health care. In our study alone, we had three women among our African American respondents who had had stillbirths. While this is a self-selected sample, this is indicative the original motivator for the study: the disproportionately high rates of infant mortality among African Americans on Long Island compared to whites and to other parts of the country. One woman’s tragic experience illustrates this connection of poor credibility and poor care in the minds of these respondents: My doctor told me nothing. First of all if he keep checking me he will see I would lose 10 lbs. This is not normal when you are pregnant. I went into the emergency room. They had to put a tube down my nose to pump out my blood, in my stomach. He’s reading the information to me telling me you’re fine. I think every other day I was in the emergency room. And he would tell me, go ahead, go home. I would throw up on the table, he would see the bile, coming out of the lining of the stomach, he would see me throwing this up, and he would say you be ok, you’re ok. And when they told me my son was dead inside me, it was like, … it was so unprofessional. I wouldn’t want a dog to go there.

When discussing their medical care, then, the women in the study identified not just frustration as a result of not being believed, but a great deal of anxiety and worry over the health of the baby. Under these conditions, a lack of concern expressed by providers – which was probably intended to be calming – actually fed into their perpetual fear that they were receiving inadequate care. If the provider did not communicate with them, they did not interpret this as meaning there was nothing wrong, rather they saw it as not paying attention to them or not believing them. Several women railed against the distant and unemotional treatment of their providers, considering it poor care: It’s like you worry about your pregnancy they throw like a wall, they probably think like “oh she is another one of your pregnant women.” Or you know, “she be black” or whatever. So when I went for my visit it was like I was talking to the doctor, you know I’m worried that something is going to be wrong with my baby. They had no care in the world. Sick the whole nine months of my pregnancy, I had morning sickness. I wasn’t thinking my baby was going to be all right and all they did was check the heartbeat and my blood pressure anytime I went my blood pressure would go up and down, up and down. And that was it.

As a group, the women appeared to have a widely shared definition of reality that included uncaring doctors and substandard care. They encouraged each other in their worry and anxiety, and showed a culture of deep mistrust of all medical

CH19-P374095.indd 348

5/26/2008 4:57:34 PM

Emotion and Inequality in Maternity Care

349

providers. This distrust of medicine among African Americans has been recognized for many years (Gamble 1997). What comes from this study, however, is a picture of how their expectations and experiences are interpreted by the women to form a set of norms of patient emotion and behavior. In their supportive interactions with each other (of which there were many in each group), the efforts of a worried woman’s peers were not directed at allaying her fears, making her feel better, or assuring her that everything was going to be fine. Rather the other women sought to validate the woman’s fears – thereby increasing her worry – and encourage her to take action on them. Such action usually took the form of demands for care, as in the case of Ursula: URSULA:

So there is just a lot with my family, just really stressing me out bad and I been stressed for three months … I just want the baby to be ok. KIERA: They could give you a stress test. You know the stress test, you know to test the baby’s stress level. TANYA: They will only give her a sonogram; they won’t give her a stress test. KIERA: You have to beg, for a stress test. I waited four months. LATOYA: They will get tired of hearing you. It’s a shame it has to be that way. But you have to do what you have to do. URSULA: A stress test. I can’t take it anymore. As long as the baby is ok. … [T]hey say [stress] can’t hurt the baby. KIERA: Don’t listen to them. URSULA: I know. In the above example, Ursula’s peers are most certainly not encouraging her to trust her doctor; they are telling her not to believe him. In the process, they are also encouraging her to demand care – in this case, an inappropriate medical test. A stress test uses ultrasound technology to determine if the uterine environment is deteriorating due to a medical condition (such as hypertension or gestational diabetes), or due to being past the due date. It does not measure emotional stress in the sense that these women are discussing. Nonetheless, one can speculate on the resulting conversation with a provider – who will no doubt deny the request for the test, leaving the woman worried once again that she is not receiving adequate care. ASSERTIVENESS AND DEMANDS FOR CARE

For these women, then, the appropriate way to deal with one’s anxiety over pregnancy was to be assertive and demanding of better care. And better care, to them, meant quite literally more care, and more intrusive care. Interestingly, a corollary of not feeling believed by their providers was not believing their providers in return. There is a repeated message among these women that doctors should be expected to prove it when they say that the women’s symptoms are nothing

CH19-P374095.indd 349

5/26/2008 4:57:34 PM

350

Social Structure and Emotion

to worry about – and proof should be in the form of medical tests and intrusive interventions. To deal with one’s worry, then, the appropriate action was to be assertive, to demand care, specifically to demand proof that what the doctor was doing was good enough. We can see this kind of action encouraged in this quote: ROBERTA: Right now, I go just about every week. When I go back I’m going to demand a sonogram. I don’t care what she got to say. ALL: Yeah! ROBERTA: I don’t care what she got to say! Because, she wants to over talk me. Every time I try to tell her something. So go to the head of that office, try to start calling now you know PATTY: and making a fuss now! ROBERTA: She don’t have to tell me something, I know my body! So the appropriate response to anxiety is defined as assertive demands. This is a form of managing each others emotions, evoking anger in the face of perceived injustice, rather than allowing the person to remain in a position of powerlessness. ANGER AND AGGRESSION TOWARD PROVIDERS

The next step up the continuum from anxiety to assertiveness, of course, is anger, and anger is very frequently expressed by these women. When their demands for care, especially expensive and intrusive care go unheeded, quite understandably they get mad. From their perspective, they are fighting for their babies’ lives and well-being in a system that does not respond to them. Confrontations like the one below were not uncommon: Everybody says County Hospital they don’t treat you right, in there they have you sitting there for hours. My cousin was sitting there and her water broke, and the lady just left her sitting there. I had to actually scream on that lady and they don’t try. She just sitting there and her water was like flowing out of her and she wasn’t in the bed or nothing, she just standing there. And she cursing, and I’m trying to calm her down and like you said, County Hospital bad.

Here we see that lack of response is directly interpreted as bad care, and as justifying an assertive, even angry response. From the women’s perspective, this is being a good patient – making sure you get the treatment that medical institutions want to deny you. The problem here is that this behavior of confrontation and demand directly contradict the expectations for patient behavior that most medical staff bring to the encounter. While we have nothing more than the most indirect data on provider reactions to these women, common sense tells us that such reactions are unlikely to be positive. Yet the women continued to define being assertive, argumentative and demanding of services as appropriate patient behavior. Their definition of the situation called for mistrust and anger, and acting on these emotions was the right thing to do.

CH19-P374095.indd 350

5/26/2008 4:57:34 PM

Emotion and Inequality in Maternity Care

351

On a few occasions, assertiveness crossed over into overt aggression against a provider, especially when faced with apparent proof of error or incompetence. One woman described her interaction with a nurse upon arrival in the hospital to give birth: You see, you have to see a doctor in [the hospital], because when I had my daughter there, my water broke, “oh my God my water broke!” She goes, “she think her water broke.” You know that I took the needle, that IV, and I just threw it right at her. She was like, “miss, miss …” I was like, “listen I told you my water broke, don’t you listen, don’t you see all this water?” “Oh I’m sorry, I thought you was just, you know.” It was like, you didn’t know what your water was!

Yet more aggressive was the story of another woman in dealing with a singularly insensitive intern. I had this one intern. He was rough, he was just winding his finger right in you and my placenta was ripped, because I was bleeding for like seven months. So I’m going to the hospital since April and all the doctors all saw that my placenta was ripped. He was the only one that could not find it. I had to actually show the man, “oh this is great.” I don’t know what I was doing, I’m showing the man on the sonogram, “right here, do you not see this is ripped?” And he is just running his finger, and I’m screaming there like “Ouch! Ouch!” And I said “you do that one more time and I’ll kick you to that wall!” And he was just strumming his fingers and I hauled back and kicked him, and his whole back of his head hit the wall. And I didn’t care, he was bleeding and everything.

What was remarkable about both these stories – other than the stories themselves – was the pride with which they were related, and the appreciation they received from the other women. The storytellers were not only unapologetic; they were quite unabashedly pleased with themselves for their actions. In both cases the attitude of the other women in the group was unilaterally supportive, as though these women had done something highly admirable. My own interpretation of the reaction is that it was akin to what would follow someone standing up to a bully and making the bully back down. There was no denying that the storytellers got positive reinforcement from the other women for their stories.

NEGATIVE CASE

Of course, not every woman had bad experiences in her prenatal care. A few were happy with their doctors and the medical services they had received. We cannot make any generalizations about the proportion of women who had good or bad care because there is probably bias in a self-selected sample: those with something to complain about may have been more likely to participate. However, to check on the emergent patterns among those unhappy with their care, we sought negative cases among those who were pleased. In particular, we wanted to see if these patterns were widely accepted, or if they were constructed only by the women dissatisfied with their care. We were unable to confirm or disconfirm some findings, as the women simply did not always talk about the same things. But one finding was supported: the equation of highly attentive and intrusive care with good care.

CH19-P374095.indd 351

5/26/2008 4:57:34 PM

352

Social Structure and Emotion

Those who were happy with their care had doctors who were unusually responsive and cautious. The quote below illustrates this pattern, and is particularly interesting because of the fact that the speaker is one of the women with a prior stillbirth: I can kiss my doctor, and recommend him to anybody here. Five years ago I lost a baby in delivery and I was four months pregnant. [This doctor] took a sonogram right away and they saw that my cervix was opening up and I guess years ago, my baby was out of my cervix and so my cervix was incompetent. This time I had to go the next day for emergency surgery and they had to stitch my cervix up … and I was in the hospital for about five weeks on a slant, like the Titanic and I couldn’t do anything and basically he saved the baby! I’m better off with the doctor that I’m with now, it was as if I was his child or something the way he took care of me. I’m grateful and I’m blessed.

Most middle class women would not feel blessed by enforced bed rest, but this woman makes it clear she is very happy with this treatment. Unfortunately, the current state of our health care system discourages doctors from being attentive and responsive in general, and insurers would certainly sanction any provider who ordered tests just because his patient wanted them. Over-service is not a viable way for providers to earn the trust of a mistrustful patient population, and since refusal of services seems to feed into that distrust, the situation is unlikely to go away soon.

SUMMARY To summarize, in this chapter, we have found that economically disadvantaged African American women have frequent experience with poor credibility and lack of communication in medical settings. Among women whose fear of poor health outcomes has a real basis, the sense of not being believed, of having their symptoms discredited, ignored or explained away in unintelligible jargon generates anxiety about their pregnancy and care. The lack of response to their anxiety – regardless of whether the anxiety is justified or not – appears to these women to be a symptom of inadequate care. The African American women in our study responded to this anxiety among their peers not by encouraging trust in their doctors, but by encouraging distrust. Group emotion management defined doctors’ lack of response as injustice, calling for anger and selfassertion. Demands for care, especially medical tests and other intrusive care, were deemed the appropriate patient behavior by these women. Assertiveness, even outright aggression in the face of poor care, was a positive action, and was met with approval by peers. Thus the approved emotion and expression norms of medical settings were, for these women, in direct contradiction to the expectations into which the doctors and providers were socialized.

DISCUSSION This study speaks to the role of emotion in inequality. This is not a new idea, it goes back to the work of Hochschild (1983). And certainly both strain theory

CH19-P374095.indd 352

5/26/2008 4:57:34 PM

Emotion and Inequality in Maternity Care

353

(Merton, 1938) and stress theory (Pearlin 1989; Thoits 1995) clearly show that negative emotion is a result of inequality. But the findings in this study illustrate a pattern of emotional response to perceived inequity. As Schwalbe et al. (2000) assert, inequality is reproduced not by reified institutions, but by patterned interpersonal processes. This study explores some of these processes. Anger is a potent emotion, and it can serve to empower the person experiencing it. Indeed, anger can give the angry person at least the illusion of having some control over their circumstances (Francis 1997). Anger, as Lively and Heise demonstrate (2004) is a short emotional distance from fear – and fear is where these women begin their journey. Unfortunately, anger is not an emotion deemed appropriate for patients in medical settings (Baker, Yoels, and Clair 2003; Olesen and Bone 1998). The cultural authority of medical doctors (Starr 1982) provides them with enormous status and power, most particularly in their own bailiwick of the examining room. Patients, correspondingly, have little power in medical encounters. Theories of emotion which take into account such discrepancies in identity attributions can illustrate the expected patterns for medical encounters. According to MacKinnon’s (1994) theorems of Affect Control Theory, for instance, a doctor is a powerful, highly evaluated, and moderately active actor. A patient would be correspondingly weak, passive, and slightly bad. In their interactions, the doctor might be predicted to feel confidence toward him or herself, and pity toward a patient. A patient might feel fear or anxiety in general, and hope toward a doctor. Sentiments such as these are in keeping with established cultural norms of medical interaction. Run through Interact (Heise 2007), then, an act such as “patient kicks doctor” would result in a fairly extreme deflection. This raises the question of how the women in this study acquired sentiments calling for anger and aggression in medical encounters. Lupton (2003) argues for the crucial role of trust in shaping patient–doctor interaction. In contradiction to the existing medical assumption of the purely rational and self-reliant patient, Lupton asserts that emotion has a crucial place in the doctor–patient relationship. Without some degree of trust in their doctor’s abilities and goodwill, patients would most likely experience even more fear, uncertainty and vulnerability than they already do. As we have seen in the above study, these women do indeed lack trust in their providers and the medical establishment more generally. More specifically, where they lack trust is in the goodwill of the providers, rather than in their abilities. The women believe that needed services are being withheld from them. Thus the main fundamental change in these women’s sentiments is that doctors are no longer evaluated positively, but negatively. And the likely respond of a weak person to a strong bad person is fear – and that, indeed, is what they express. What we do not know is exactly how medical providers react to these poor black women who are assertive, angry, even aggressive to them, particularly since from their own perspective, they most likely do not see they have done anything to provoke such a reaction. We can, however, speculate that the interaction,

CH19-P374095.indd 353

5/26/2008 4:57:34 PM

354

Social Structure and Emotion

premised on two entirely disjunctive definitions of the situation, will usually not proceed smoothly and constructively for all concerned. Future research needs to focus on understanding this venue of reproducing inequality. From these data, it is clear how these women perceive their care in medical and social services settings – and their picture is not pretty. But it is a picture of a reality constructed in the closed context of focus groups by the women participating. I bring this up because the power of the women’s words makes it easy to overlook the fact that they are just one piece of a puzzle. Focus groups are somewhat artificial, in that one particular definition can come to dominate the discussion. It need not be the only definition possible, not even the only one held by everyone in the room. Once a particular definition of the situation gets established in a focus group, it takes some courage to speak out and disagree with the group. The role of a focus group facilitator is to try to encourage just that, but we have no guarantee that we were successful. Nonetheless, this fact does not undermine the validity of the constructed definition. It undoubtedly reflects a real appraisal, as the same perspective arose again and again in multiple independent groups, among women who did not know each other. So to say that it is a constructed definition in no way invalidates its impact on these women. Is it a true reality? Does it have objective existence? For the purposes of this study, it probably does not really matter. These women have intersubjectively interpreted their experiences and have created an explanation of reality that for them can be agonizingly true in its consequences.

CONCLUSIONS From this study, we can see that the emotion and expression norms of these women understandable in context of their experiences and their interpretations of behavior in the light of those experiences. However, their emotion and expression norms encourage actions that are almost diametrically opposed to the traditional ones of medical environments – screaming at one’s doctor is not widely considered good patient behavior. We do not have data on provider’s views of poor black women in public health care settings, but this study certainly calls for such. What are the consequences of these conflicts in terms of medical staff attitudes, not to mention actual care, toward poor African American women? Of immediate interest, however, is the fact that divergent definitions of the situation between patient and professional produce vastly different sets of understandings of what is going on. The resulting emotions, and the emotion management efforts done with peers, generate definitions that are even further from those of the doctors providing care to these women. The structural constraints, cultural history and socialized expectations for emotional expression of both parties combine to create potential barriers to care. The expression (and nonexpression) of emotion in medical interaction itself, then, can become a means of perpetuating disparities in care and outcomes.

CH19-P374095.indd 354

5/26/2008 4:57:34 PM

Emotion and Inequality in Maternity Care

355

REFERENCES Baker, Patricia S., William C. Yoels, and Jeffrey M. Clair. 2003. “Emotional Expression During Medical Encounters: Social Disease and the Medical Gaze.” In: V. James and J. Gabe (eds), Health and the Sociology of Emotions, pp. 173–200. Oxford: Blackwell. Francis, Linda E. 1997. “Ideology and Interpersonal Emotion Management: Redefining Identity in Two Support Groups.” Social Psychology Quarterly 60:153–171. Gamble, Vanessa Northington. 1997. “Under the Shadow of Tuskeegee: African Americans and Health Care.” American Journal of Public Health 87:1773–1778. Heise, David R. 2007. Interact. Retreived from www.indiana.edu/~socpsy/ACT/interact.htm, on August 24, 2007. Hochschild, Arlie R. 1983. The Managed Heart. Berkeley: University of California Press. Lively, Kathryn J. and David R. Heise. 2004. “Sociological Realms of Emotional Experience.” American Journal of Sociology 109(5):1109–1136. Lupton, Deborah. 2003. “Your Life in Their Hands: Trust in the Medical Encounter.” In: V. James and J. Gabe (eds), Health and the Sociology of Emotions, pp. 157–172. Oxford: Blackwell. MacKinnon, Neil J. 1994. Symbolic Interactionism as Affect Control. Albany: State University of New York Press. Merton. 1938. “Social Structure and Anomie.” American Sociological Review 3:672–682. Olesen, Virginia and Debora Bone. 1998. “Emotions in Rationalizing Organizations: Conceptual Notes from Professional Nursing in the USA.” In: G. Bendelow and S. Williams (eds), Emotions and Social Life: Critical Themes and Contemporary Issues, pp. 313–329. London: Routledge. Pearlin, Leonard I. 1989. “The Sociological Study of Stress.” Journal of Health and Social Behavior 30:241–256. Schwalbe, Michael, Sandra Godwin, Daphne Holden, Douglas Schrock, Shealy Thompson, and Michele Wolkomir. 2000. “Generic Processes in the Reproduction of Inequality: An Interactionist Analysis.” Social Forces 79(2):419–452. Starr, Paul. 1982. The Social Transformation of American Medicine. New York: Basic Books. Thoits, Peggy A. 1995. “Stress, Coping, and Social Support Processes: Where Are We? What Next?” Journal of Health and Social Behavior (extra issue):53–79.

CH19-P374095.indd 355

5/26/2008 4:57:34 PM

Index

A Ability referential structure, 46 ABO. See Actor-Behavior-Object Abstract intelligence, 97 ACT. See Affect control theory Acting. See Deep acting; Surface acting Activation, 68 Activities, 42 Actor-Behavior-Object (ABO), 99 Affect binding negotiated exchange and, 161 cooperation and, 158f definition of, 143 duration of, 135 emotion and, 43, 135, 182 measurement of, 154 non-binding negotiated exchange and, 158–159 reciprocity and, 158 trust and, 149, 162–163, 184 uncertainty and, 158–162 Affect control theory (ACT), 4, 68–70, 257 consistency and, 98–99 criminal behavior and, 70 dictionaries of, 101 meanings and, 101 models of, 101 negative emotion and, 69 positive emotion and, 69 sentiments and, 68 social intelligence and, 98–102 Affect theory of social exchange, 57, 143 Affective consequences, emotion and, 134 Affective regard definition of, 172 exchange and, 168 generalized exchange and, 175 negotiated direct exchange and, 175 reciprocal direct exchange and, 175 reciprocity and, 179 African-Americans inequality and, 354 leadership and, 22 Age, status and, 23

Agent, 95, 95n, 1 Aggression, 350–351 Alienation, 1, 324 distal bias with, 335 sadness and, 334 Americans, trust with, 44 Anger, 350–351 assertion and, 334 depression and, 336 elaborations on, 323t fairness and, 186, 196 guilt and, 334 humiliation and, 339 identity and, 81, 86–87 injustice and, 204, 206 legitimacy and, 207 measurement of, 84 as moral emotion, 231–232 moral identity and, 243 power and, 207 as primary emotion, 143, 182 procedural justice and, 221 sadness and, 336 shame and, 334, 339 status and, 3, 248, 290 Anticipatory emotions, 14 Anxiety, 84 Artificial intelligence, of virtual actors, 96–97 Assertion, anger and, 334 Assertiveness, 349–350 Assurance, 172 Attribution, 149–151, 339–340 as defense mechanism, 326–327 measurement of, 155 reciprocity and, 160 Authenticity, 271 Autonomic nervous system, emotion and, 135–136 Avatar, 95, 95n, 1 Aversion. See Fear

B Bad faith, 256 Badness, vs. goodness, 68

357

IDX-P374095.indd 357

5/26/2008 4:59:23 PM

358 Bias negative, 326, 333, 335 proximal, 327 Bilateral exchange, 171 conflict and, 180 Binding negotiated exchange, 146, 152 affect and, 161 Bureaucracies, 1 Bush, George H.W., 22 Bush, George W., 22 Bywater, Bill, 29–30

index reciprocity and, 176 trust and, 156f, 157f Core emotions, 136 Corporate units, 320 Credibility, 347–348 Criminal behavior, ACT and, 70 Culture emotion, 3 emotional labor and, 259–282 vs. social structure, 279 Cybernetic control model of perception, 68

C Campbell, Kim, 22 Categoric units, 320 Categorical referential structures, 46–47 Chain-generalized exchange, 170 Charity, 241 empathy and, 245 Class, inequalities in, 39 Clinton, Bill, 22 Clinton, Hillary, 22 Coercion, 288–289 Cognition emotion and, 137–138 risk and, 183–184 social change and, 309–311 trust and, 183–184 Cohesion, 42 emotion and, 168 humor and, 27 Collective effervescence, 1 Commitment, 43–44 in exchange, 144–146 identity theory and, 103–105 risk and, 44 trust and, 144 Competency emotion and, 116–117 positive emotion and, 119, 122 power and, 41 Compliance, 257 power and, 13 Concentration, social face and, 108 Confidence, social face and, 108 Conflict bilateral exchange and, 180 exchange and, 145, 185 fairness and, 184 salience of, 179–180, 184–185 Conflict theory, 38 Conformity, emotional expression and, 104 Connected networks, 151, 151n, 7 Consistency, ACT and, 98–99 Context. See Social context Cooperation affect and, 158f in exchange, 153–154 negotiation and, 180

IDX-P374095.indd 358

D Das Kapital (Marx), 263 Death, 3 Decision-making emotional intelligence and, 98 fairness in, 207–208 social exchange and, 134, 186 Deep acting, 14, 256 emotional labor, 269 performance of, 277–279 social context and, 270 Defense mechanism, 325–326 attribution as, 326–327 Deference leadership and, 23 status and, 13 Deflection, 69 Denial, 326 Department of Social Services (DSS), 347 Dependency, 6, 133 Depression anger and, 336 education and, 87 fairness and, 196 fear and, 336 measurement of, 84–85 occupation and, 88 sadness and, 84 status and, 3, 86–87 Differentiation, integration and, 171–172 Diffuse status characteristic, 46 Direct exchange, 146–148, 169 vs. indirect exchange, 170–171 negotiated, 169, 175 reciprocal, 170, 175 Disagreement, status and, 15 Disappointment. See Sadness Disgust, as primary emotion, 143 Displacement, 326 Display rules, 136 Dispositional attribution, 149–151 Distal bias with alienation, 335 of negative emotion, 326, 333 Distress, 84 identity and, 88

5/26/2008 4:59:23 PM

Index injustice and, 206 measurement of, 85 Distributive justice. See Fairness Dole, Bob, 22 Durkheim, Émile, 134

E Education, 272 depression and, 87 love and, 88 status and, 80 Embarrassment, 68 identity and, 81 Emotion, 42. See also Moral emotion; Negative emotion; Positive emotion; Primary emotion affect and, 43, 135, 182 affective consequences and, 134 anticipatory, 14 arousal of, 136–137 autonomic nervous system and, 135–136 classification of, 321–324 cognition and, 137–138 cohesion and, 168 competency and, 116–117 components of, 142–143 core, 136 definition of, 142–143 as dependent variable, 138 duration of, 135 emotional labor and, 269–271 equity model and, 195–196 in exchange, 133–189 fairness and, 206–207 gender and, 127, 222 identity and, 67–128 impression management and, 71 as independent variable, 138 inequality and, 343–354 injustice and, 198–199, 203–223 integrative bonds and, 167–189 intensity of, 182 moral character and, 71, 119 motivation from, 311–314 need for, 187–189 non-binding negotiated exchange and, 161–162 norms and, 255–256 procedures and, 197 reciprocity and, 167–189 rules and, 275 secondary, 142–143, 182n, 4, 324 situational, 14 social, 135–136 social change and, 309–354 social exchange theory and, 43 social face and, 108 social intelligence and, 97–98 social movements and, 309–311 in social order, 16–18

IDX-P374095.indd 359

359 social structure and, 136–138, 136f, 319–341 sociology of, 2–4 solidarity and, 168 status and, 71 structural, 14, 15 trust and, 184 valence of, 182 values and, 275 work and, 107–128 Emotion cultures, 3 Emotion management, 207–208, 263–266 emotional labor and, 288–289 hierarchy and, 288 with others, 294–298, 302 Emotion spiral EPA and, 113 measurement of, 114f Emotion work, 15, 136, 255–256, 270 others and, 287–303 Emotional deviance, 3 Emotional expression, 70, 212–214 conformity and, 104 expectations and, 104 measurement of, 216 negative emotions and, 293n, 4 social order and, 45 status and, 289–290 Emotional intelligence, 97 decision-making and, 98 Emotional labor, 14, 255–303 culture and, 259–282 deep acting, 269 emotion and, 269–271 emotion management and, 288–289 gender and, 280 with nurses, 259–282 social context and, 271–273 social structure and, 259–282 surface acting, 269 Emotional processes, 134 Empathy charity and, 241, 245 guilt and, 233 as moral emotion, 231–232 Encounters, 320 Endorsement, 213–214, 215–216 EPA. See Evaluation, Potency, and Activity Equity model, 195–196 Evaluation, Potency, and Activity (EPA), 99–100 dictionaries of, 101 emotion spiral and, 113 men and, 115 women and, 115 Exchange, 37. See also Direct exchange; Indirect exchange; Negotiated direct exchange; Reciprocal direct exchange; Social exchange affective regard and, 168

5/26/2008 4:59:23 PM

360 Exchange (Con’d) bilateral, 171, 180 chain-generalized, 170 commitment in, 144–146 conflict and, 145, 185 cooperation in, 153–154 emotion in, 133–189 fairness and, 145 forms of, 146–148, 169–171, 175–176 generalized, 170, 175 indirect, 146–148 integrative outcomes of, 171–173 interactional fairness and, 168 negative emotion and, 185 negotiation in, 145–147 power and, 145 reciprocity in, 145, 147, 153, 168 relational solidarity and, 168 risk and, 145 status in, 54f trust and, 150, 162t, 168 uncertainty in, 144–146, 147, 149 unilateral, 171, 179 Exchange theory tradition, 5, 15 Excitement, status and, 51 Exclusive humor, 31 Expectation states theory, 15 order in, 45–47 shortcomings of, 53 Expectations, 179 emotional expression and, 104 performance, 45–46 positive emotions and, 325 reward expectations states theory, 5, 46 shame and, 337 shared, 99 Exploitation, 148 trust and, 178 Expression rules, 256 Expressive value of indirect reciprocity, 179 of reciprocity, 178–179

F Face. See Social face Face-to-face settings, vs. virtual environments, 96, 104–105 Fair process effect, 207 Fairness, 142, 185–186, 195–249. See also Injustice; Procedural justice anger and, 186, 196 conflict and, 184 in decision-making, 207–208 depression and, 196 emotion and, 206–207 exchange and, 145 happiness and, 196 inequality and, 186

IDX-P374095.indd 360

index interactional, 168, 172–173 power and, 145 procedures and, 196–197 Fear, 324 depression and, 336 elaborations on, 323t guilt and, 334 as primary emotion, 143 risk and, 184n, 5 sadness and, 336 shame and, 334 Feeling rules, 14, 207, 256 negative, 271 positive, 271 social change and, 310 First-order elaborations, 322 Frustration, procedural justice and, 221–222

G Gandhi, Indira, 22 Gender. See also Men; Women emotion and, 127, 222 emotional labor and, 280 moral emotion and, 228, 247 procedural justice and, 222 race and, 39n, 1 social face and, 119 status and, 23 General Social Survey (GSS), 287 Generalized exchange, 170 affective regard and, 175 trust and, 175 Gifts, 44 reciprocity and, 159–160 Good faith, 256 Goodness, 247 vs. badness, 68 moral identity and, 230 Grief, 3 Group processes research, 5–7, 15–16 Group-value model, 197 GSS. See General Social Survey Guilt, 246–247, 324 anger and, 334 charity and, 241 empathy and, 233 fear and, 334 injustice and, 206 as moral emotion, 232–233 over-reward and, 196 status and, 3–4, 235

H Happiness, 324 elaborations on, 323t fairness and, 196 over-reward and, 196

5/26/2008 4:59:23 PM

361

Index as primary emotion, 143 status and, 4 Hatred, 338 Hierarchy emotion management and, 288 of identity, 102–103 others in, 290–291 of status, 6, 28, 29 Homo moralis, 199 Homo socialis, 199 Humiliation, anger and, 339 Humor of Clinton, B., 22 cohesion and, 27 inclusive, 31 influence and, 26 leadership and, 21–34 masculinity and, 26 motivation and, 27–30 negative, 32 non-traditional leaders and, 21 positive, 26–27, 32 self-deprecating, 33–34 shared, 29 solidarity and, 27–29 status and, 17, 26, 30 status hierarchy and, 28, 29 stereotypes and, 25–26 stress and, 27 traditional leaders and, 21 women and, 25–27, 30–31

I Identity anger and, 81, 86–87 distress and, 88 embarrassment and, 81 emotion and, 67–128 hierarchy of, 102–103 identity control theory and, 70–71 love and, 90 meanings and, 77–78 model of, 77f multiple roles in, 76, 231 negative emotion and, 76, 89–90 resources and, 71, 76, 78–80, 89 self and, 102 shame and, 81 standards, 78 status and, 75–91 stress and, 103 verification of, 78, 80–81 Identity control theory, 4, 68 identity and, 70–71 Identity standard, 69 Identity theory, 67, 102–103 commitment and, 103–105

IDX-P374095.indd 361

power and, 58 status and, 58 structural symbolic interaction framework and, 75 Ideologies, 229 Impression management, 71 Impression-formation equations, 100–101 Impropriety, social change and, 56 Incentives, 43 Inclusive humor, 31 Income, moral emotion and, 228, 247 Indirect exchange, 146–148, 169 vs. direct exchange, 170–171 Indirect reciprocity, expressive value of, 179 Inequality African-Americans and, 354 emotion and, 343–354 fairness and, 186 legitimacy and, 45, 56–57 performance expectations and, 45–46 status value theory of power and, 55 Influence, 23 humor and, 26 Injustice, 5 anger and, 204, 206 distress and, 206 emotion and, 198–199, 203–223 guilt and, 206 toward others, 203–223 perceptions of, 205n, 1 resentment and, 204 satisfaction and, 206 Institutional domains, 320 Integration, differentiation and, 171–172 Integrative bonds, 135, 136f emotion and, 167–189 forms of, 175–176 negotiation and, 175 reciprocity and, 175, 176 structure of, 167–189 Integrative outcomes of exchange, 171–173 of reciprocity, 181–182 Intelligence. See also Social intelligence abstract, 97 artificial, 96–97 emotional, 98 mechanical, 97 Interaction, 42 Interactional fairness, 172–173 exchange and, 168 Interpersonal ties, 210–211, 215 negative emotion and, 222

J Japanese, trust with, 44 Job autonomy, 272

5/26/2008 4:59:23 PM

362 Johnson-Sirleaf, Ellen, 22 Joint tasks, 57 Joy. See Happiness Justice. See Fairness

K Kennedy, John, 22–23 King, Martin Luther, 22–23

L Leadership African-Americans and, 22 choosing of, 81 deference and, 23 humor and, 21–34 legitimacy and, 17 motivation and, 28 negative humor and, 32 non-traditional, 17, 21 positive humor and, 32 power and, 28 solidarity and, 28 status and, 28 traditional, 17, 21 transformational, 27 women and, 22, 24–25 Legitimacy, 6, 204–205, 211–212 anger and, 207 inequality and, 45, 56–57 leadership and, 17 resentment and, 207 satisfaction and, 207 Liking, 182 Love, 84 education and, 88 identity and, 90 measurement of, 85 occupation and, 88 race and, 88 spousal identity and, 86

M The Managed Heart (Hochschild), 2, 14, 263 Manipulation, 289 Marital partners, 5 Marx, Groucho, 31 Marx, Karl, 38, 263 Masculinity, humor and, 26 Maternity care, 343–354 Meanings, 230 ACT and, 101 equilibrium in, 257 identity and, 77–78 of moral identity, 239 resources and, 78 Mechanical intelligence, 97

IDX-P374095.indd 362

index Meir, Golda, 22 Men EPA and, 115 in minorities, 22–23 as nurses, 280 Microorder, 42–48 Minorities. See also African-Americans; Race men in, 22–23 Mixed-motive structures, 179 Mood, 114–116 Moral character emotion and, 71, 119 negative emotion and, 125 social face and, 108, 113 Moral emotion, 195–249 anger as, 231–232 empathy as, 231–232 gender and, 228, 247 guilt as, 232–233 income and, 228, 247 race and, 228, 247 shame as, 232–233 social structure and, 234–235, 238 status and, 227–249 Moral identity, 227–249 anger and, 243 goodness and, 230 meanings of, 239 morality and, 228–231 norms and, 233–234 salience of, 240 shame and, 243 Moral superiority, 207 Morality, moral identity and, 228–231 Motivation from emotion, 311–314 humor and, 27–30 leadership and, 28

N NAFTA. See North American Free Trade Agreement National Advanced Driving Simulator (NADS), 96 Negative emotion, 15 ACT and, 69 arousal of, 324–325 distal bias of, 326, 333 emotional expression and, 293n, 4 exchange and, 185 identity and, 76, 89–90 interpersonal ties and, 222 moral character and, 125 sanctions and, 325 self-presentation and, 68 social change and, 313 social face and, 121 social structure and, 330–331t, 333–340 spousal identity and, 86

5/26/2008 4:59:23 PM

363

Index status and, 125 work and, 212–213 Negative feeling rules, 271 Negative humor, leadership and, 32 Negotiated direct exchange, 169. See also Binding negotiated exchange; Non-binding negotiated exchange affective regard and, 175 relational solidarity and, 175 trust and, 175 Negotiation cooperation and, 180 in exchange, 145, 146–147 integrative bonds and, 175 Network exchange, 134 Network exchange theory, 5–6, 41, 138 Neutral behaviors, 127 Non-binding negotiated exchange, 146–147, 153 affect and, 158–159 emotion and, 161–162 Nonreciprocity, risk of, 176–178, 183–184 Non-traditional leaders, humor and, 17, 21 Normative order, 227–249 Norms, 228 emotion and, 255–256 moral identity and, 233–234 North American Free Trade Agreement (NAFTA), 29 Nurses emotional labor with, 259–282 men as, 280

O Occupation depression and, 88 love and, 88 status and, 80 Opportunism, 148 Order. See also Social order in expectation states theory, 45–47 microorder, 42–48 normative, 227–249 in social exchange theory, 43–45 Order producing effects, 6 Organization, of social exchange, 134–135 Organizational face theory, 107–128 Others emotion management with, 294–298, 302 emotion work and, 287–303 in hierarchy, 290–291 injustice toward, 203–223 Over-reward, 196

P Partners, 148 marital, 5 trust and, 150, 155–158 uncertainty and, 148 Perceived trustworthiness. See Trust

IDX-P374095.indd 363

Performance expectations inequalities and, 45–46 status characteristics theory and, 46 Pleasure as primary emotion, 182 status and, 51 Positive emotion ACT and, 69 arousal of, 324–325 competency and, 119, 122 expectations and, 325 proximal bias of, 327 social face and, 121 social structure and, 327–333, 329–330t Positive feeling rules, 271 Positive humor leadership and, 32 women and, 32 Potency, 68 Power, 3, 13–59, 133, 145, 169 anger and, 207 competency and, 41 compliance and, 13 definition of, 23, 41 early work on, 13–15 exchange and, 145 fairness and, 145 in group process research, 15–16 identity theory and, 58 inequalities in, 39 leadership and, 28 microorder and, 42–48 reciprocity and, 180–181 relational cohesion theory and, 45 resentment and, 207 satisfaction and, 207 social exchange theory on, 41 in social order, 16–18 status and, 17–18, 41, 256 status value theory of, 39, 40, 55 theories of, 41–42 Power-dependence theory, 5 Pride, social face and, 108 Primary emotion, 142–143, 143n, 1, 321–324 anger as, 182 pleasure as, 182 Procedural justice, 215 anger and, 221 frustration and, 221–222 gender and, 222 resentment and, 221–222 satisfaction and, 213, 221 self-esteem and, 207 Procedures emotion and, 197 fairness and, 196–197 Projection, 326 Propriety, 56

5/26/2008 4:59:23 PM

364 Proximal bias, of positive emotion, 327 Psychodynamics, 327–340

R Race gender and, 39n, 1 love and, 88 moral emotion and, 228, 247 status and, 23 Reaction formation, 326 Reagan, Ronald, 22 Reason-unreason dichotomies, 310 Reciprocal direct exchange, 170 affective regard and, 175 relational solidarity and, 175 trust and, 175 Reciprocity affect and, 158 affective regard and, 179 attribution and, 160 causal model for, 176–181, 177f cooperation and, 176 emotion and, 167–189 in exchange, 145, 147, 153, 168 expressive value of, 178–179 gifts and, 159–160 integrative bonds and, 175, 176 integrative outcomes of, 181–182 power and, 180–181 in social exchange theory, 135 structure of, 167–189 trust and, 141–164, 179 uncertainty and, 160 unilateral exchange and, 179 Referential structures, 46 Reich, Robert, 29–30 Relational cohesion theory, 44 power and, 45 Relational solidarity, 172 exchange and, 168 negotiated direct exchange and, 175 reciprocal direct exchange and, 175 Repression, 1, 326 of shame, 338 social change and, 311–314 Resentment injustice and, 204 legitimacy and, 207 power and, 207 procedural justice and, 221–222 Resources identity and, 71, 76, 78–80, 89 meanings and, 78 status and, 71, 80, 84, 89 Respect, 182 Revenge, 336 Reward expectations states theory, 5, 46 Richards, Ann, 22 Righteousness, 336

IDX-P374095.indd 364

index Risk, 145 cognition and, 183–184 commitment and, 44 exchange and, 145 fear and, 184n, 5 of nonreciprocity, 176–178, 183–184 trust and, 183–184 Role theory, 67 Roles social, 99 spousal, 82–85 Rubin Love Scale, 85 Rules, 275

S Sacrifice, 42 Sadness, 324 alienation and, 334 anger and, 336 depression and, 84 disappointment and, 334 elaborations on, 323t fear and, 336 as primary emotion, 143 status and, 3 Salience, 102–103 of conflict, 179–180, 184–185 of moral identity, 240 Sanctions, 324–325 negative emotions and, 325 Sarcasm, 22 Satisfaction elaborations on, 323t injustice and, 206 legitimacy and, 207 power and, 207 procedural justice and, 213, 221 social face and, 108 status and, 51 Secondary emotions, 142–143, 182n, 4, 324 SEI index, 84 Self, identity and, 102 Self-deprecating humor, 33–34 Self-esteem, fairness and, 207 Self-handicapping, 33–34 Self-presentation, 5 negative emotion and, 68 Self-verification, 77–78 Sentiments, 42, 143n, 1 ACT and, 68 Shame, 246, 324 anger and, 334, 339 charity and, 241 expectations and, 337 fear and, 334 identity and, 81 as moral emotion, 232–233 moral identity and, 243 repression of, 338

5/26/2008 4:59:23 PM

Index social face and, 108 status and, 235, 248 Shared expectations, social roles and, 99 Signs, 76n, 1 Situational attribution, 149–151 Situational emotions, 14 Social capital, 172 Social change cognition and, 309–311 emotion and, 309–354 feeling rules and, 310 impropriety and, 56 negative emotion and, 313 repression and, 311–314 social reproduction and, 260–261 Social context, 266–267 deep acting and, 270 definition of, 266n, 4 emotional labor and, 271–273 surface acting and, 270 Social emotions, 135–136 Social exchange decision-making and, 134, 186 organization of, 134–135 status-stratified systems and, 37–59 Social exchange theory, 133 emotion and, 43 order in, 43–45 on power, 41 reciprocity in, 135 shortcomings of, 52–53 trust and, 43, 135 Social face, 107–128 concentration and, 108 confidence and, 108 emotion and, 108 gender and, 119 moral character and, 108, 113 negative emotion and, 121 positive emotion and, 121 pride and, 108 satisfaction and, 108 shame and, 108 status and, 108, 113 Social identity theory, 67 Social intelligence ACT and, 98–102 definition of, 97 elements of, 97–98 emotion and, 97–98 virtual environments and, 95–105 Social interactional theory, 13, 15 Social movements, 309–311 Social order emotion in, 16–18 emotional expression and, 45 power in, 16–18 status in, 16–18 Social place, 288

IDX-P374095.indd 365

365 Social reality, 321f Social reproduction, social change and, 260–261 Social roles, shared expectations and, 99 Social structure, 79–80 vs. culture, 279 emotion and, 136–138, 136f, 319–341 emotional labor and, 259–282 moral emotions and, 228, 234–235 negative emotion and, 330–331t, 333–340 positive emotion and, 327–333, 329–330t relations in, 133–134 Solidarity emotion and, 168 humor and, 27–29 leadership and, 28 Spousal identity, 86 Spousal role, 82–85 Standards, 230 Status, 13–59 age and, 23 anger and, 3, 248, 290 deference and, 13 definition of, 41 depression and, 3, 86–87 disagreement and, 15 early work on, 13–15 education and, 80 emotion and, 71 emotional expression and, 289–290 in exchange systems, 54f excitement and, 51 gender and, 23 in group processes research, 15–16 guilt and, 3–4, 235 happiness and, 4 humor and, 17, 26, 30 identity and, 75–91 identity theory and, 58 inequalities in, 39 interest and, 51 leadership and, 28 microorder and, 42–48 moral emotions and, 227–249 negative emotion and, 125 occupation and, 80 pleasure and, 51 power and, 17–18, 41, 256 race and, 23 resources and, 71, 80, 84, 89 sadness and, 3 satisfaction and, 51 shame and, 235, 248 social face and, 108, 113 in social order, 16–18 spousal identity and, 86 theories of, 41–42 Status characteristics theory, 5 performance expectations and, 46 shortcomings of, 53

5/26/2008 4:59:24 PM

366 Status construction theory, 5, 40, 80 Status hierarchy, 6 humor and, 28, 29 Status legitimation theory, 5 Status processes, 37 Status value, 42 Status value theory of power, 39, 40 inequalities and, 55 Status-stratified systems, social exchange and, 37–59 Stereotypes, humor and, 25–26 Stratification systems, 320 Stress humor and, 27 identity and, 103 Structural emotions, 14, 15 Structural social psychology, 39–40 Structural symbolic interaction, 4–5 Structural symbolic interaction framework, 75 Sublimation, 326 Suppression, 270 Surface acting, 14, 256 emotional labor, 269 performance of, 277–279 social context and, 270 strategies of, 270 Symbolic interaction, 67 Symbols, 76n, 1 Systems of societies, 320

T Task groups, 15, 29 Thatcher, Margaret, 22 Theory of relational cohesion, 42 Traditional leaders, 17 humor and, 21 Transformational leaders, 27 Trust, 149, 353 affect and, 149, 162–163, 184 with Americans, 44 cognition and, 183–184 commitment and, 144 cooperation and, 156f, 157f definition of, 172 emotion and, 184 exchange and, 150, 162t, 168 exploitation and, 178 generalized exchange and, 175 with Japanese, 44 as long-term emotional disposition, 182 measurement of, 154–155 negotiated direct exchange and, 175

IDX-P374095.indd 366

index partners and, 150, 155–158 reciprocal direct exchange and, 175 reciprocity and, 141–164, 179 risk and, 183–184 social exchange theory and, 43, 135 uncertainty and, 144–145, 147–148, 150, 155

U Uncertainty, 43–44 affect and, 158–162 in exchange, 144–147, 149 partners and, 148 reciprocity and, 160 trust and, 144–145, 147–148, 150, 155 unilateral exchange and, 179 Unilateral exchange, 171 reciprocity and, 179 uncertainty and, 179

V Validity, 56 Value congruence, 271 Values, 229 emotion and, 275 Virtual actors, 95n, 1 artificial intelligence of, 96–97 creation of, 104 Virtual environments vs. face-to-face settings, 96, 104–105 social intelligence and, 95–105 Virtual Soldier Project, 96

W Weber, Max, 134 White males, vs. female African-Americans, 38, 50 Whole societies, 320 Women African-American, 38, 50 EPA and, 115 humor and, 25–27, 30–31 leadership and, 22, 24–25 positive humor and, 32 Work. See also Emotion work emotion and, 107–128 negative emotions and, 212–213 Work-identity discrepancy, 272 Workload, 272

Y Yerkes-Dodson Law, 137

5/26/2008 4:59:24 PM