Figure 3.1 Winner focuses on the goal, loser focuses on the winner. Final lap of relay race, which runner E is about to win.
EMOTIONAL ENERGY AND THE TRANSIENT EMOTIONS
EMOTION IS A central ingredient and outcome of IRs. It is time now to examine emotions more closely. Among other benefits of doing so is to highlight the contribution that sociology of emotions makes to macrosociological theory. And we shall see, via a circuitous route, the emotion-laden view of macro-sociological structure and hence of the place of individuals within it will give us some leads for a sociological theory of differences in personality.
Emotion implicitly occupies a crucial position in general sociological theory. As we attempt to make sociological concepts more precise and more empirically grounded, we find that many of the most important rest to a considerable extent upon emotional processes. Durkheim raised the central question of sociology: What holds society together? His answer is the mechanisms that produce moral solidarity; and these mechanisms, I have argued, do so by focusing, intensifying, and transforming emotions. Parsonian sociology, which took the most reified, agentless side of Durkheim, put the argument in equivalent terms: society is held together by values. But values, to the extent that they exist (and leaving open the issue of how far they are shared, and under what conditions), are cognitions infused with emotion. On the conflict side of sociological theory, Weber’s central concepts also imply emotion: the legitimacy that underlies stable power, the status group ranking by which stratification permeates everyday life, the religious worldviews that motivated some crucial periods of economic action. When we attempt to translate any of these concepts into observables, it is apparent that we are dealing with particular kinds of emotions. Marx and Engels are perhaps furthest away from theorizing about emotional processes: in their analysis, everything is structural (even alienation, which for Marx is an ontological relationship, not a psychological one). But it is apparent that in Marxian analyses of class mobilization and class conflict, emotion must play a part—whether it is the mutual distrust within fragmented classes that keeps them from mobilizing, or the solidarity that dominant classes have and that oppressed classes acquire only in revolutionary situations. In these respects, Marx and Engels’s conflict theory comes close to a dynamic version of Durkheim’s themes.
The sociology of emotions thus bears upon the central questions of sociology. What holds a society together—the “glue” of solidarity—and what mobilizes conflict—the energy of mobilized groups—are emotions; so is what operates to uphold stratification—hierarchical feelings, whether dominant, subservient, or resentful. If we can explain the conditions that cause people to feel these kinds of emotions, we will have a major part of a core sociological theory. There is, of course, a structural part of such a theory, and a cognitive part; but the emotional part gives us something essential for a realistic theory—its dynamics.1
These classic sociological theories implicitly concern emotions, but they do not usually refer to them explicitly. This is because our theories have a macro-primacy, or at least deal with social life at a level of considerable abstraction and aggregation. We are told of something called “legitimacy,” and of “values,” floating somewhere in a conceptual sky beyond the heads of real people in ordinary situations. If we attempt a micro-translation of sociology—not a micro-reduction, but a grounding of macro-concepts in real interactions across the macro-dimensions of time and space—we are led to see the importance of emotional processes. In other words, the micro-translation of macro-concepts yields emotion.
For the most part, this is not what most micro-theories have stressed. Mead and symbolic interactionism emphasize process, emergence, and cognition; Schutz and phenomenology emphasize routine and cognition; exchange theory emphasizes behaviors and payoffs; expectation states theory again stresses cognition. Emotion of course could be brought into these theories, but it is central to none of them.2 On the other hand, there is a burgeoning field of sociology of emotions, but until recently it has been largely treated as a specialized enclave, cut off from general issues of sociology.3 But several prominent versions of microsociology do not have to be pressed very far to yield the central micro-dynamics of emotion as a social process—a process that will serve to unpack the macro-sociological issues mentioned at the outset.
One of these is Garfinkel’s ethnomethodology. At first sight, it seems to be pitched on a different level. With its concern for the construction of mundane reality, and its heavy use of phenomenological abstractions, it seems to be essentially a cognitive theory. Cicourel (1973) even called his own version “Cognitive Sociology.” Nevertheless, I want to suggest that ethnomethodology reveals emotion at its core. Garfinkel’s most important contribution is to show that humans have intrinsically limited cognitive capabilities, and that they construct mundane social order by consistently using practices to avoid recognizing how arbitrarily social order is actually put together. We keep up conventions, not because we believe in them, but because we studiously avoid questioning them. Garfinkel demonstrated this most dramatically in his breaching experiments, in which he forced people into situations that caused them to recognize indexicality (i.e., that they rely on tacit acceptance of what things mean contextually) and reflexivity (that there are infinite regresses of justifying one’s interpretations). Interestingly enough, the reactions of his subjects were always intensely emotional. Usually it was an emotional outburst: becoming nervous and jittery, shaken, displaying anxiety and sometimes shock (Garfinkel 1967, 44, 221–26) Sometimes it was depression, bewilderment, or anger at having been put in a situation where they constructed a reality they later discovered to be false. In short, when people have to recognize that they are tacitly constructing their social worlds, and in an arbitrary and conventional way, rather than simply reacting to a world that is objectively there, they show intense negative emotions.
Garfinkel’s breaching experiments reveal something very much like Durkheim’s world. In this case, conventional social reality is a sacred object. Garfinkel’s experiments, violating the sacred object, call forth the same effects as violating a ritual taboo would have for a tribal member, desecrating the Bible for a Christian, or defaming the flag for a patriot. In Durkheim’s theory, moral sentiments attach to sacred objects. When they are violated, this positive sentiment of moral solidarity turns negative, into righteous anger directed against the culprit. Just so in Garfinkel’s experiments: there is outrage against the violator of everyday cognitive conventions. Garfinkel’s strategy parallels Durkheim’s: to show the conditions that uphold a social fact by revealing the opposition that occurs when it is broken. Durkheim used suicide and crime as means of highlighting the social solidarity that is their opposite; Garfinkel extended the method to reality-construction as a whole.
Ethnomethodology’s lack of explicit focus on emotions is misleading. One could well say that everyday life reality-construction is an emotional process, and that the emotions that uphold reality come forth in intense form when the social reality is broken. Furthermore, Garfinkel has shown that human cognition is limited; social order cannot be based on rational, conscious agreement. Durkheim (1893/1964) argued the same, but in the context of criticizing utilitarianism. If cognition does not hold society together, then, what does? Garfinkel tends to leave this on the level of cognitive practices (mostly borrowed from Schutz); but it is a peculiar form of cognition—cognitive practices for how to get by without too much cognition. Ethnomethodology seems to have a mysterious x-factor underlying social order, which the very notion of indexicality prohibits us from probing. But let us take the plunge: leave the cognitive plane, and recognize the x-factor as emotion.
Interaction ritual theory gives the most fine-grained picture of how emotions are transformed in the process of interaction: rituals begin with emotional ingredients (which may be emotions of all sorts); they intensify emotions into the shared excitement that Durkheim called “collective effervescence”; and they produce other sorts of emotions as outcomes (especially moral solidarity, but also sometimes aggressive emotions such as anger). This puts us in a position to use the flow of emotions across situations as the crucial item in the micro-to-micro linkage that concatenates into macro patterns. The most important of these patterns of IR chains is what from a macro viewpoint appears as stratification. Social order is produced on the micro level: that is to say, all over the map, in transient situations and local groups, which may well be stratified by class, race, gender, or otherwise divided against each other. Interaction ritual produces pockets of moral solidarity, but variably and discontinuously throughout a population. Now if we trace individual human bodies moving from one encounter to the next, we see that the history of their chains—what sociologists have conventionally referred to as their positions in the social structure—is carried along in emotions and emotion-laden cognitions that become the ingredients for the upcoming encounter. And then as the IR does its work, it intensifies, transforms, or diminishes those emotional ingredients so that those human bodies come out of the situation charged with emotional outcomes, which in turn set up what will happen in their next situations.4 In what follows, I will show that research on stratification gives us clues as to how emotional ingredients and outcomes are shaped. Stratification theory contributes to a theory of the distribution of varying emotions; and the microsociology of emotion contributes to the patterns of stratification.
DISRUPTIVE AND LONG-TERM EMOTIONS, OR DRAMATIC EMOTIONS AND EMOTIONAL ENERGY
A necessary first step is to widen our conception of emotion. Ordinary usage refers to emotions as experiences that are, for the most part, sudden and dramatic. “Don’t be so emotional” is advice predicated on this conception. The famous emotions are the most dramatic ones: fear, terror, anger, embarrassment, joy, and so forth. Some people and some cultures are regarded as too “unemotional” (as in the late-twentieth-century disparagement of “WASP” culture). But both Goffman and Garfinkel force us to see that there are also emotions that are undramatic; they are long-lasting, underlying tones or moods that permeate social life. Garfinkel’s mundane reality, for example, is characterized by the feeling—I stress that this is a feeling rather than an explicit cognition—that “nothing out of the ordinary is happening here.” This is an uninteresting emotion, from the point of view of the actor; but if Garfinkel is right, considerable work went into producing that feeling of ordinariness, and, into keeping ourselves from seeing that work itself. Mundane reality is a members’ accomplishment.
In Goffman and Durkheim, the ordinary-life, long-lasting feelings are more apparent. These theories stress solidarity, feelings of membership, and in Goffman’s case, feelings about one’s self. These are, if everything goes well, smoothly persistent sentiments; though in some important cases they may have an “up” feeling tone, or a “down,” depressed tone. Solidarity feelings, moral sentiment, the enthusiasm of pitching oneself into a situation, or being carried along by it, and, at the other end, depression, alienation, embarrassment—these are recognizably longerlasting kinds of emotions. Garfinkelian mundanity is merely a generic emotional quality at the middle of the plus-minus scale.
My aim is not to enter into terminological controversy. It would be useless for us to define emotions in such a way that we can talk only about the dramatic, disruptive emotions. Whatever we call them, we must also be able to talk about the long-term emotional tones, even the ones that are so calm and smooth as not to be noticed. In theoretical terms, it is the long-lasting ones (that I discuss as emotional energy, EE) that are of greatest importance. But I will also attempt to show that the dramatic, short-term emotions are best explained against the backdrop of the long-term emotions.
There are four emotions that virtually all researchers agree are found in all societies, and that may be considered the primary emotions (for a summary of research, see Turner 2002, 68–79). These four are anger, fear, happiness, and sadness / disappointment. Mammals share with humans the primary emotions of fear and anger / assertiveness. In humans, these emotions have their physiological base in the amygdala, an evolutionarily primitive part of the brain. Happiness, however, is not based in a particular part of the brain, but is spread out, not only in the primitive amygdala, but in the cortical and subcortical areas, which are evolutionarily later; that is to say, happiness is physiologically generalized, across the major regions of the brain including those involved in human symbolic functioning. Similarly for sadness, which has no distinctive brain location; it operates physiologically through the failure of neurotransmitters and in the flow of hormones in the endocrine system.
Happiness and sadness can be expressed in a number of terms: joy, elation, enthusiasm, effervescence—in contrast to disappointment, dreariness, and depression. These are related to the basic psycho-physiological pattern that I am calling high and low emotional energy. From the point of view of IR theory, it is not surprising that these two emotions lack a specific location in the brain. They are distinctively human blends of emotion and cognition, implicating the entire workings of the cognitive regions of the brain. High and low EE come from the entrainment of communicative gestures and emotional rhythms that are distinctive to human intersubjectivity; from an individual viewpoint, they are tightly woven together into the human self. Thus what from a narrower viewpoint may be considered an expression of joy—as a momentary emotional experience—is carried over as a long-term mood of emotional energy, of varying duration and degree of intensity. EE gives energy, not just for physical activity (such as the demonstrative outbursts at moments of acute joy), but above all for taking the initiative in social interaction, putting enthusiasm into it, taking the lead in setting the level of emotional entrainment. Similarly, sadness or depression is a motivational force when it is a long-term mood, reducing the level of activity, not only bringing physical listlessness and withdrawal (at its extreme, the avoidance of being awake), but making social interaction passive, foot-dragging, perfunctory.
Emotional energy, in IR theory, is carried across situations by symbols that have been charged up by emotional situations. Thus EE is a central part of the arousal of symbols that humans use to talk and to think with. Here again, the findings of physiological research bolster IR theory: “joy” in the narrower sense of short-run experience, high EE in the larger sense of long-term mood, is not a specific part of the brain firing but an overall activity of the brain’s cognitive and emotional functioning. Similarly, “sadness,” taken more broadly and in the longterm as low EE, is an overall decline in the functioning of the entire neuro-endocrinological system. To say that symbols are carried on EE is not merely a metaphor. The physiology buttresses the sociology.5
Interaction Ritual as Emotion Transformer
The basic model of ritual interaction (IR) is spelled out in chapter 2 as the mutual-focus / emotional-entrainment model. Let us review all the places that emotions occur in the model.
One initiating ingredient is that participants share a common mood. It is unessential which emotion is present at the outset. The feelings may be anger, friendliness, enthusiasm, fear, sorrow, or many others. This model posits an emotional contagion among the persons present: because they are focusing attention on the same thing and are aware of each other’s focus, they become caught up in each other’s emotions. As a result, the emotional mood becomes stronger and more dominant; competing feelings are driven out by the main group feeling. On the ultra-micro level, this happens by the process of rhythmic entrainment physiologically. That is to say, activities and emotions have their own micro-rhythm, a pace at which they take place. As the focus of interaction becomes progressively more attuned, the participants anticipate each other’s rhythms, and thus become caught up “in the swing of things.” Participants feel sadder in the course of a funeral, more humorous as part of a responsive audience at a comedy show, more convivial during the buildup of a party, more engrossed in a conversation as its rhythms become established. All these are versions of “collective effervescence”—even if that has a connotation of happy excitement, the more general condition is a high degree of absorption in emotional entrainment, whatever the emotion may be.
The outcome of a successful buildup of emotional coordination within an interaction ritual is to produce feelings of solidarity. The emotions that are ingredients of the IR are transient; the outcome however is a long-term emotion, the feelings of attachment to the group that was assembled at that time. Thus in the funeral ritual the shortterm emotion was sadness, but the main “ritual work” of the funeral was producing (or restoring) group solidarity. The emotional ingredients of a party may be friendliness or humor; the long-term result is the feeling of status group membership.
I refer to these long-term outcomes as “emotional energy” (EE). It is a continuum, ranging from a high end of confidence, enthusiasm, good self-feelings; down through a middle range of bland normalcy; and to a low end of depression, lack of initiative, and negative self-feelings. Emotional energy is like the psychological concept of “drive,” but it has a specifically social orientation. High emotional energy is a feeling of confidence and enthusiasm for social interaction. It is the personal side of having a great deal of Durkheimian ritual solidarity with a group. One gets pumped up with emotional strength from participating in the group’s interaction. This makes one not only an enthusiastic supporter of the group, but also a leading figure in it. One feels good with the group, and is able to be an energy-leader, a person who stirs up contagious feelings when the group is together.
At the low end of the emotional energy continuum, the opposite is the case. Low emotional energy is a lack of Durkheimian solidarity. One is not attracted to the group; one is drained or depressed by it; one wants to avoid it. One does not have a good self in the group. And one is not attached to the group’s purposes and symbols, but alienated from them.
This is not the way the term “emotion” is commonly used, and commonsense categories have difficulty in grasping that EE is emotion at all. Folk-categories usually point at emotions only when they are dramatic shifts, disruptions of the normal flow of social energy. We are particularly likely to overlook middle levels of EE, in which the flow of energy toward social situations allows everything to proceed normally and hence is taken for granted. But without this emotional energy flow, social interactions could not take place.
There are more differentiated variants of emotional energy as well, besides this up / down, high / low in solidarity and enthusiasm. We will see that there are two major dimensions of stratification (power and status) that produce specific qualities of emotional energy. But while we are considering the main, generic level of emotional energy, I will mention one more Durkheimian feature. Emotional energy is not just something that pumps up some individuals and depresses others. It also has a controlling quality from the group side. Emotional energy is also what Durkheim (1912/1954) called “moral sentiment”: it includes feelings of what is right and wrong, moral and immoral. Persons who are full of emotional energy feel like good persons; they feel righteous about what they are doing. Persons with low emotional energy feel bad; though they do not necessarily interpret this feeling as guilt or evil (that would depend on the religious or other cultural cognitions available for labeling their feelings),6 at a minimum they lack the feeling of being morally good persons that comes from enthusiastic participation in group rituals.
Feelings of moral solidarity generate specific acts of altruism and love; but there is also a negative side. As Durkheim pointed out, group solidarity makes individuals feel a desire to defend and honor the group. This solidarity feeling is typically focused on symbols, sacred objects (like a tribal totemic emblem, a holy scripture, a flag, a wedding ring). One shows respect for the group by participating in rituals venerating these symbolic objects; conversely, failure to respect them is a quick test of nonmembership in the group. Members of the ritual group are under especially strong pressure to continue to respect its sacred symbols. If they do not, the loyal group members feel shock and outrage: their righteousness turns automatically into righteous anger. In this way, ritual violations lead to persecution of heretics, scapegoats, and other outcasts. Such events bring out clearly yet another transformation of emotion by rituals: from specific initiating emotions to their intensification in collective effervescence; from collective effervescence to emotional energy carried in individuals’ attachment to symbols; and from symbol-respect to righteous anger.
Detailed microsociological evidence of such emotional transformations is provided in the work of Scheff and others (Scheff 1990; Scheff and Retzinger 1991; Samson 1997). Scheff’s theoretical model builds also on Durkheim, but gives emphasis to the emotions experienced by individuals as touching their selves. For Scheff, intact social bonds (which, from the point of view of IR theory, are the result of carrying out a successful IR) give participants a feeling of pride; broken social bonds (an unsuccessful IR) results in a feeling of shame. Scheff and his collaborators examine social interactions in micro-detail by using video and audio recordings (largely from marriage counseling sessions, as well as from family interactions). Pride and shame are documented in the patterns of body alignment, eye gaze, speech hesitations or flow, loudness as well as overt expression of emotions. These data show the ups and downs of mutual focus and emotional entrainment on the second-to-second level.
Scheff goes on to point out that shame—the sense of broken social attunement—can either be immediately expressed and brought into the interaction as a topic; or it can be by-passed, repressed from conscious verbal attention. By-passed shame, he argues, is transformed into anger. This sets up a cycle of repeated failed interactions: for example, a married couple or parent and child may shame one another by breaking the attunement of interactions, but ignoring the shame; it thereby comes back in angry moves later in the same encounter, or in later encounters. Emotional dynamics recycle through the IR chain, since each episode of broken attunement generates more shame and more anger, which comes out in yet further patterns of interaction.
The negative effects of broken attunement can also be read in a comparative light, as a demonstration of the importance of attunement. Scheff shows that Durkheimian solidarity, operating on the micro-level of situational encounters, is highly attractive to individuals, and is experienced as pride, a favorable social self. The failure of solidarity, down to the minute aspects of coordinating mutual participation in a conversation, is felt as a deep uneasiness or affront, which Scheff refers to as a feeling of shame. In the Durkheimian model, violation of solidarity brings the reaction of righteous anger; this results in yet another highly ritualized interaction, a ritual of punishment. Durkheim’s theory of crime (1895/1982) holds that punishment has the effect of reinforcing the group’s commitment to its symbolic ideals, whether or not it is successful in deterring the violator from future transgressions. In Durkheim’s view, punishing criminals is carried out not as a utilitarian act to manipulate the reinforcement schedule of the criminal, but as a ritual to maintain the group’s solidarity. Scheff shows a similar dynamic operating through individual emotions: violation of solidarity brings anger; but the ritual expression of anger does not bring a return of solidarity in an alienated relationship, but rather leads to a further round of shame, anger, and ritual retaliation. Durkheim stops his analysis at the point where the punishment ritual takes place, and does not inquire what it does to the criminal’s future behavior. Scheff extends the Durkheimian model into a chain, a vicious cycle.
But there is yet another way in which the emotions might be transformed. The failed interaction—the breakdown of solidarity that generates shame—can be followed by a different sequence. The failure itself can become the explicit focus of attention for an interaction in which the shamed or violated person gets to express his or her feeling of outrage directly to the perpetrator; if the latter acknowledges it, social solidarity is reestablished. This is the model of “restorative justice” implemented by the criminologist Braithwaite and others (Braithwaite 1989; Strang and Braithwaite 2000). Criminals are confronted at group meetings by their victims as well as other members of the social networks on both sides. These encounters have often been remarkably successful in reconciling the contending parties and in reducing repeated offenses. In terms of IR theory, these reconcilation circles work because they are high intensity IRs; all of the ingredients of figure 2.1 are present to a high degree. The mutual focus of attention is enforced, in part, because a police officer makes the offender pay attention to what the victim is expressing. The initiating emotional ingredient is high: the strong feelings of shame and anger; these feelings are shared and transformed, because all the persons in the circle get to express their opinions and feelings, and are swept into a common mood. The result is that the offender is shamed and ritually punished, but then is reintegrated into the group by participating in the group emotion of collective solidarity. Restorative justice groups are a striking example of how an IR can take any topic and any initiating emotion, and transform it into solidarity.
STRATIFIED INTERACTION RITUALS
The model of interaction rituals gives us the general process of interaction. IRs themselves are variable, insofar as rituals can be successful or unsuccessful, that is, in terms of how much focus and emotional contagion actually takes place, and hence how strongly the participants become attached to membership symbols. Because of these variations, interactions are stratified: some persons have the power to control others through rituals, while others are passive or resistant; some persons are in the center of attention, while others are marginalized or excluded. These are the two dimensions of power and status. As we shall see, just where people are located in such IRs is a major determinant of individual personalities.
Power Rituals
Power operates on the micro-interactional level by all those factors that bring together individuals who are unequal in their resources such that some give orders and others take orders, or more generally dominate the immediate interaction. This is an interaction ritual, insofar as it involves focusing attention on the same activity, and becoming aware of each other’s involvement; and it has a shared emotional focus, which builds up as the ritual successfully proceeds. (As always, it is also possible that the ritual will not proceed successfully, that it will break down into avoidance or conflict; but let us deal with that variant separately.) The focus of a power ritual is the process of giving and taking orders itself. As many organizational studies show (especially the classic studies of informal work groups, many of which are used as an empirical base by Goffman [1959]), the order-takers do not necessarily carry out the bosses’ orders; for that matter, the bosses do not always expect them to do so, or do not even know very clearly what they want done. But the crucial item of attention is showing respect for the ordergiving process itself. Order-givers are in charge of a Goffmanian frontstage performance; they take the initiative in it, and if they are successful, they uphold the organizational chain of command. For this reason, the order-giving classes have a Goffmanian “frontstage personality”; they are attached to their frontstage roles. In Durkheimian terms, order-givers enhance or sustain their emotional energy by dominating during power rituals; and their ritual stance makes themselves loyal to the symbols of the organization. Their cognitions are of the “official” sort (see evidence summarized in Collins 1975, 62–87).7
People who are order-takers participate in these rituals in a different way. They are required to take part: whether by the raw coercion of military force (as in the army, a prison camp, or in feudal / aristocratic societies), or by the slightly more long-range coercion of a paycheck, fines and privileges, or chances of promotion wielded by bosses, teachers, and other persons in authority. The situation of taking orders, of being coerced, is in itself alienating. But persons subject to authority usually cannot evade it directly; their resistance usually occurs in situations when they are out of the direct surveillance of an order-giver—for example, in Goffmanian backstages where they criticize or ridicule their bosses, or in their normal work routine, in which they put in a perfunctory performance. In this sense, the order-taking classes have a “backstage personality.”
Order-takers nevertheless are required to be present at order-giving rituals, and are required to give at least “ritualistic” assent at that moment. They and their boss mutually recognize each other’s position, and who has the initiative in the ritual enactment. Power rituals thus are an asymmetrical variant on Durkheimian interaction rituals. There is a focus of attention, in this case, on the order-giving process. But the emotions that are invoked are constrained; there is a tone of respect, of going along with what the order-giver is demanding. The more coercive and extreme the power differential, the more emotional contagion there is. The medieval peasant, or the child who is being beaten, is forced to put him or herself into a state of compliance, of going along with what the master / parent / authority figure wants. It is a coerced focus of attention; the order-takers have to try hard to anticipate what the order-giver wants. Conversely, the order-giver uses coercion precisely to feel this mastery over the subordinates’ minds, to “break their will.”8 Less coercive forms of order-giving have correspondingly less powerful ritual effects.
According to this theory, a successful order-giving ritual coerces a strong mutual focus of attention, and produces a situationally dominant emotional mood. But it is a heavily mixed emotion. Insofar as there is successful role-taking on both sides (and that is at the core of any successful ritual), the order-giver feels both his / her own sentiment of mastery, and the order-taker’s feeling of weakness. On the other side, the order-taker has a mixture both of his / her own negative emotions—weakness/depression, fear—and the mood of the dominator, which is strong emotional energy, dominance, anger. This explains why persons who are severely coerced (concentration camp inmates, marine corps recruits, beaten children) tend on one level to identify with the aggressor, and will enact the aggressor’s role when possible in the future: they have an emotional complex of fear and anger, although situationally the fear side is dominant when they are taking orders. Conversely, order-givers who use extreme coercion acquire sado-masochistic personalities, because of the role-taking that goes on, thus blending anger / dominant feelings with a sense of the fear and passivity that they invoke in their subordinates. Thus the experience of momentary, situationally dominant emotions gives rise to long-term emotional styles, which is a large part of what is meant by the term “personality.”
Power rituals produce complex emotions. Order-givers and ordertakers share the dominance / anger / fear / passivity complex, but in very different proportions. Considered analytically, power rituals appear to be less effective than status rituals in generating large amounts of EE for dominant individuals; for subordinates, on the other hand, power rituals have serious emotional consequences. Exercising ordergiving power increases one’s EE insofar as it coincides with being in the center of attention of a situation of emotional entrainment rising to a palpable level of collective consciousness, which is what I call a status ritual: intense versions of this coincidence include military officers in combat, athletic coaches in the course of a contest, and somewhat less dramatic occasions in business and professional activities where there is a shared level of intensity among the participants. When the power ritual does not coincide with a status ritual, the person exercising power does not usually experience much EE gain, but at any rate it keeps the power holder from losing EE. Order-takers, however, generally lose EE, especially when the power ritual does not bring about a solidarity ritual.
Order-givers and order-takers also share an orientation toward dominant symbols, but again with a different blend of emotions. Ordergivers identify themselves with the sacred objects of their organization; they respect these symbols as ideals, and are foremost in requiring other people to kowtow to them too. This is the conservatism of dominant classes, their self-appointed motivation as upholders of tradition, as restorers of law and order, and as righteous uprooters of heretics and deviants.
Order-takers, on the other hand, have an ambivalent attitude toward the dominant symbols. They are alienated from these symbols, and privately speak and think of them cynically, if they can get away with it.9 Thus the modern working class is generally alienated from the business ideals of their bosses, and troops ridicule the rhetoric of their commanders. These symbols become, so to speak, negative sacred objects; when and if rebellion is possible, a suddenly liberated order-taking class wreaks vengeance on the symbols that they formerly had to bow to. (Kids without career chances in the academic system, who are forced order-takers in schools, thus tend toward acts of vandalism and other forms of “deviance” directed precisely at the “sacred objects” in whose name they are subordinated: see Cohen 1955.) It is also possible that order-takers hold the dominant symbols in a kind of superstitious respect; that is, if they are so tightly coerced that there is little opportunity for distancing themselves, no backstages into which they can retreat from their masters’ surveillance, they are ritually forced to show respect for the sacred symbols at all times. Thus arises the “loyal retainer” mentality, found among long-time servants and peasants (and in a different context, among children who are strongly coerced by their parents, but also strongly controlled, and given no opportunities to rebel). The difference between these two kinds of order-takers’ attitudes—alienated or subservient—depends primarily upon ecological structures: whether coercive control is continuous, or allows breaks into backstage privacy.
I have schematically outlined two polar types of participation in power rituals: order-giving and order-taking. But power rituals are a continuum. There are several kinds of positions in the middle between the extremes: persons who are order-transmitters, who take orders from someone above them and give orders to others below; these persons tend to blend the order-givers and order-takers culture into a narrow and rigid “bureaucratic personality.”
There is another kind of midpoint between extremes: the person who neither gives nor takes orders, but who interacts with others in egalitarian exchanges. Analytically, this is a point within the power dimension where there is no power; hence the effects of order-giving and order-taking are both neutral. To explain what will happen at this neutral level of power, in “horizontal” relations among equals, we must turn to the status dimension.
Status Rituals
I am using the term “status” not as a general term for hierarchical differences of all kinds, but in a restricted sense of belonging or not belonging. At the micro-level of the encounter, status is the dimension of inclusion or exclusion. This, too, is a continuum; in everyday life, it appears as popularity versus unpopularity.
This dimension of membership versus nonmembership is analytical, in the sense that any individual (and any interaction) can be classified both as to where it stands in terms of status membership, and in terms of power inequality. That means that every interaction is producing both status membership effects and power effects, and every individual is subjected to both of these kinds of effects from one situation to the next. The power effects, however, might be zero, if there is no order-giving and order-taking in that situation; on the other hand, even extreme situations of order-giving also have a status dimension, insofar as the group is assembled and some membership feelings are being generated.
In what ways can individuals differ in their status group participation? Here we need to tease apart four aspects. Two of these are characteristics of the micro-situation itself and the individual’s location within it. Two are meso-level characteristics of the IR chains: what happens over time as situations repeat.
First, on the micro-level, we must ask, How successful is the interaction ritual? In other words, does it build up to a high level of collective effervescence, a moderate level, or little emotional entrainment at all? The higher the ritual intensity, the more emotion is generated both in the immediate present and for long-term effects. Ritual intensity thus operates as a multiplier for the other three aspects of ritual effects.
Again, on the micro-level: Where is the individual located as the IR takes place? There is a continuum from persons who are on the fringes of the group, just barely members, barely participating; others nearer the core; at the center is the sociometric star, the person who is always most intensely involved in the ritual interaction. This person is the Durkheimian participant of the highest degree, and experiencing the strongest effects of ritual membership: emotional energy, moral solidarity, attachment to group symbols. At the other end, there is the Durkheimian nonmember, who receives no emotional energy, no moral solidarity, and no symbolic attachments. This is the dimension of central / peripheral participation.
Next, on the meso-level, as IRCs string situations together: What proportion of their time do people spend in each other’s physical presence? This is the dimension of social density. At one end of the continuum individuals are always in other people’s presence, under their eyesight and in their surveillance; this leads to a high degree of conformity, a feeling of social pressure on onself, but also a desire to make other people conform as well. At the other end of the continuum individuals have a great deal of privacy (social and physical spaces where others do not intrude; Goffmanian backstages) or of solitude (other people are simply not around). Here pressures for conformity are low. Social density is a quantitative matter, an aggregate of a chain of situations over time. An individual might occasionally be in other people’s presence, perhaps even in very intense IRs, but their effect is quite different than if he or she were almost always in such situations. That is to say, a person with a high degree of privacy or solitude (low overall social density) might treat these occasional high ritual intensity episodes as sharp breaks from ordinary consciousness, either as wonderful and longed-for experiences, or as unwelcome intrusions and threats to his or her privacy. Which of these is the individual’s response depends on additional features (his or her peripheral / central position and location in the power dimension).
Again on the meso-level: Who are the participants who come together in the aggregate of IR chains? Is is always the same persons, or a changing cast of characters? This is the dimension of social diversity, which might also be called the dimension of localism / cosmopolitanism. Specifying the argument of Durkheim’s Division of Labor in Society, low diversity should produce local solidarity, strong attachment to reified symbols, literal-mindedness, and a strong barrier between insiders and outsiders. There is high conformity within the group, along with strong distrust of outsiders and alien symbols. At the other end of this subdimension, there is participation in a loose network consisting of many different kinds of groups and situations. Durkheimian theory predicts the result of cosmopolitan network structure is individualism, relativistic attitudes toward symbols, abstract rather than concrete thinking.
Stated in terms of emotions, this implies that persons in cosmopolitan networks have relatively weak feelings of conformity to group symbols; emotional coolness of tone; and generalized trust in a wide range of interactions. When symbols are violated or ritual procedures go badly, members of tight, localized groups respond with anger and fear (especially if rituals are backed up by coercion on the power dimension). Can there be ritual violations in loose cosmopolitan groups, where there is less intensity and conformity? Yes, because there can be violations of the appropriately casual and sociable tone of interaction.10 Goffman (1959, 1967) concentrated most of his analysis on situations of cosmopolitan interactions, and depicted just such violations and their sanctions. Following Goffman, I would suggest that persons in these situations respond by amusement to minor ritual violations by others, and with embarrassment, contempt, and a desire to exclude perpetrators of more serious violations of the sociable order. The persons who commit these Goffmanian sacrileges feel anxiety and embarrassment.
Durkheim’s (1893/1964) pioneering analysis did not pull apart these various dimensions built around the mechanism of ritual solidarity. His terminology conflates all four into an overall level of what he called “moral density.” His most differentiated argument distinguished “mechanical” and “organic” solidarity, which was a move in the direction of seeing multiple causes. In effect, “mechanical solidarity” is the overlap of high social density and low social diversity (localism), with an implication that there is also high ritual intensity, and that most individuals experience relatively central participation—which also seems to assume that the group is lacking in power differences. To be sure, this overlap would constitute extremely high degrees of solidarity, conformity, and attachment to the group as the sole source of emotional energy. “Organic solidarity” is a situation of high social diversity (cosmopolitanism; i.e., the modern division of labor, as contrasted to undifferentiated small tribal or rural communities); but he left it unclear what variation there might be in the other dimensions. Durkheim seems to have envisioned relatively high ritual intensity, so that organic solidarity would provide sufficient solidarity, morality, and conformity to keep modern society together. But he (and his followers and critics) were never satisfied with the organic solidarity theory. An underlying problem was the failure to distinguish enough subdimensions to recognize all the different combinations that might exist, and that indeed are found all across the historical landscape.
This, then, is my set of hypotheses about how the various dimensions of interaction ritual affect emotions. By way of summary, let us recapitulate the model, first in terms of the effects on long-term emotions (emotional energy), and then in their effects on short-term, transitory emotions.
Effects on Long-Term Emotions: Emotional Energy
The IR chain model proposes that individuals acquire or lose emotional energy in both power and status interactions. Order-givers maintain and sometimes gain EE, order-takers lose it; being in the focus of attention and thereby successfully enacting group membership raises EE, experiencing marginality or exclusion lowers it. Interaction rituals are connected in chains over time, with the results of the last interaction (in emotions and symbols) becoming inputs for the next interaction; thus EE tends to cumulate (either positively or negatively) over time.
Emotional energy is an overall level of being “up” or “down,” ranging from enthusiasm to depression. Between interactions, EE is carried in the individual’s stock of symbols, in the cognitive part of the brain; it is an emotional mapping of the various kinds of interactions that those symbols can be used in, or that can be thought about through symbols. Thus emotional energy is specific to particular kinds of situations; it is a readiness for action, that manifests itself in taking the initiative in particular sorts of social relationships or with particular persons.11 Thus there is EE specific to power situations—expecting to dominate, or be dominated—as well as an EE specific to status situations—expecting to be a central member, or a marginal one, or not to be accepted at all. Furthermore, these emotional energies tend to be specific to particular networks and groups, or to particular kinds of them: some persons feel full of confidence and initiative in a gathering of professional acquaintances, but not in a sexual situation; some feel confidence in a business negotiation, but not a political one; persons who dominate the center of attention in an intellectual gathering may fade into shyness at a drinking party. It is in this sense that, as we will see, sexual drive is a form of EE.
People move through the chain of encounters that make up their daily lives on an up-and-down flow of EE. They are more attracted to certain situations than others, and sometimes feel disinterest or repulsion. In each situation as it unfolds, their own emotional and symbolic resources, meshing or failing to mesh with those of the people they meet, determines to what extent the IR will be successful and unsuccessful. These outcomes, in turn, raise or lower EE. The end result is motivation to repeating those sorts of encounters with particular persons and to avoid them with others.
Emotional energy manifests itself both physically and psychologically; but its underlying basis—the form in which it is “stored,” so to speak—is not as physical energy per se. EE has a cognitive component; it is an expectation of being able to dominate particular kinds of situations, or to enact membership in particular groups. The cognitive side of this is that symbols (particularized memories as well as generalized ideas or emblems) have emotional energy attached to them, in the sense that the symbols call forth a high or low degree of initiative in enacting social relationships using those symbols. But this is not ordinarily a process of conscious calculation, of the actor thinking “I will get a good feeling of power or status if I interact with so-and-so.” Instead, certain symbols come to mind, or appear in the external environment, and spark off propensities (positive or negative) for social action. The “expectation” may work on a subconscious level. It is an anticipation of being able to coordinate with someone else’s responses, of smoothly role-taking in the ongoing flow of the interaction, and thus anticipating the buildup of emotional force that goes on within a successful IR. The process of rhythmic entrainment of the ultra-micro aspects of interaction is the mechanism by which emotional contagion occurs within a successful interaction. Thus there is a very finegrained, micro-anticipation that happens within the interaction itself (on a level down to fractions of a second), as well as a more long-term expectation of being able to enter into such micro-coordination with particular kinds of people. Emotional energy exists as a complex of these kinds of expectations, a priming for successful ritual interaction in particular settings.
The low end of EE is depression, manifested in withdrawal, both from expressiveness and activity. Depression appears to be a more complex process than high EE.12 Experience at the low end of the power dimension brings depression: low energy, loss of motivation. But this may happen only when order-takers experience a strong degree of being under someone else’s control. When their lack of control is only moderate, they may typically respond by anger—by a temporary increase in the output of EE, as vigorous reactance against the situation that is controlling them (Frijda 1986, 290). The middle level of negative interactional experience—in temporal terms, an episodic and atypical experience of being subordinated—thus has a distinctive emotional effect.
Negative experience on the status dimension has a similar contour: declining EE, with a flare-up of anger in the middle range and where the flow of emotional expectations from the IR chain is episodically dashed. Over the long run, I suggest that failure of membership in a group ritual brings a degree of depression commensurate with the degree of social exclusion. Kemper (1978), however, argues that low status brings anger as well as shame. Scheff (1990; Scheff and Retzinger 1991) present evidence that exclusion on the micro-level of the encounter, breaking attunement, brings shame, which may get into a spiral with anger. From the point of view of IR theory, shame is a form of low EE, with a distinctive cognitive component directed toward one’s social image (i.e., social membership) in a particular group. Anger occurs when there is an abrupt negative change in expected social membership feelings. It is a short-term emotion due to the disruption of expectations; the long-term effect of membership loss is nevertheless depression. Hence there is no long-term increase in vigor of the sort that an angry reaction brings for moderate levels of put-down on the power dimension, that is, when there are structural opportunities for mobilizing rebellion.13
Scheff’s model is a valuable complement to IR theory because it specifies emotions generated by both high and low levels of Durkheimian solidarity. Successful interactional attunement or an intact social bond generates pride; breaking the bond generates shame. In the terms of IR theory, pride is the emotion attached to a self energized by the group; shame is the emotion of a self depleted by group exclusion.14 As we will see below, nonverbal and paralinguistic measures of pride and shame can be useful as measures of high and low EE. Pride is the social attunement emotion, the feeling that one’s self fits naturally into the flow of interaction, indeed that one’s personal sense epitomizes the leading mood of the group. High solidarity is smooth-flowing rhythmic coordination in the micro-rhythms of conversational interaction; it gives the feeling of confidence that what one is doing, the rewarding experience that one’s freely expressed impulses are being followed, are resonated and amplified by the other people present. When Scheff speaks of shame as the broken social bond, I take this to mean that the rhythm is impaired, that one’s spontaneous utterances are choked off—even for fractions of seconds—that there is a hesitancy about whether one is going to be understood, and hence about whether it is possible to formulate a clear or understandable utterance at all. The shared rhythm is what enables each person to anticipate what the other will do, not in specific contents, but in rhythmic form: a certain rhythm of talk is launched, characterized by a certain energy, a certain emotional flow. The conversational ritual generates high solidarity to just the extent that its participants pick up the same rhythm, molding their utterances to the rhythm that they have established in the past few moments, and riding its waves to anticipate just how their next set of utterances will flow one upon the other. By observing these rhythms, we can see emotional energy in the process of being manifested in the micro-situation.
The main long-term emotional energies resulting from stratified interaction, then, are: high levels of enthuasiasm, confidence, initiative, and pride, resulting from controlling the attunement of interaction in either a power or status situation; low levels of the same (i.e., depression, shame), resulting from being dominated in a power situation, or excluded from a status situation. There is one other long-term emotional disposition: the amount of trust or distrust of other people. At the trust end of the continuum, this simply manifests itself as high EE, willingness to take initiative toward certain social situations. At the distrust end, it comes out as fear of particular situations. Distrust / fear is attached to particular structural configurations, namely distrust of those who are outsiders to the local group; it is the result of the structural subdimension of status group interaction, in which there is tight local closure of group boundaries.
Emotion Contest and Conflict Situations
In power situations, gains of emotional energy by one person and EE loss by the other person are reciprocally related. This may also happen in sociable situations. Some persons act as energy drainers, bringing other persons down while dominating the situation. Consider the micro-mechanisms of an interaction ritual: the common focus of attention, the rhythmic coordination that intensifies emotions. Persons who control the situation can frustrate this process. They can break the micro-rhythm, by not responding to the signals the other person is putting out (by changing or reframing the topic, starting new activities, ignoring and overriding nonverbal contact cues). This is one way that order-givers establish their dominance, perhaps most likely when there are signs of challenge to their control. It also makes up the substance of aggressive status contests that happen in sociable conversations (i.e., what Goffman [1967, 24–25] calls face-work contests).
Such contests are an activity that breaks the focus of ritual microcoordination and prevents the circular buildup of anticipations on both sides. In smoothly running situations, one’s own ability to use symbols in thinking and talking depends upon anticipating the other’s reactions, and feeling the surge of symbolic group membership with each successful use of a commonly recognized symbol. Dominance contests break this down (whether by the deliberate intention of one person, or just inadvertent lack of interest in the other—i.e., the dominant or more attractive person’s emotional energies are directed elsewhere). The result for the person who is unable to carry through their intentions and anticipations is that there is a blockage in the smooth flow of their own thoughts, words, and actions; they are unable to project a micro-future in this situation, and this is what it means to lose emotional energy.
If the failure of an interaction ritual to achieve coordination and emotional buildup debilitates one person, though, why shouldn’t it bring down the other person emotionally too? According to the basic IR model, the emotional flow is a group process; what one side fails to get, the other side should fail to get also. But in some kinds of situations the result may be unequal. Consider larger group structures in which particular micro-interactions are embedded: the boss confronting a rebellious worker within an organization; or an athletic contest before a group of spectators. The person who dominates the microsituation has the possibility (which may be overt or only subjectively felt) of gaining recognition in the larger group context. And conversely, this individual may bring along previously generated feelings of membership in the larger group structure, the emotional energy of being a dominant figure capable of mobilizing an enforcement coalition (in a formal organization), or of being a popular person (before an audience of fans).
Chambliss (1989) has studied this interaction in the case of athletic contests (competitive swimmers), and has found that there is a major difference in outlook between high-level performers (consistent winners) and lesser performers (losers). The difference is manifested in the details of behavior: winners are meticulous in performing their routines in ways that they have deliberately developed; they have built up their own rhythms and stick to them in the face of competitive opposition. The winners make themselves the focus of attention; they set the expectations around themselves. Losers, however, let the winners become the focus, and adapt their micro-behavior toward them. This implies that a winner (perhaps dominant persons generally, in dominance contests more widely as well as in athletics) has a sense of control throughout the situation: winners maintain and build up their own rhythmic coordination, their anticipation of what they will do, setting the micro-rhythmic pace. Losers (and persons who are subordinated in dominance contests) allow someone else to break their own flow of anticipation of what will happen in their own activities. These dominated persons can cope with the situation, can maintain some anticipation about what will happen only by focusing on the other person as the lead, rather than by projecting their own volitional future. In effect, such a person can recoup some emotional energy from the situation by becoming a follower, attaching themself to someone else’s lead.15 The more they resist such attachment, the less emotional energy they will have.
Figure 3.1 Winner focuses on the goal, loser focuses on the winner. Final lap of relay race, which runner E is about to win.
In terms of the IR model, one could also say that the dominant person makes oneself the focus of the interaction. He or she becomes, in some sense, a Durkheimian sacred object. Microsociologically, that is just what a “sacred object” means—it is the object upon which attention of the group is focused, and which becomes a symbolic repository of the group’s emotional energies. When someone feels oneself in this position, they have a store of emotional energy for their own use; it makes that person “charismatic.” For others, the person who is a “sacred object” compels attention. They become spectators to that person. Their attitudes as spectators can vary. If they throw themselves completely into acquiescence, they become compliant admirers, who want to attach themselves and draw some flow of the “sacred” emotional energy for themselves (like fans asking for an autograph).
At the other extreme, they may be resentful would-be or failed competitors. But even their resentment is a feeling based upon recognizing that the other person has a special status as “sacred object” that they do not have. Chambliss (1989) describes this difference as the “mundanity of excellence.” Persons inside the social realm of winning / dominance experience a mere routine, in which they have smooth anticipated control of situations—that is, a great store of “emotional energy” available to them in contest situations. But persons on the outside looking in see a mystifying difference, a gulf to greatness that they feel they cannot cross. These differences are, of course, most exaggerated in highly publicized contest situations, like the Olympic athletes Chambliss studied. In lesser degrees, dominant persons are also little “sacred objects” at least in certain small local situations, while subordinated persons are left with the choice of being participating spectators of their dominant energy, or feeling the energy drain of opposing them.
There is fine-grained micro-situational evidence of this process in a study by Erickson and Schultz (1982) of video-taped interactions between junior college counselors and students. Typically, these dyads fell into the same rhythm of syllables by pitch and loudness, both in their own turns and in the turn-taking rhythm as they shifted from one speaker to the other; these beats were often synchronized on the microkinesic dimension with body movements. This pattern could be interpreted as a baseline of IR solidarity. At times, one person takes the lead in holding the rhythm, while the other person flounders for fractions of seconds (usually in small-scale upsets on the order of a few quarterseconds) in an unrhythmic pattern, then follows the pattern maintained by the first. These recordings display situational dominance on the ultra-micro level of seconds. There are also instances where the tapes show the two persons’ rhythms mutually interfering with each other; or where both go along in different rhythms, as if deliberately in opposition to one another. In follow-up interviews where the subjects were shown tapes of themselves and asked to describe what was going on, they tended to comment on the “uncomfortable moments” where the rhythmic coordination broke down, but not on the moments when their rhythms were in sync. It appears that the participants took solidarity for granted, and only noted its absence. For the most part, the subjects seem to experience the pattern of the interaction subliminally; it was only after repeated viewing of the tape recording, and repeated discussions with the experimenter about their reactions to “uncomfortable moments,” that students began to become consciously angry about what they came to see as their dominated position in the interaction.16
Power derives from a variant on the basic IR model. In its Durkheimian formulation, successful rituals produce group solidarity. Teasing apart the mechanisms and fine-grained processes of an IR, we could say instead that successful IRs produce heightened mutual focus and bodily emotional entrainment. Power is an asymmetrical focus of attention upon such a situation, so that one side battens on the energy that all the participants have mutually produced. In a power ritual, the social battery is revved up, but the benefit goes largely to one side.17
SHORT-TERM OR DRAMATIC EMOTIONS
Most research on emotion has focused on the short-term, dramatic emotions: the “phasic” rather than the “tonic,” the outbursts that disrupt the ongoing flow of activity (Frijda 1986, 2, 4, 90). My argument is that the short-term emotions are derived from the baseline of emotional energy; that it is against the backdrop of an ongoing flow of emotional energy that particular disruptive expressions are shaped. Surprise, for example, is an abrupt reaction to something that rapidly and severely interrupts the flow of current activity and attention. This is also the general pattern of more important short-term emotions.
The positive emotions become intense largely because of a contagious buildup during an interaction ritual. This is the case with enthusiasm, joy, and humor: all of these build up in social situations as the result of a successful ritual. Psychological analysis tends to take these emotions from the individual viewpoint. For example, joy is explained as the result of the momentary expectation of success in some activity (Frijda 1986, 79). This is sometimes true; but joy and enthusiasm are particularly strong when an assembled group is collectively experiencing this expectation or achievement of success (e.g., fans at a game, political partisans at a meeting). Further, the group itself by a successful emotional contagion can generate its own enthusiasm (which is what the flow of conversation at a party does).
These kinds of positive emotional outbursts are relatively short and temporary in their effects. They happen upon a baseline of previous emotional energy: for a group to establish this kind of rapport, its members need to have previously charged up some symbols with positive attraction, so that these symbols can be used as ingredients in carrying out a successful ritual. A previous cumulation of emotional energy is thus one of the ingredients in making possible the situational buildup of positive emotion. Frequently, the positive emotions (joy, enthusiasm, humor) are generated by a group leader, an individual who takes the focus, who is able to propagate such a mood from his or her own stores of emotional energy. This individual thus serves very much like an electric battery for group emotional expressiveness. Persons who occupy this position in IR chains are what we think of as “charismatic.” In general, “personality” traits are just these results of experiencing particular kinds of IR chains. (This is true at the negative end as well, resulting in persons who are depressed, angry, etc.)
The negative short-term emotions are even more clearly related to the baseline of emotional energy.
Anger is generated in several ways. Psychologically, anger is often regarded as the capacity to mobilize energy to overcome a barrier to one’s ongoing efforts (Frijda 1986, 19, 77). This means that the amount of anger should be proportional to the amount of underlying effort; and that is the amount of emotional energy one has for that particular project. High emotional energy may also be called “aggressiveness,” the strong taking of initiative. This can have the social effect of dominating other people, of lowering their emotional energy, of making them passive followers. This implies that there is a connection between the generic quality of high emotional energy—especially the EE generated in power situations—and the expression of the specific emotion of anger.
The disruptive form of anger, however, is more complicated. That is because anger in its intense forms is an explosive reaction against frustrations. Truly powerful persons do not become angry in this sense, because they do not need to; they get their way without it. To express anger is thus to some extent an expression of weakness. However, persons who are powerful can afford to become angry; their power-anger is an expression of the expectation that they will get their way against the obstacle. In the case of a social obstacle—the willful opposition of some other person—it is an expression of the powerful person’s confidence that he or she will be able to mobilize an enforcement coalition to coerce the opposition into compliance, or to destroy the resistance. Previous stores of EE thus determine when and how someone will express explosive anger.18
The most violent expression of anger occurs when one feels strong in overcoming a strong frustration. If the frustration itself is over-whelmingly strong, the feeling is fear, not anger. Persons who are weak do not manifest anger in the same way. It is only when they have enough resources to be able to mount some resistance (or at least some social privacy, a separate social circle in which they can utter symbolic threats) that weak persons, order-takers, have anger. This follows from the principle that the core of anger is the mobilization of energy to overcome an obstacle. It is only when there are enough social bases of support to generate EE that one can react to a frustration (in this case, being dominated) by mobilizing anger. Persons who are too weak (i.e., in their IR chains they lack resources or space in which to mobilize any other socially based EE), do not react angrily to domination but succumb to depression.
In between these two situations there are selective outbursts of anger. This is the targeted anger that individuals feel against particular other persons. It occurs because these individuals are structural rivals in the market of social relationships: for example, two women competing for the same man, or two intellectuals competing for the same audience. Here one does not feel angry against someone who is stronger than oneself (rebellious anger), nor against someone weaker (dominance anger); rather, this is a case of someone frustrating one’s own projects. The anger here is not really “personal”; there is no role-taking (as in the dominance / subordination forms of anger), although the target is a person, and the underlying structure is a social one; it is only an accident that the obstacle to one’s goals happens to be a person.
An especially Durkheimian form of short-term emotion is righteous anger. This is the emotional outburst, shared by a group (perhaps led by particular persons who act as its agents) against persons who violate its sacred symbols. It is group anger against a heretic or scapegoat. Such anger only happens when there is a previously constituted group. One can predict that righteous anger is proportional to the amount of emotional charge of membership feelings around particular symbols. The amount of such charge, in turn, is highest where the group has high social density and a local (rather than cosmopolitan) focus. Where the group networks are diffuse and cosmopolitan, on the other hand, the short-term emotion felt at disruption is embarrassment on behalf of the disrupter—resulting in status exclusion, unwillingness to associate with that person, rather than in a violent ritual punishment to restore symbolic order.19
Righteous anger has great importance in political sentiments as well as in the dynamics of local communities (scandals, witch-hunts, political hysterias). The theoretical difficulty is understanding just how this kind of anger relates to the power and status dimensions of group structure. In the Durkheimian model, it seems to be the group in general, and all its adherents, who are outraged at the violation of its symbols. But anger, and hence violence as a punishment (burning a witch or heretic at the stake, throwing drug dealers or gamblers or abortionists in jail) is related to the power dimension, since the use of violence is the ultimate sanction of power. To explain righteous anger, we need to observe the power and status dimensions in conjunction—in instances where the status group structure is dense enough and locally closed enough so that there is a strong sense of group membership, attached to reified symbols; and where this ritual community has a power hierarchy within it, which regularly exercises coercive threats to enforce obedience to orders. Under these circumstances ritual violations (violations of membership symbols on the status dimension) are taken as a threat to the power hierarchy as well.
Righteous anger is a particularly intense emotion because it is expressed with a strong sense of security: the individual feels that they have the community’s support, and not merely in a loose sense. Righteous anger is an emotion that is an evocation of the organized network that has been previously established to use violence. Persons who feel righteous anger are evoking their feeling of membership in an enforcement coalition.
As evidence, I would point to the fact that the most violent punishments for ritual deviance (witch-burning, public tortures, and executions in medieval patrimonial states; violent atonement for taboo violations in tribal societies) occur where the political agents are both highly coercive in their ordinary operations, and are active in enforcing group cultures (Collins 1974; Douglas 1966). Heresy trials and violent ritual punishments have declined in keeping with the degree of the separation between church and state; it is where these spheres (the power hierarchy and the status community) are fused that righteous anger is most prevalent. In some degree, however, the political hierarchy still remains the focus of status rituals—through its claims to be a community as well as an organization for wielding power. This makes it possible to mobilize deviance-hunting as a form of status intrusion into the political sphere, even in relatively differentiated modern societies. And it is advocates of a return to the fusion of community with polity who are most strongly involved as “moral entrepreneurs” in modern deviance-hunting. Such advocates often come from the localized sectors of modern society, especially the remnants of traditional and rural communities. In addition, the attempt of socialist regimes to keep up a high level of collective solidarity helps to explain their concern for rituals of conformity.
Fear is another short-term negative emotion. The most intense and briefest forms of fear are those that most sharply disrupt activities; at the extreme, intense fear experience is next to a startle response. Crying is an expression of fear in a more complex sense: it is a social call for help in distress. Adults do not cry very much, because their horizon widens out. Instead of relatively short-term and simply physical threats or discomforts, the most important form of fear becomes fear of social consequences: fear of being coerced or fear of social exclusion, which are more long-term experiences. Furthermore, since the problem is itself the social situation, crying (which is a communication of helplessness) is subordinated by more complex adjustments of EE. One cannot usually so readily call on others for sympathy, if one is being coerced or excluded.20 Crying, as a form of emotional communication, is upstaged by a more direct emotional response in the form of fear and avoidance.
In social relationships, fear is generally a response to someone else’s anger. It is an anticipatory emotion, the expectation of being hurt. Thus it is most directly related to long-term emotional energy deriving from subordination on the power dimension. It occurs in similar circumstances to depression, but it has a more confrontational structure. Whereas depression is a withdrawal of EE (i.e., withdrawal of attention from particular activities), fear is a kind of social cringing before the consequences of expected actions. Depression is a sinking of EE level because of the bludgeoning effects of negative social situations;21 fear is a negative anticipation of what will happen, which assumes enough EE to take some initiative, or at least remain alert to situations that carry social dangers. Hence one can experience fear of status loss (membership exclusion), as well as fear of power coercion. On the power dimension, fear is mobilized together with anger in cases where a person is able to mobilize anger, but has low confidence in being able to win positive results from its expression.
Transformations from Short-Term Emotions into Long-Term Emotional Energy
The results of various short-term emotional experience tend to flow back into the long-term emotional makeup that I have called “emotional energy.” Emotional energy, though, does not have to depend upon the dramatic emotions; situations of uncontested domination or belonging add to one’s store of confidence and sense of attraction toward particular kinds of situations; undramatic feelings of subordination and unpopularity have similar negative effects. The dramatic short-term emotions also spill over, though it is an unexamined question whether their very quality as dramatic makes them more important for long-term emotions, or brackets them as a sort of exception. In the case of positive short-term emotions (joy, enthusiasm, sexual passion), it seems likely that these experiences should build up the store of EE, although perhaps in a very situation-specific way (i.e., one becomes attached to repeating just those situations with particular partners).
In the case of negative emotions, there is a long-standing clinical tradition that sees traumatic situations as the major determinant of longterm social and psychological functioning. Particular experiences of intense anger, fear, or shame are regarded as controlling one’s whole subsequent functioning. This may well be true, to a degree; but it should be seen against the background of the overall level of emotional energy. A person who generally has favorable, if undramatic, experiences on the power and status dimensions of their everyday interactions, will likely get over an episode of extreme anger, fear, or shame. It is only when the individual’s overall “market position” of interactions is on the negative side that particularly intense dramatic experiences are stored up and carried over as “traumas,” especially in highly charged memories of the sort that Freudian therapy is designed to ventilate. Max Weber’s conception of stratification as inequality of life chances in the market thus extends not only to material economic chances but to the realm of emotional health.
Scheff’s model reformulates Freudian theory as a carryover of emotions through an interactional chain. There is a shame / rage cycle in which an individual who experiences a shaming situation feels rage against the perpetrator, which can lead to further conflicts; these typically have unsatisfactory outcomes, resulting in further shame and rage. Rage at oneself can also become part of a self-reflective loop, intensifying this process. Scheff presents evidence that the traces of previous emotional arousals, especially anger, can remain at an unconscious, trace level; and that there are unconscious shame behaviors that are manifested in the micro-details of interactions. The limitation is that Scheff and Retzinger (1991) have chosen a sample of cases—couples in marital counseling—in which these shame / rage cycles are well established; but they have not considered the cases in which the cycle does not occur or quickly terminates. That is to say: Scheff concentrates on conflictual social relationships among individuals who are relatively equally matched, who are at the middle levels of dominance and popularity, such that they can continue long cycles of shaming and raging at each other. More extreme differences in power would not allow a conflictual cycle to go on; and if persons are not confined to the same network of status interactions (i.e., their market possibilities are more open) they may cut short a shame cycle by leaving that interaction and finding another where the resource lineups may be different.
THE STRATIFICATION OF EMOTIONAL ENERGY
IR chains often have a circular, self-perpetuating form. Persons who dominate rituals gain EE, which they can use to dominate future IRs. Persons who are at the center of attention gain EE, which they can use to convene and energize still further gatherings, thereby making themselves yet again the center of attention. In this way, powerful persons re-create their power from situation to situation, while those whom they dominate re-create the low energy level that makes them followers and subordinates. Status group leaders re-create the energy that makes them popular; groupies, fringe members, and outcastes are carried along in their positions by the repeated flow of lower EE.
Changes, of course, are possible chiefly if and when the composition of the persons encountering one another shifts, since in a perfectly closed cycle there would no way of getting out of a low-EE situation, or of failing to confirm one’s high EE. Thus even high-EE persons (like political leaders, or sociability leaders, or in specialized kinds of EE, sexual stars or intellectual dominants) move into an arena where they become overmatched by someone else with still greater EE (and hence become a medium fish in a bigger pond); and low-EE persons may find a different arena where they avoid old situational match-ups and find others that generate more solidarity (e.g., by graduating from high school). These are matters of how the entire array of IR chains, which makes up the population of a society, are arranged across time and space; and thus we widen out perspective to relatively more mesorather than micro-analysis.
We may visualize the stratification of society, not as a matter of who owns what material resources, or occupies what abstract position in a social structure, but as an unequal distribution of emotional energy. Positions in a social structure are macro-level abstractions; we can see stratification in a more empirically realistic way, as well as keep ourselves focused on its processual dynamics, by looking closely at exactly what stratification is enacted in micro-situations. Material “resources” are often repetitively available from one interactional situation to another, but what makes them “resources” hinges upon the micro-interactions that allow someone to appropriate them; and that is a question of who takes the initiative to take them and use them, and who passively accepts that these material objects are so used. Material property, as enacted in situations, is really the EE that particular persons have in acting upon those objects.22 Where the right to property is conceded, the distribution of emotions is asymmetrical, in that someone’s high EE in appropriating those objects is matched by someone else’s low EE in allowing them to be appropriated or at least standing by watching the other person display them. Similarly, Bourdieu’s “cultural capital” is too static a conception if it is taken merely as the counterpart in the hierarchy of culture to a hierarchy of economic capital.23 Another way to say this is that the key to stratification is not material property, nor cultural differences, but inequalities in emotional energy. It is the processual flow of EE that enables people to wield material and culture, or lets others wield those over them.
The simplest version of stratification is an energized upper class, lording it over a depressed lower class, with moderately energized middle-class persons in between. Take this pattern as an ideal type; it does yield a crucial point, that stratification generally works because those who dominate have the energy to dominate situations in which they encounter other persons. The winning generals are usually the most energetic ones; so are the richest financiers; in the specialized realm of intellectual domination, the stars of world science, philosophy, and literature generally are what I have called “energy stars” (for evidence on generals, see Keegan 1987; on philosophers, Collins 1998). To say this is not to make a moral judgment about any of these people: first, because what they are doing may well be manipulative, destructive, or selfish; and second, because their energy is not their own, in the sense that it arises interactively from chains of IRs, and thus from the network positions that put them into a positively accumulating, upward series of EE-enhancing encounters. My argument is far from holding that the upper classes are uniquely energetic individuals; they are products of processes that affect all of us, and in which all of us (very likely) are pretty much interchangeable. About any such dominant energy star, it is possible to say, there but for the grace of God (i.e., the luck of IR chain trajectories) go you or I. Dominant persons are not intrinsically heroes, but it is socially significant that they often appear as such. Persons with lower amounts of EE are impressed by those who have accumulated a lot of it; such people have an EE-halo that makes them easy to admire. They are persons who get things done; they have an aura of success surrounding them. And since having high EE allows one to focus attention, one can get a certain amount of rise in one’s own EE by following them, becoming part of their entourage, taking orders from them, or even viewing them from afar. Thus high EE gives dominant persons a kind of micro-situational legitimacy. This is not necessarily the same thing as the ideologies of legitimacy that Weber typologized (although it may undergird this formal legitimacy); I would hold that micro-situational legitimacy is by far the kind most worth having.
The stratification of EE thus makes the other aspects of stratification particularly solid and hard to dislodge. When the upper class has really high EE, no one even thinks of dislodging them, or even wanting to. That is, of course, an ideal type. A crucial point follows: A portion (perhaps a large portion) of what we conventionally call the “upper class” may consist of persons who have inherited their wealth, rest on their laurels from an earlier period of action, or otherwise not show very much EE. In such cases, the real distribution of EE differs from the formal, ideological conception of stratification. What we want to look for as sociologists is the real distribution of EE, and how it matches up against this surface appearance.
A perfectly self-reproducing stratification of EE is an ideal type. Patterns resembling this in degree have existed at various historical times. But these can break down in a variety of ways: some of these can shift very rapidly, since the mechanisms that generate EE are quite volatile, and conflict generates its own immediate patterns of EE. The mobilization of collective EE in social movements is a prime case of this. A stable hierarchy of EE can also break down in a different sense: not so much in the case of political action in which there are massive collective struggles, but in the case where energizing situations become fragmented. Instead of a hierarchy resembling the energized upper class, the depressed lower class, and a plodding middle class, we may get a purely local, episodically shifting situational stratification of EE in which almost any encounter is up for grabs. These topics will be taken up in later chapters.
APPENDIX: MEASURING EMOTIONAL ENERGY AND ITS ANTECEDENTS
It is sometimes raised as a criticism against IR theory that emotional energy is merely a hypothetical construct, or even a tautology. In reply, I wish to underscore the point that EE is an empirical variable.
We must be careful to distinguish EE from other kinds of emotions that are displayed. First, EE is not simply a matter of showing a lot of excitement, agitation, loudness, or bodily movement. These are characteristics of the dramatic or disruptive emotions: shouting or lashing out in anger, squealing and gesturing with joy, shrieking or running around in fear. EE instead is a strong steady emotion, lasting over a period of time, not a short-term disruption of a situation. A general characteristic of EE is that it gives the ability to act with initiative and resolve, to set the direction of social situations rather than to be dominated by others in the micro-details of interaction. And it is an emotion that allows individuals to be self-directed when alone, following a smooth flow of thoughts, rather than a jerky or distracted inner conversation. (For more detail on the latter, see chapter 5.)
Second, EE is a long-term consequence of IRs that reach a high degree of focused emotional entrainment, which we can also call attunement, collective effervescence, or solidarity; but EE is not the attunement itself. In figure 2.1, the ingredients and processes on the left side and middle of the diagram happen earlier than the outcomes on the right side; EE is a consequence that carries over after the individual has left the situation. Thus we must be able to measure it apart from the collective arousal itself. But it is also important for us to be able to measure the degree of collective effervescence or solidary entrainment within a situation, since this is the causal condition that produces EE.
Thus we wish to measure (a) the level of collective attunement reached at the height of an interaction, and see if it predicts the level of (b) emotional energy carried away by individual participants. With good measures of (b), we can also examine how long EE lasts, and test Durkheim’s proposition that EE fades away over a period of time unless a sufficiently intense ritual attunement is reenacted.
The following briefly overviews the different kinds of verbal and nonverbal phenomena that we can use as measures of EE, and as measures of the chief causal variable, situational attunement or solidarity. A clue is that attunement is a collective pattern, EE an individual one.
Self-report. I have defined EE as the continuum from enthusiasm, confidence, and initiative at the high end, down to passivity and depression at the low end. EE exists empirically in one’s flow of consciousness and in one’s bodily sensations: it is the most important item in one’s own everyday experience. It is not difficult to observe rises and falls in one’s own EE in different situations; with close self-observation, one can notice it rise or fall in a matter of seconds within any particular situation. Patterns of EE could be systematically studied by having individuals give reports on their subjective experience in various kinds of situations.
EE can also be measured objectively, by outside observers. Here the best measures are for the most part unobtrusive, although they do call for close observation of micro-details.
Bodily postures and movements. High EE is generally expressed in an erect posture, moving firmly and smoothly, and taking the initiative in relation to other persons. Low EE is indicated in postures and movements that are shrinking, passive, hesitating, or disjointed. Since high EE is social confidence, it is manifested in movements toward other people, especially movements that take the initiative and that lead to establishing a pattern of rhythmic coordination. Low EE, conversely, is found in movements and postures of withdrawal, and low initiative; low-EE persons in a social situation show a pattern of following others’ nonverbal leads, or a freezing of movement. Conflict at moderate levels of EE may be indicated by a rapid or jerky alternation between orienting toward and away from the others. Scheff and Retzinger (1991) describe this pattern, which they interpret in terms of the self-oriented emotions of pride (turning toward the other person) and shame (turning away).
We need to be careful to distinguish bodily measures of EE from those bodily movements that represent the process of collective entrainment within a social situation, although the one can lead directly into the other. High or low EE is visible in body postures and movements when an individual is alone. When an individual enters an interaction, EE is visible in the moments leading up to the high point of entrainment (whatever level that may be). That is to say, the high-EE person takes the initiative in setting the tone of the interaction, and the low-EE person lags behind or follows passively. EE must be observed in the dynamics of how the individuals lead or lag in the interaction, apart from the observation of how much entrainment finally results. This peak level of entrainment is a measure of collective effervescence.
At peak moments the pattern tends to be jointly shared among all participants: in high solidarity moments, bodies touch, eyes are aligned in the same direction, movements are rhythmically synchronized (see Figures 2.2, 2.3, 2.4, 2.5, 8.6, and 8.7). At moments of failure of the interaction, bodies turn away from each other, heads turn downward or inward toward one’s body, eyes look down or away. (For an example, see Scheff and Retzinger 1991, 54–56.)
Bodily measures also express the dramatic short-term emotions, which need to be distinguished from high and low EE more generally. On specific emotions, see Ekman (1984) and O’Sullivan et al. (1985), who also indicate the extent to which the body can be controlled so as to mask emotions, and which body movements tend to be involuntary and thus are genuinely unguarded expressions of emotion.
Eyes. Solidarity is directly expressed in eye contact. As Scheff and Retzinger (1991) show, persons in a situation of high attunement look at each other. This occurs in a rhythmic pattern, viewing the other person’s face, responding with micro-expressions, then periodically looking away (to avoid staring). In moments of intense solidarity (such as group triumph or erotic entrainment) the mutual gaze is longer and more steady. In a situation of low attunement, persons lower their eyes and turn away for prolonged periods. These are measures of high or low attunement or collective effervescence, and they tend to be symmetrical across participants. EE is seen in the eyes, as in the case of bodily postures and movements, as a temporal pattern for each individual as they approach the situation. Initiative or lack of initiative can be seen in establishing eye contact; high or low EE is manifested in dominating or avoiding mutual gaze (Mazur et al. 1980; Mazur 1986).
Voice. The amount of enthusiasm, confidence, and initiative (high EE) versus apathy, withdrawal, and depression (low EE) can be measured paralinguistically, that is, in the style rather than the content of talk. (See Scherer 1982, 1985, for studies of the emotional dimensions of recorded speech.) Since the flow of speech in an interaction is also a measure of the amount of attunement or collective solidarity, we must be careful to observe in micro-detail the patterns of the individuals as they approach the vocal interaction, as distinguished from the degree of attunement that is reached collectively.
A refined study that separates out these several aspects is Erickson and Shultz (1982: see especially 85–96, 103–117). This study demonstrates measures of voice rhythms charted at twenty-four frames per second, but typically visible at quarter-second intervals. These fall into five patterns: (i) a shared rhythm, with the beat falling at about onesecond intervals; this may be interpreted as normal solidarity; (ii) “individual rhythmic instability,” as one individual follows the previously set mutual rhythm while the other is momentarily disorganized: an indicator of dominance or interactional centrality on the part of one speaker in relation to the other; (iii) mutual rhythmic instability, as both speakers slow down or speed up the rhythm for a brief period before returning to the baseline rhythm: a temporary failure of the interaction ritual: a display of low solidarity; and (iv) mutual rhythmic interference and (v) mutual rhythmic opposition, which are two types of micro-interactional conflict: in (iv) the conflict is ongoing, whereas in (v) a dominance struggle is won by the speaker who overcomes the rhythm of the previous speaker and gets acquiescence to the new rhythm.
The characteristics of (iv) and (v) are worth quoting directly from the authors:
Mutual Rhythmic Interference [iv] A kind of mismatch between the behavior of one individual and the other, lasting a few moments, involving the persistence by each party in rhythmic patterns that are regular for each individual but different across individuals, for example individual A’s behavior is patterned in a rhythmic interval of 1 second duration, while individual B’s behavior over the same period of time is patterned in a rhythmic interval of .75 second duration.
Mutual Rhythmic Opposition [v] …. Momentary rhythmic disintegration between the behavior of one individual and the other, involving deviation of 4–5 twenty-fourths of a second from the previously established periodic interval. Coming in this much too soon or too late at turn exchange has the effect of “tugging” at the underlying rhythm. This tugging is seemingly competitive; at the very least it indicates a lack of cooperation or integration in the mutual behavior of speakers, since one speaker does not participate in the rhythm used by the previous speaker. After the momentary tug occurs, however, the previous speaker adapts to the new rhythmic interval, and so the lack of temporal integration between them involves momentary opposition rather than continuous interference (Erickson and Schultz 1982, 114–15).
The troubled moments (ii, iii, and v) also tended to coincide with shifts in body posture and proxemics, or, changes in body orientation between the speakers.
These voice rhythms thus show variations in solidarity, as well as fine-grained indications of who sets the rhythm and who follows it. For our purposes, (ii) and (v) indicate taking the initiative and setting the pattern, which are indicators of EE—high EE for the individual who sets the rhythmic pattern, low EE for the individual whose rhythm is determined by the other. Pattern (i) is an indicator of high solidarity; (iii) and (iv) are indicators of low solidarity.
Measures of interactional solidarity are also available from the ultramicro analysis of the sound-wave frequencies at subliminal levels, using Gregory’s (1994; Gregory et al. 1993) Fast Fourier Transform (FFT) analysis. FFT analysis finds rhythms of vocal coordination at a level below .5 KHz (KiloHertz, thousand cycles per second), a region in the sound spectrum that is heard only as a low-pitched hum. Although participants are not aware of the sounds they are making on this level, their voice rhythms converge in conversations that they subjectively rate as more satisfactory interactions with a higher level of rapport.
Comparing Gregory’s measures at the level of thousands of cycles per second and Erickson and Schultz’s measure at the level of quarterseconds, it is apparent that several levels of rhythmic coordination overlay each other, at different orders of time-frequency. The relationships among these different time-orders are yet to be investigated, as well as their connection with IR ingredients and outcomes.
Other indicators of conversational attunement or solidarity have been displayed in chapter 2: a close pattern of turn-taking with minimal gap and overlap; rhythmic entrainment in shared laughter, applause, and other simultaneous vocalizations. Conversely, gaps between turns, and prolonged overlaps among speakers contesting the floor, indicate low solidarity.
Indicators of conversational solidarity are easier to tease out than indicators of EE, since the latter involves showing who takes initiative in establishing the pattern of the interaction. Some aspects of individual voices are probably not good measures: loudness of tone and speed of talking are too easily confounded with specific disruptive emotions such as anger. Better indicators of EE are fluidity, hesitation pauses, and false starts on the part of each particular individual. Ability to get the floor, versus incidence of contested speech turns, is another indicator; methods are demonstrated in Gibson (1999).
Hormone levels. Mazur and Lamb (1980; see also Kemper 1991) have shown that the experience of dominating an interaction has continuing effects upon hormone levels (especially testosterone). These hormones may provide a physiological substrate for medium-run flows of EE across situations. It should be noted that testosterone is found in females as well as in males, although in lower amounts (Kemper 1991); hence the pattern could be operating for both sexes. The important comparison is for shifts in hormone levels within the same person across situations, not necessarily for relative hormone levels across different individuals. Studying hormone levels requires intrusive measures, which are especially intrusive when this is done by drawing blood samples, and thus such studies have been done largely by volunteers who are trained medical personnel; saliva measurements have also been used. It would be worth seeing how shifts in hormone levels relate to shifts in other measures of EE. It is not clear whether EE shifts are related to absolute or relative levels of testosterone, and of other physiologically active substances. In any case, whatever physiological substrate is involved must interact with the cognitive components by which EE is carried along as a propensity to respond positively or negatively toward particular kinds of interactional situations; and with the level of mutual focus and emotional entrainment that constitute the immediate process of social action.
Facial expression. I do not place emphasis on facial expressions as indicators of EE. Ekman and Friesen’s (1975/1984, 1978) manual shows the ways in which specific emotions are expressed in the several zones of the face, such as joy, anger, fear, sadness, and disgust. But these are indicators of the short-term, disruptive emotions. It is not clear that there are specific facial indicators for high and low EE. It is possible that facial measures of EE could be developed. High EE should be found in facial expressions of confidence and enthusiasm; low EE as expresssions of apathy and depression. These should be distinguished from facial indicators of momentary happiness and sadness, since high and low EE should be prolongations across situations.
Even if facial measures are not the best way to measure EE, I would urge microsociologists to study Ekman’s facial indicators of emotions and make use of them in situational observations; they provide useful auxiliary information, and may show patterns of short-term emotional expression that are related in various ways with flows of EE across situations. Ekman’s research (1984) is valuable also because it indicates which zones of the face are most easily controlled by deliberate efforts to mask emotions, while other zones tend to express spontaneous emotions.
It would be useful to study all or several of these measures simultaneously. Especially worthwhile would be to compare each of the objective measures—body posture and movement, eyes, voice, etc.—with self-reports of high or low confidence and initiative. Arriving at objective measures is desirable insofar as they are less intrusive, and thus easier to use in observational research. The result of such multi-measure studies should be to show which measures are redundant, and which are most highly correlated with long-term patterns (i.e., with the flow of EE across situations).
The two approaches to measuring EE—subjective self-observation and objective observation of other people—may also be used together. If subjective measures are pursued, persons can become better self-observers by training in objective measures, enabling them to attend to one’s own bodily sensations, movements, and postures in detail, as well as to those around them.
What I would like to stress, in using either subjective or objective measures, is that these processes always happen in micro-interactional situations; the level of EE should always be studied in relation to the kind of situation that is occuring at the moment, and within the chain of situations from the immediate past. It is less useful, in using subjective measurements (such as questionnaires, interviews, or time-diaries), to ask for a global assessment: “how much enthusiasm, confidence, and energy (or depression, apathy) have you been experiencing in your life?” Such information gives an indication of the overall drift of situational outcomes, but it is more valuable to be able to show what the situational conditions are in which these observations take place.24
To study shifts of EE in real-life situations, it would be desirable to follow people’s experiences across a chain of interactions. A mediumterm design would be necessary. Possibly this could be constructed in a laboratory situation lasting several days. Observation in natural conditions would also be desirable, especially to estimate how long emotional effects of interactions may last. I suspect, however, that the time-decay of emotional energy, if it is not reinvested and reinforced by subsequent interactions, may be less than a few days.