Meta-Emotion

How Families Communicate Emotionally

John Mordechai Gottman, Lynn Fainsilber Katz, Carole Hooven

• Publisher: Routledge
• Year: 2013
• ISBN:
  9781134795970
• Edition: 1
• Topic: Psychologie
• Subtopic: Psychologie sociale

This book describes research on the emotional communication between parents and children and its effect on the children's emotional development. Inspired by the work, and dedicated to the memory of Dr. Haim Ginott, it presents the results of initial exploratory work with meta-emotion--feelings about feelings. The initial study of meta-emotion generated some theory and made it possible to propose a research agenda. Clearly replication is necessary, and experiments are needed to test the path analytic models which have been developed from the authors' correlational data. The authors hope that other researchers will find these ideas interesting and stimulating, and will inspire investigation in this exciting new area of a family's emotional life.

Part I

The Emotional Life of Families

Introduction to the Concept of Meta-Emotion

This chapter introduces our concepts of “ meta-emotion” and “meta-emotion structure,” and provides some historical context for the evolution of these ideas in our laboratory.

History and Preface

For the past 20 years, our laboratory has had two lines of research. We have been studying children’s friendships and children’s peer relations since 1972 (e.g., Gottman, 1983; Gottman & Parker, 1986). We have also been studying the social interaction processes related to marital satisfaction (Gottman, 1979), to longitudinal change in marital satisfaction (Gottman & Krokoff, 1989), and, more recently, to marital dissolution (Buehlman, Gottman, & Katz, 1992; Gottman, 1993; Gottman, 1994; Gottman & Levenson, 1992). Much of this latter work emerged from a continuing collaboration with Robert Levenson that began in 1979, when Gottman and Levenson decided to study the role of emotion in marriage using methods of social psychophysiology. These methods proved to be quite productive (e.g., Gottman, 1994; Gottman & Levenson, 1985, 1988, 1992; Levenson & Gottman, 1983, 1985).

In 1984, Gottman decided to try to bring his work with Levenson on marital psychophysiology and his research on children’s friendships and children’s social development together into a single series of studies. The idea was to build a new family psychophysiology laboratory, to study the linkages between the marital, parent-child, and child-peer systems, extending the work that Levenson and Gottman had employed in studying marriage. To accomplish this task, it was necessary for Gottman to become a psychophysiologist and also to learn about parent-child relationships. It also required designing and building a new laboratory, which could not have happened without the critical assistance, guidance, and technical help of Robert Levenson. Lynn Fainsilber Katz, then a graduate student at the University of Illinois, joined the project at its inception. Gottman and Katz wrote the grants that eventually funded the work. In 1985, Levenson came to the University of Illinois and put together a family psychophysiology laboratory. Levenson guided Gottman in this first entry and later development as a psychophysiologist, a process that took about 5 years. Gottman’s training was assisted by his NIMH Research Career Development Award, by informal instruction by Stephen Porges, by membership in the Society for Psychophysiological Research, and by formal instruction in John Cacioppo’s National Science Foundation’s training program in psychophysiology during the summer of 1987. The Family Psychophysiology Laboratory was completed in 1985 and our first study began in 1986.

For the past decade, our laboratory has been studying the relationship between the parents’ marriage, parent-child interaction, and the emotional-social development of children. We have suggested that these linkages are mediated through the development of emotion-regulation abilities in the child (Gottman & Katz, 1989; Katz & Gottman, 1991). By emotion-regulation abilities we had a specific set of abilities in mind, namely the child’s ability to inhibit inappropriate negativity, the child’s ability to self-soothe, and the child’s ability to focus attention in the service of an external goal. Steve Porges was then at the University of Illinois, and his thinking about regulation in the cardiovascular system was an important influence. In fact, Porges and Gottman team taught a course on time-series analysis to graduate students that introduced them to the measurement of vagal tone by spectral time-series techniques.

The area of parent-child relations was new to us. As fledglings in this area, we decided to build on Philip and Carolyn Cowans’ balanced approach to parenting, much of which had been guided by the work of Diana Baumrind. This collaboration was made possible by the NIMH Family Research Consortium, which was set up in 1980 by Joy Schulterbrandt. In selecting a theoretical basis for our studies, building on the work of the Cowans and Baumrind, we decided to focus on the emotional life of families and the development of emotional regulation in children. The focus on emotional regulation was inspired by Eleanor Maccoby’s (1980) seminal book on social-emotional development.

This book presents the results of our initial work with meta-emotion. We are excited by these results, but we offer this work only as our initial, exploratory work with meta-emotion. Clearly replication is necessary, and experiments are needed to test the path analytic models we have developed from our correlational data. This book explores our initial study, which includes meta-emotion, generates some theory, and in Chapter 18 proposes a research agenda. We think that the research agenda is as important as our initial findings. We hope that other researchers will find these ideas interesting and stimulating, and we offer them in the hopes that others will join us in investigating this exciting new area of a family’s emotional life.

Much of our laboratory’s initial work was designed to parallel the work of Carolyn and Philip Cowan. In the design of our studies, we added to the Cowans’ approach a social psychophysiological emphasis, an emphasis on the parents’ marriage, an emphasis on emotion regulation in children, and an emphasis on children’s relationships with other children.

Before the first study in 1984 was launched, Gottman visited Robert Leven-son, who was then on sabbatical in Paul Ekman’s laboratory, and Ekman introduced him to Hochschild’s (1983) book, The Managed Heart. This book inspired Gottman to think of the idea of meta-emotion, and to the development of a “meta-emotion interview” in conjunction with Lynn Fainsilber Katz (Katz and Gottman, 1986).

In this first study we planned for our family psychophysiology laboratory, each parent was to be separately interviewed about their own experience of sadness and anger, their philosophy of emotional expression and control, and their attitudes and behavior about their children’s anger and sadness. Their behavior during this interview was audiotaped and later coded with a meta-emotion coding system designed by a central member of our laboratory staff, Carole Hooven, a coauthor of this book. Initially, the goal of the meta-emotion interview was to examine each parent’s feelings about being emotionally expressive, but this idea was later expanded.

In pilot work for our first study of the effects of the parents’ marriage on children, we discovered a great variety in the experiences, philosophies, and attitudes that parents had about their emotions and the emotions of their children. One pair of parents said that they viewed anger as “from the devil,” and that they would not permit themselves or their children to express anger. Their child was quite docile in her interactions with her parents but appeared quite angry and bossy in her interactions with her best friend. A similar negative view toward anger was echoed by other parents. Some parents in our study said that they put their children in Time Out for being angry, even if there was no child misbehavior. Other parents felt that anger was natural, but ignored the experience of anger in their children. Other parents encouraged the expression and exploration of anger. There was similar variety with respect to sadness, and the information we gathered about sadness was not redundant with the information we gathered about anger. Some parents minimized sadness in themselves and in their children, saying such things as, “I can’t afford to be sad,” and “What does a kid have to be sad about?” Other parents thought that emotions like sadness in themselves and in their children were important and viewed themselves as emotion coaches of their children about the world of emotion. In our pilot work there also appeared to be gender differences: Fathers seemed less likely to be aware of their own sadness or to assist when their children were sad; fathers who were oriented toward emotion seemed more interested in their children’s anger than in their sadness. Mothers seemed to be more concerned with their children’s sadness than fathers. These were our initial impressions.

We decided to focus on studying parents’ feelings about feelings, which we called their meta-emotions. The notion of meta-emotion we had in mind paralleled the area of metacognition, which referred to the executive functions of cognition (Allen & Armour, 1993; Bvinelli, 1993; Flavell, 1979; Fodor, 1992; Nelson, 1992; Olson & Astington, 1993). Metacognitions are thoughts about thinking. In social interaction research, the term “metacommunication” referred to communication about communication. We began to also use the term meta-emotion structure to refer broadly to similar executive functions of emotion, ones that included concepts, philosophies, and metaphors about emotions, as well as emotions about emotions. Concepts similar to meta-emotion have been discussed by Salovey and Mayer (1990) in their ideas about emotional intelligence, and by Mayer and Gaschke (1988) in their state and trait meta-mood scales.

What we mean by the meta-emotion structure construct, specifically, is the parent’s awareness of specific emotions, their awareness of these emotions in their child, and their coaching of the emotion in their child. Emotion coaching refers to a set of processes that include such elements as talking to the child about the emotions, helping the child to verbally label the emotions being felt, accepting the child’s emotions, discussing the situations that elicited the emotions, and having goals and strategies for coping with these situations. In chapter 4, we delineate the five components of an emotion coaching meta-emotion structure. If this construct of emotion coaching is to be useful, it should relate to how the family actually functions, and our goal is to test this notion in this book. At that time we did not expect the concept of meta-emotion to become the centerpiece of our work. That result was entirely serendipitious. The data led us to modify our initial theorizing and to realize that we had a particularly appealing and parsimonious theory and set of initial findings. Once we made meta-emotion the centerpiece of our work, this became a data set that said yes to almost all the theoretical questions we asked of it in building our theoretical model. Lynn Fainsilber Katz’s thinking was instrumental to the development of the theoretical models. We also wish to acknowledge the help and feedback of Stephen Porges in building our theoretical model. He took a look at our initial results the same year he gave his spectacular presidential address on the vagus nerve and vagal tone to the Society for Psychophysiological Research, in Atlanta, in 1994.

Definitions

Meta-Emotion

By meta-emotion, we mean emotion about emotion. The idea is analogous to metacognitions, which are cognitions about cognition (see Nelson, 1992). Researchers have searched for uniformity with respect to emotion, thinking that, for example, the emotion of anger is the same for most people, but they have tended to overlook how people feel about a particular emotion such as anger. Some people are ashamed or upset about becoming angry, others feel good about their capacity to express anger, and still others think of anger as natural, neither good nor bad. Any experiment that involves the induction of emotion will also involve meta-emotions, even if they are not studied; hence, in emotion induction experiments, meta-emotions are not controlled. Thus, the effects of inducing the emotion of anger may vary across subjects, not only because anger is not induced uniformly across people, but because people’s emotions about their own anger vary so much. A good paradigm to take as an example is the startle experiment (Ekman, Friesen, & Simons, 1985; Landis & Hunt, 1939). In a startle experiment the experimenter will fire off a high intensity stimulus, like a gunshot, behind the subject’s head. After displaying a startle response, people will have various emotions to having been startled (Ekman, personal communication; Sutton & Levenson, personal communication.) Some laugh with pleasure, some laugh with embarrassment, some become afraid, some are disgusted, and some become angry. This latter emotional response is a meta- emotion: Quite simply, it is an emotion about an emotion. (We ignore here the issue of whether the startle is a reflex or an emotion.) This discussion is clearly not limited to the startle experiment. Whenever we elicit emotion, we are also dealing with emotions about having experienced or felt the emotion. We always engage the person’s metaemotion structure, whether we study it or not.

Meta-Emotion Structure

What do we mean by meta-emotion structure? By a meta-emotion structure we mean an organized and structured set of emotions and cognitions about the emotions, both one’s own emotions and the emotions of others. Thus, a meta-emotion structure is an organized set of feelings and concepts about emotion, and this idea includes the idea of an emotion philosophy. For example, one parent may be disgusted by his or her 5-year-old’s anger and believe that children of that age should not express anger, that it is destructive and bad. Another parent may view his or her 5-year-old’s anger as acceptable, and as an important moment for talking about the child’s emotions and understanding what the child is feeling.

We analyze people’s metaphors about emotion (Lakoff & Johnson, 1980) as one way of getting at people’s concepts about meta-emotion. We are only beginning this study. An example may help clarify the importance of this idea of meta-emotion metaphors. One father in our research referred to sadness as if it were a limited resource, like how much money one has. He thought that his children should not “waste” their sadness on trivial things but instead “invest” their sadness only in important things. For this father it was a sin for his child to be sad about missing mom one morning, because that was wasting sadness on a trivial, unimportant, everyday event. If his daughter’s pet had died, he would have thought it would be a wise “investment” of sadness to express sadness about that event, but merely missing mom was showing poor judgment on the child’s part on what to “spend” this valuable and limited resource on. This “limited resource” metaphor for sadness led this father to act disapproving some of the time, perhaps most of the time, when his daughter was sad, but other times to be compassionate and understanding of his daughter’s sadness. Would his daughter understand his concept, would she see only his disapproval, or would she see only his compassion at moments selected by him? At any rate, it seemed clear that his metaphor for sadness was likely to have a profound effect on his parenting and on his relationship with his daughter. In our research this metaphor is an example of a pattern of meta-emotions, and an organized set of thoughts about these patterns of meta-emotions, and we refer to each of these various patterns as a metaemotion structure.

1

Research on Parenting and Meta-Emotions

This chapter is a brief review of the literature on parenting research. We think that meta-emotion is not an independent dimension but, rather, it is contextualized within a network of parenting dimensions, and we suggest two parenting dimensions that will be employed in our model building.

Because we think that our meta-emotion variables are embedded in a set of variables that describe parenting, this chapter briefly reviews the parenting research literature. We do not attempt to provide a thorough review of this large literature. Rather, we offer this review with the goal of outlining the kinds of variables that have been considered to date. Our purpose, in part, is to convince the reader that there is something new in considering meta-emotion as part of parenting, despite the fact that we believe that meta-emotion variables do not stand alone but are contextualized in a web of variables that describe parenting.

Parenting has been classified and studied primarily in terms of the predominant parental affects toward the child and predominant parental discipline techniques. What we think is missing is how the parent feels about and relates to specific emotional displays by the child, and how this might relate to the parent’s feelings about his or her own emotions. This area of family functioning is likely to be more general than the parents’ predominant affect or the parents’ discipline techniques.

In early multidimensional parenting research that employed factor analysis, two major independent dimensions emerged. Maccoby and Martin (1983) pointed out that these variables emerged despite the initial theoretical bases of the investigations that derived them, rather than because of the theories; these orthogonal dimensions seem to have been born from the myriad of variables considered. They are: (a) a permissive/restrictive dimension, which refers to the amount of autonomy parents permit their children to have, and ...