You are on page 1of 366

CONTENTS

LIST OF CONTRIBUTORS vii


PREFACE
Jonathan H. Turner ix
INTEGRATING EMOTION INTO IDENTITY THEORY
Sheldon Stryker 1
IDENTITIES, EVENTS, AND MOODS
Peter J. Burke 25
EMOTIONS IN IDENTITY THEORY: THE EFFECT OF
STATUS
Jan E. Stets 51
PHYSIOLOGICAL MEASURES OF THEORETICAL
CONCEPTS: SOME IDEAS FOR LINKING DEFLECTION
AND EMOTION TO PHYSICAL RESPONSES DURING
INTERACTION
Dawn T. Robinson, Christabel L. Rogalin
and Lynn Smith-Lovin 77
VIOLENT MALES: A THEORY OF THEIR
EMOTIONAL/RELATIONAL WORLD
Thomas J. Scheff 117
EMOTIONS, SENTIMENTS, AND PERFORMANCE
EXPECTATIONS
Robert K. Shelly 141
v
vi
THE ENHANCEMENT IMPERATIVE AND GROUP
DYNAMICS IN THE EMERGENCE OF RELIGION AND
ASCRIPTIVE INEQUALITY
Michael Hammond 167
TOWARDS A UNIVERSAL POWER AND STATUS THEORY
OF EMOTION
Robert Thamm 189
THE DIFFERENTIAL IMPACT OF EMOTIONS ON
RATIONAL SCHEMES OF SOCIAL ORGANIZATION:
READING WEBER AND COLEMAN
Theodore D. Kemper 223
CONSCIOUSNESS, EMOTIONS, AND SCIENCE
Jack Barbalet 245
A THEORY OF THE SELF, EMOTION, AND CULTURE
Erika Summers-Efer 273
DEFENSIVE STRATEGIES: THE FORMATION AND SOCIAL
IMPLICATIONS OF PATTERNED SELF-DESTRUCTIVE
BEHAVIOR
Erika Summers-Efer 309
EMOTIONS AS COMMENTARIES ON HUMAN CONCERNS
Margaret S. Archer 327
LIST OF CONTRIBUTORS
Margaret S. Archer Department of Sociology, University of
Warwick, UK
Jack Barbalet Department of Sociology, University of
Leicester, UK
Peter J. Burke Department of Sociology, University of
California, Riverside, USA
Michael Hammond Department of Sociology, University of
Toronto, Canada
Theodore D. Kemper Department of Sociology, St. Johns
University, USA
Dawn T. Robinson Department of Sociology, University of Iowa,
USA
Christabel L. Rogalin Department of Sociology, University of Iowa,
USA
Thomas J. Scheff Department of Sociology, University of
California, Santa Barbara, USA
Robert K. Shelly Department of Sociology, Ohio University,
USA
Lynn Smith-Lovin Department of Sociology, Duke University,
USA
Jan E. Stets Department of Sociology, University of
California, Riverside, USA
Sheldon Stryker Department of Sociology, Indiana University,
USA
Erika Summers-Efer Department of Sociology, University of
Pennsylvania, USA
Robert Thamm Department of Sociology, San Jose State
University, USA
vii
PREFACE
My goal in this volume is to communicate the range and diversity of work in the
sociology of emotions. Most of the papers fall within the normal parameters of
the Advances in Group Process series, but I have also included several articles
that normally would not be part of the series. My intent is, in essence, to stretch
the denition of group to larger-scale social structures, although the focus is on
how culture and social structure interact with the activation of emotions in people.
I have also sought to stretch the study of emotions in the opposite direction:
toward the physiological processes generating emotional responses. There is a
clear bias toward theoretical papers, although several articles report research
ndings as they bear on a particularly theory.
The rst three papers revolve around identity dynamics and emotional arousal.
In the rst article, Sheldon Stryker expands upon his earlier theorizing on
emotions. After placing into context his theoretical approach, Stryker offers a
series of propositions that connect identity processes (salience, commitment, and
role performance) to emotional arousal as it ows from role performances and
from networks of relationships. In the second paper, Peter Burke presents data
on the effects of identity-disrupting events on mood and the unease/distress that
individuals experience. And, in the third paper of this set, Jan Stets presents data
on how diffuse status characteristics effect emotions in identity processes.
The next paper by Dawn Robinson, Christabel Rogalin, and Lynn Smith-Lovin
is also in the symbolic interactionist tradition but seeks to incorporate the
physiological dimensions of emotions into affect control theory and, by extension,
the identity theories of Stryker and Burke. After reviewing the theoretical tradition
from which affect control theory emerged and the various efforts to incorporate
the biology of emotions into theory and research, these authors offer a provisional
mapping of physiological measures that can be employed today in experimental
studies of affect control and identity theories.
The article by Thomas Scheff also draws from the symbolic interactionist
tradition but adds ideas from psychoanalysis. Scheff presents a theory of male
violence, seeing violence as an outcome of repressed shame. Scheff then uses
this theory to develop a more macro theory of how political leaders are able to
mobilize whole populations to collective violence.
In Robert Shellys paper, emotional dynamics are analyzed within
the expectation-states theoretical tradition. Building upon his earlier
ix
x
conceptualizations of sentiments, Shelly argues that sentiments can oper-
ate, by themselves, as a force in forming expectation states and in the process of
status differentiation.
The next paper by Michael Hammond is one of those that might not typically
nd its way into Advances in Group Processes. In this article, Hammond argues
that larger-scales social structures such as religion and stratication can be un-
derstood as outcomes of biologically based imperatives to overcome habituation
through emotional arousal.
The article by Robert Thamm extends Theodore Kempers power and status
theory into a theory about how the expectations and sanctions operate to change
power/status relations and the emotional reactions of individuals to such changes.
Thamms paper is followed by an article from Theodore Kemper who develops a
critique of models of rationality that fail to recognize the importance of emotions
in generating commitments to the organizations and institutional systems of
human societies.
Jack Barbalets paper shifts the unit of analysis to the institution of science,
arguing that emotions are an important for the activities of scientists. Emotions
operate non-consciously to motivate scientists and to frame objects of study. In de-
veloping this more institutional viewpoint, Barbarlet explores the phenomenology
of consciousness and non-consciousness as they inuence behaviors.
The next two papers are by Erika Summers-Efer. The rst seeks to synthesize
elements of existing theories and research into an entirely new theory of the rela-
tionship among emotions, self, and culture. The second of these two papers extends
Randall Collins interaction ritual theory to account for how individuals, who do
not posses resources to enhance emotional energy, will try to minimize the loss of
emotional energy and, as the do so, will sustain self-destructive social relationships.
Margaret Archer expands upon her earlier theory by emphasizing that clusters
of emotions represent commentaries on the concerns and problems of social
relationships that are evident in three orders of reality: the natural, practical, and
discursive. Because of the reexivity provided by the discursive domain of reality,
humans develop second-order emotional meanings that provide further clues to
the concerns of individuals in society.
These papers collective provide a sense for the range of work in the sociology
of emotions. It is now clear that the study of emotions is at the forefront of not
just micro sociology but more macro-level approaches as well. No one volume
can capture the full range of work, but these article demonstrate how vibrant the
sociology has become over the last three decades.
Jonathan H. Turner
Editor
INTEGRATING EMOTION
INTO IDENTITY THEORY
Sheldon Stryker
ABSTRACT
Of contemporary interest in the sociology of emotions and social psychology
of self is the question of the reciprocal relation of affect or emotion and self.
This question is pursued by asking how affect or emotion impacts the identity
theory variables of commitment, identity salience, and role performance
as well as by asking how these identity theory variables impact persons
affective responses. Brief reviews of the general literatures on emotion,
the symbolic interactionist literature (from which theoretical frame identity
theory derives) on emotion, identity theory itself, and characteristics of social
sentiments and emotional outbursts allows the development of expectations
about the interrelation of sentiments and emotional outbursts and identity
theoretic variables.
INTRODUCTION
Important theoretical and empirical problems drawing attention in the sociology
of emotions and in the social psychology of self concern the reciprocal relation
of affect or emotion
1
and self. This assertion holds across social psychological
frameworks and theories incorporating a concept of self;
2
it holds in particular for
the symbolic interactionist frame and related theories.
3
Such problems are at the
Theory and Research on Human Emotions
Advances in Group Processes, Volume 21, 123
2004 Published by Elsevier Ltd.
ISSN: 0882-6145/doi:10.1016/S0882-6145(04)21001-3
1
2 SHELDON STRYKER
heart of Hochschilds (1979, 1983) accounts of emotion norms and management,
especially the self-alienating costs of the latter. They are central to Thoits (1985,
2003) account of howpersons sense their affective responses deviate fromexisting
emotion norms and how this sense enters self-labeling processes. They underlie
the development of Affect Control Theory (Heise, 1979; Smith-Lovin & Heise,
1988), are integral to Burkes (1999; Burke & Stets, 1999; Cast & Burke, 2002)
arguments dealing with the relation of identity disruption and stress, and are
pervasive in the interactionist literature on the social construction of affect and the
socialization of affective responses (Cahill, 1999; Denzin, 1984; Gordon, 1981,
1989; Hafferty, 1988; Lively, 2002; Pollak & Thoits, 1989; Shibutani, 1961;
Shott, 1979; Taylor, 2000).
4
I approachthese issues fromaninterest insymbolic interactionismandinidentity
theory, seeking to integrate affect into identity theory.
5
A quick review of identity
theory is a necessary prelude to the attempted integration. I next briey examine
the general literature on emotions in psychology and sociology, afterward looking
at the symbolic interactionist and identity theory literatures, asking what these
literatures in particular may say that is relevant to my problem. I then turn to
relating affect and identity theory. As previously put, the paradigmatic question
identity theory seeks to answer is: why, on a free weekend afternoon, does one
man choose to take his children to the zoo and a second choose to play golf with
his buddies?
Identity theory (Stryker, 1968, 1980,
6
2000b, 2001) derives from Meads
(1934) formula (paraphrased), society impacts self impacts social interaction. I
argued each term of that formula had to be specied to allow testable predictions.
I claimed that role-choice is signicant for interaction, warranting explanation.
I noted contemporary sociology conceptualized society as highly differentiated
yet organized, a view that, by Meads dictum, self reects society, requires a
highly differentiated yet organized concept of self. I offered a concept of self as
consisting in part of multiple identities, internalized meanings attached to roles in
many role relationships, which identities are organized in a salience hierarchy.
7
I noted that in contemporary society people do not live their lives in society as
a whole but in multiple smaller or small networks. I suggested that the salience
of an identity depends on commitment, the strength of persons ties to networks
in which they have roles and related identities. So specied, Meads formula then
reads: Commitments impact identity salience impacts role-related behavioral
choices. The papers arguments were initially developed in Stryker (1968) and
further developed over the years (e.g. in Ervin & Stryker, 2001; Stryker, 1980,
1986, 1989, 1994, 2000a). Tests have supported the basic theory (e.g. Serpe,
1987; Serpe & Stryker, 1987; Stryker & Serpe, 1982, 1994).
Integrating Emotion into Identity Theory 3
A QUICK LOOK AT THE GENERAL
LITERATURE ON EMOTION
There are nothing but conceptions of emotions, identifying them at one extreme
as evolutionary bred, instinctive physiological reactions located in the viscera or
facial musculature (Ekman, 1982; Izard, 1969), at the other identifying them as
totally socially constructed or determined by culture (Averill, 1980; Denzin, 1990;
Illouz, 1997; Rosaldo, 1980, 1984). Current social science usually takes an eclectic
view, regarding emotions as a conjoint of social and more general environmental
conditions, physiological reactions, social denitions, and experiential responses
(Gerth & Mills, 1953; Gordon, 1981; Kemper, 1987; Lazarus, 1992; Scheff, 1983;
Thoits, 2003).
The literature asserts that emotions are adaptive in the interactional short term
(Cahill, 1999; Clark, 1990) or evolutionary long term. Historically, emotions were
for some epiphenomenal or evolutionary anomalies destined to disappear with the
triumph of evolution-bred rationality (see Klinnert et al., 1983 for a description
and critical evaluation of such views).
8
The predominant current view argues
the adaptational value of negative emotions such as hate (e.g. Weinrich, 1980).
Darwins (1872) idea that some emotions (e.g. fear) have survival functions,
warning organisms of endangering predators, was sharply asserted by Freud
(e.g. 1948) in his argument that emotions warn persons of intolerable internal or
external mental health dangers. Hochschild (1983) generalizes Freuds treatment
of anxiety as having a signal function to the full range of emotions.
The literature asserts the adaptation functions of emotional expression both
for persons who experience and express emotion and for persons witnessing
the expression. People read their emotional responses as guides to their further
performances (e.g. Rosenberg, 1987) and they read emotional reactions of others
in interacting with those others (e.g. Klinnert et al., 1983). The extensive literature
on emotion management of self and others (e.g. Lively, 2000, 2002; Thoits,
1996), again with its source in Hochschilds work, testies to the validity of
this observation.
Partly explicit and partly implicit in this literature are characteristics of emotion
potentially relevant to its integration into identity theory. Prime is the centrality of
self in considering emotion. For some (e.g. Denzin, 1983) the linkage of emotion
to self is, in essence, a matter of denition. However, the relevance of self for
emotions goes far beyond matters of denition. Emotion theorists working from
a various perspectives argue that emotions depend on the implications others are
perceived to have for ones personal survival and welfare, needs, goals and plans.
Thus, Hochschild (1983) regards her treatise on the commercialization of feelings
4 SHELDON STRYKER
as an exploration of the idea that emotion functions as a messenger from the self,
more specically that emotions are messages from self reporting the connection
between what one observes in the world and what one expects to observe along
with what one is ready to do about any discrepancy. Berscheid (1983) observes
that since Darwin emotion theorists regard emotion as depending importantly if
not exclusively on the implications others are perceived to have for ones survival
and welfare, needs, goals and plans. Arnold (1960) suggests that the appraisal of
an event in term of its implications for the appraisers well-being is what instigates
emotional expression. Plutchik (1980) take the problem of learning who we are
and what groups we belong as one of four universal problems of adaptation in
a psychoevolutionary theory of emotion. Higgins (1987, 1989) hypothesizes that
particular emotions reect discrepancies among persons and others perceptions
of actual, ideal and ought selves.
EMOTION IN SYMBOLIC INTERACTIONISM
Into the 1980s a criticism of the symbolic interactionist frame was that it
downplayed or neglected entirely the import of affect in social life (Gordon,
1981; Meltzer et al., 1976; Shibutani, 1961; Stryker, 1980). The criticism reects
the frames stress on the construction and communication of signicant symbols
as the basis for social interaction and the correlative image of human actors as
building, testing, and selecting from alternative hypotheses behavior linked to the
solution of interactional problems. The complaint, to the extent it was justied,
followed from Meads overwhelming role in shaping interactionist thought. There
is little warrant in Mead for a serious interest in emotion. His few references to
the topic in his writings (more accurately, his posthumously edited and published
lecture notes) include an argument contra Darwin that emotions arise in the course
of social conduct rather than expressing conditions prior to and independent of
that conduct, reected in Blumers (1969) treatment of human attitudes. Another
makes the argument that emotional expressions are not fully communicative
because they typically do not arouse the same response in persons emitting them
and others and so are essentially not signicant symbols.
However, the 18th century Scottish Moral Philosophers, precursors of symbolic
interactionism, did attend to emotion. David Hume (1748) and AdamSmith (1759)
understood that linking inner feelings to the fate of others was key to effective
social control, and Smiths doctrine of sympathy asserted the social importance of
sharing the feelings of others. The early interactionist Cooley (1902) argued the
transformation of instinctual bases of emotion through social life, adopted Smiths
concept of sympathy in analyzing the communication that is society, and
Integrating Emotion into Identity Theory 5
promoted sympathetic introspection as the proper method of sociology. Symbolic
interactionists of the mid-1900s emphasized ways in which emotional arousal
and expression reected requirements of social situations and are responsive to
denitions of the situation (Lindesmith & Strauss, 1949). Gerth and Mills (1953)
dealt with emotions through linking feelings with socially recognized gestures and
with self. Shibutani (1961) treated the transformation of emotional arousal into
social sentiments through the development of sustained feelings in interpersonal
relationships. The failures in the social control of emotions are a major theme
in Blumers (1951) analyses of elementary collective behavior. The interactional
bases of particular emotions discrediting identities in the case of embarrassment
(Goffman, 1967; Gross & Stone, 1964), inability to behave in accord with
ideal self conceptions in the case of shame (Riezler, 1943) were argued in
interactionist analyses.
From such beginnings, symbolic interactionist considerations of emotion as
product of socially located interaction and as social constructions, as well as
having personal, interpersonal and more general societal consequences, exploded
in the last quarter of the 1900s and continues apace at the present.
9
One of the few
systematic propositional analyses of any kind using the symbolic interactionist
frame as its grounding was produced early in this period by Shott (1979) in her
theory of emotion. That theory accepts the basic premise that emotions are the
outcomes of societal and social interactional processes, develops the concept of
role-taking emotions, and then goes on to anticipate social outcomes of emotional
states based on role-taking processes. Scheff (2000, 2003) addresses both the
failure to follow through on earlier work on the import of shame in society
and seeks to stimulate a revival of that interest. In her work on emotion norms
and management, Hochschild (1979, 1983), while focusing on the personal and
social consequences of commercialization (by which she intends the institutional
invasion of the formerly private albeit social processes of emotion management),
recaptures the idea impregnating the evolutionary literature on the emotions,
the idea of emotion as signal. She thus comes to the discrepancy that can occur
between emotion norms and emotional responses, and to emotion management as
the resultant effort to bring the two into line with one another. In that discussion
is embedded the idea that expressions of emotion signal to others who and what
the person expressing the emotion is, that is, signal aspects of self pertinent to
interaction. Also embedded here is the idea, not well developed, that persons
emotional responses signal to themselves whom or what they may be.
The idea that conversations of self and other are internalized and become con-
versations within the self is ingrained in interactionist thought. That idea becomes
explicit in three important theoretical efforts premised on a symbolic interactionist
framework: Heises Affect Control Theory (1979) in which discrepancies between
6 SHELDON STRYKER
or among the affect attached to self, others and behavior relating self and other
motivate attempts to restore affective balance; Burkes (1996, 1997; Stryker &
Burke, 2000) stress on an equivalent control process that operates to bring persons
behavior into line with their identity standards; and Thoits (1985) theory relating
emotional deviance to self-labeling processes in which repeated self-perceived
deviations from emotion norms are seen to lead to persons labeling themselves
as emotional deviants in need of professional help. Gordon (1989) suggests,
based on Ralph Turners (1976) distinction between institution and impulse as
contrasting bases of self-denitions in contemporary society, that this variation
will affect persons interpretations of events and therefore their emotional
responses to the events.
Also underwritten by the idea that conversations of self and others are inter-
nalized is a comparatively recent thrust in the symbolic interactionist literature
that has to do with the relation of affect and mental health, namely an emphasis
on persons as psychological activists (Thoits, 1994). If internal conversations
can be seen as involving self and others, they can also be seen as involving
alternative selves, either earlier and later or deriving from alternative sets of
others at a given point in time. As I have argued (Stryker, 1994), the possibility of
adopting alternative standpoints in an internal self dialogue in turn makes possible
creative responses that alter life circumstances generating negative affect
changing jobs in order to spend more time with family is an example (see Kiecolt,
1994; Kiecolt &Mabry, 2000 for more general suggestions about howpersons may
effectively initiate changes in self). Such possibilities are argued by Thoits (1994,
2003), in work asserting the positive emotional potential of role accumulation that
starts with the premises that roles and related identities give existential meaning to
persons lives and provide resources to deal with problems that life may present.
An important distinction in the symbolic interactionist literature is that which
differentiates between routinized and socialized affect, the socially organized
affective responses that have been called sentiments (Cooley, 1902; Gordon, 1981;
Shibutani, 1961) and acute emotional outbursts that sometimes occur in social
interaction and that have received considerable recent attention.
10
The distinctions
between the two aspects of affect are not always as clear as their abstract denitions
suggest. Clearly, both are signicant in social behavior, and both are incorporated
into the following discussion of the relation of affect to identity theory.
AFFECT IN IDENTITY THEORY
Identity theory derives from a structural symbolic interactionism (Stryker, 1980,
2001), a frame that joins Meads premises to a more adequate sociological
Integrating Emotion into Identity Theory 7
conceptualization of social structure and its role in self processes than has been
typically entailed in symbolic interactionist formulations. Despite the attention
given social structures, the theory is in a fundamental sense a cognitive theory. The
cognitive character of the theory is seen in particular in its conceptualization of self
as a structure of identities, with identities taken to be cognitive representations of
the positions in which persons are embedded and the roles they play in social life,
that is, they are internalized meanings of structural positions in the form of role
expectations. When the theory was initially presented (Stryker, 1968), self was
viewed as a structure composed of three modalities conative (I want), cathectic
(I feel), and cognitive (I am) and the interdependencies of these modalities was
postulated. Thus, the concept of self itself incorporated an affective component
equivalent to the cognitive component of identities (reasserting the Affect Control
Theory axiom that affective meaning is part of the meaning of all social objects),
although attention was then focused on the cognitive component. Too, empirical
evidence forced an early recognition of the relevance of affect to identity theory
predictions by insisting on an independent contribution of affective commitment
to the hierarchical ordering of identities (Serpe, 1986, 1987; Stryker, 1987; Stryker
& Serpe, 1982). Stryker and Statham (1985) recognized the signicance of role
distancing, role involvement, and role satisfaction, concepts that all suggest the
emotional freight carried by identities (Serpe, 1987; Stryker & Serpe, 1982).
One more element in identity theory thinking recognizes the import of affect.
While it need not, the concept of role expectations built into the theorys vision
of social structure as well as self empirically includes the concept of norm; that
concept implies that expected behavior typically generates or reects feelings.
Others, and the actors related to those others,
11
care whether or not expectations
are met, and the meeting or failures to meet expectations generate more or less
strong and diverse forms of affective expression. Social structures, as patterned
regularities of social behaviors, as well as the identities incorporated into self
are in part constituted by affect. Again, this idea is obviously fundamental to
Affect Control Theory. It is equally fundamental to various past efforts to measure
identity in Schwartz and Strykers (1970), Burke and Tullys (1977), and Burke
and Reitzes (1981) early work as well as in Burkes more recent efforts to model
the internal dynamics of self-processes (Burke, 1999; Stryker & Burke, 2000).
A few further comments on symbolic interactionist treatments of affect help
move the discussion to the issue of the relation of emotion to identity theory
variables. First, an intrinsic relation between affect and self does not obviate affect
based in empathy or altruism, if for no other reason than that others are or can
become extensions of self (Rosenberg, 1979). Some affect, however, connects to
self in special ways: it is generated by afliations or denials of self; it provides
information about who we are and where our priorities lie; and it supplies an
8 SHELDON STRYKER
important part of the meanings for us of events we experience (Scheff, 1988; Shott,
1979; Simmons, 2001). Second, the assertion that affect is intrinsically linked to
self does not deny the signicance of the self-other relationship; indeed, it afrms
that signicance in a sense that goes deeper than is frequently recognized.
A basic tenet of a symbolic interactionist frame is that persons have selves
insofar as they respond to themselves as objects, and they respond to themselves
as objects primarily through internalizing the responses of others with whom they
interact.
12
Interactionists have long recognized that society enters sentiments,
with some largely ascribing to society the role of situational stimuli for triggering
affective responses and others regarding society as the supplier of collective
denitions of appropriate affect to be expressed, appropriate times to express
particular affect, and appropriate modes of expressing this affect. However, so
conceiving affect does not recognize the full signicance of the relationship.
For the symbolic interactionist, society is a congeries of self-other relationships;
thus a major link of society and affect is its role in self-other relationships.
Simply put, affective concomitants, whether in the form of sentiments or affective
outbursts, and whether these are expressed in the form of calm self-satisfaction,
smiles, laughs (see Francis, 1997; Francis et al., 1999 on the use of humor in
medical interactions), cries, muscle tensing, screams, or whatever, serve to inform
others about self, providing communication that makes self-other coordination,
ultimately society, possible (see Clark, 1990, 1997 on the role of sympathy in
producing solidarity).
13
This does not imply that reading affective expressions in terms of what they say
about the self of the person providing them is immediately given or unambiguous,
any more than the reading of symbols of any kind. Sentiments and emotions are
emergents from interaction as surely as they are contributors to interaction: they
serve as premises in interactive next steps, both shaping interaction and available
for reformulation as interaction unfolds. That there is sufcient accuracy in the
use of affective expressions in communication is testied to by the evolutionary
and developmental import of affect.
If affective expressions serve as signals to others and others premise their
actions toward one on the basis of their reading of the self implications of those
expressions, persons can come to read the self-implications of their own affective
expressions. The self-other dialectic becomes internalized and reproduced. Where
we are led, then, is to a view of at least some affect as peculiarly related to self in
the sense of communicating to persons important information about their selves.
Sentiments and emotions have signal functions, not only to others but to self; they
are messages from the self but also to the self, informing persons sometimes
complete with surprises about the strength of their commitments, the relative
salience of their identities, about who they really are. It is reasonable to expect
Integrating Emotion into Identity Theory 9
that, as Hochschild work (1983) suggests, it is non-normative emotional reactions
to others or ones own actions that are especially powerful in this respect.
Just as reading affective information from another is not straightforward, so
too reading information about oneself may be difcult, confusing, misleading.
Nevertheless, our experiences of affect can inform us about our selves, and the
information will be given special credence because of the nature of the experience
(Denzin, 1983). Meanings associated with the experience of affect can be
powerful deners as well as organizers of self, consequently powerful organizers
of subsequent action. This is true of sentiments, and it may be particularly true
of the acute emotions.
DIMENSIONS OF THE
EXPERIENCE OF EMOTIONS
Why so? What is the nature of the meanings contained in experiencing acute
emotions that may make the information they convey regarding self of particular
signicance? That question leads to asking how acute emotions are experienced.
Recall the special place subjective responses have in the conceptual structure
of symbolic interactionism; it is from these that rather than from the objective
character of the world that much social behavior is presumed to ow.
14
Acute emotional experiences obviously vary on a positive-negative dimension
and on intensity, or weak-strong, dimension. Occasionally becoming a central
focus of emotion theorists (e.g. in Tomkins, 1963) is the idea that emotions
amplify both the stimuli that trigger emotion and subsequent behavior as well.
Emotion, in these terms, strengthen the effects of stimuli on responses beyond
the ordinary, i.e. beyond what others or the person experiencing them regard as
normal. Thus, the signals conveyed by the experience of emotion to the self are
enhanced, and are brought starkly to attention. Amplication may be involved
in what the cognitive psychologists (e.g. Markus & Zajonc, 1985) refer to the
availability or accessibility of cognitions. Alternatively, it may be involved in how
acute emotions focus or monopolize consciousness.
In any event, emotions are often experienced as non-routine, as out-of-the-
ordinary; and it is in part from their out-of-the-ordinary quality that the meanings
of emotions are drawn. (On this point, it is important to remember
15
that
normatively required affect an example is a parents love for a child can also
be experienced as out of the ordinary.) Indeed, the experience of emotions as out
of the ordinary in itself constitutes part of the meaning of emotion, informing
people that they are being signaled about something important, something to
which they need pay attention. In work on self, it is perhaps McGuires (1973;
10 SHELDON STRYKER
McGuire & Padawer-Singer, 1976) that most clearly suggests the signicance of
this dimension of emotion for constituting self and for ordering the salience of
identities; but the lesson is available through much of the literature on perception
and cognition.
The meaning inherent in the experience of acute emotion does not stop here.
Two further aspects of such experience reinforce and extend the meanings that
can be read from emotional responses are: (1) acute emotions are experienced as
immediate, unmediated, and spontaneous; and (2) acute emotions are experienced
as outside of conscious control.
16
As Ralph Turner (1976; see also Gordon, 1989)
instructs us, in the modem era persons draw inferences about what they really
are from the perceived spontaneity of and their perceived inability to control their
responses, as do others who observe responses they characterize in these ways.
The experience of a presumably visceral response to an event accords that event
special status with respect to the meaning of the event for self. It communicates
strong messages about who and what we are, what roles we play carry particular
gratications or absence thereof, what identities are truly central to self, etc.
The question of how acute emotion enters identity theory can be pursued by
developing implications of this discussion.
INTEGRATING AFFECT
INTO IDENTITY THEORY
Those working from an interactionist perspective have emphasized the import
for social behavior of subjective cognitive responses, of emotion norms and
socially organized emotional responses, and more recently, of the import of sub-
jectively felt acute emotional behavior. While incorporating all of these symbolic
intractionist-based emphases, I focus in particular on characteristic features of the
experience of emotion, whether that experience is shaped by evolutionary residues,
by social structural, cultural and interactional positioning, or by some interplay
of these does not matter, that may critically enter the social process of shaping
commitments, organizing identities, and impacting role performances. I argue the
potential in extending the traditional arguments of symbolic interactionism about
the import of cognitive responses to the emotions, and it is in the experience of
emotions that meanings critical to the consequences of emotions for self are to
be found.
Identity theory invites a consideration of affect, both sentiment and acute emo-
tion, as these may alter the probability of networks of relationships being formed,
maintained, or dissolved; insofar as affect impacts directly on commitment to
networks; insofar as affect impacts directly or indirectly the salience of identities;
Integrating Emotion into Identity Theory 11
and insofar as affect alters the ways in which commitment and identity salience
relate to role choices. It also invites a consideration of the reciprocal effects
of commitment, identity salience, and role performance on the generation and
experience of affect.
Part of these effects must be the resultant of socialization practices that build
cultural as well as more personal sentiments into self. Much affect in social life
is routinized in ways that may never enter the consciousness of persons whose
behavior is nonetheless affected. If it does enter consciousness, such affect is
likely to be viewed as in the nature of things. If it does not enter consciousness
or is experienced as in the nature of things, it is unlikely to be open to personal
or social control.
17
Persons identities incorporate routinized, unexamined,
unconscious, in-the-nature-of-things affect that is nevertheless consequential for
the social process. For example, it is reasonable to suppose that there is selectivity
in social relationships not only in terms of the social structures of class, education,
etc., but also in bringing together persons with similar systems of affect. The vast
literature demonstrating value and attitude homogamy among members of social
groups, and demonstrating that selection mechanisms operate to bring about part
of that homogamy, supports this contention.
Persons with common affective meanings will be more likely to enter social relationships with
one another, and having entered them will be more likely to maintain those relationships.
Thus, affect in part enters identity theory in the same way social structures do,
by increasing or decreasing the probability of social networks forming and being
maintained.
18
Additionally, it seems reasonable to assert that
sentiments are likely to directly impact commitment to groups to the degree that sentiments are
shared within the groups and whatever the positive or negative quality of the shared sentiments.
This expectation derives from the premise that commitment depends on the costs
of foregoing interactions with valued others, and the value of others, other things
equal, will be higher when affective meanings are shared. This is most likely to be
the case when it is the affective meanings of the positional designations dening
self-other relationships that are involved. More, it can be expected that intense
affect will reinforce these effects.
Whether the affect generated in particular role relationships and attaching
either to the persons or the roles increases or decreases commitment to social
networks will depend on whether the affect generated is positive or negative.
In general, positive affect is likely to lead to seeking more interaction in those role relationships,
which in turn will increase commitment to the relationship.
12 SHELDON STRYKER
This restates a sociological principle, best known through Homanss (1950)
analyses of the relation between liking and interaction. Added, perhaps, is that
greater interaction in particular networks of role relationships increases the costs of
breaking off those relationships. Too, if there is anything to the idea of an economy
of time and energy, increasing interaction in particular networks must at some
point decrease interaction in other networks, decreasing commitment to the latter.
In general, we can expect negative affect generated in role relationships to have the opposite
effects,
although negative affect can reinforce commitment as well.
19
We can also expect that identities to which intense positive or negative affect is attached will
be more salient by virtue of that intense affect: messages to self contained in intense affect will
be more insistent, difcult to ignore, and, probably, easier to read.
However, the effect of identity salience on role behavior can be expected to differ
markedly depending on the quality of the affect involved.
Intense positive affect will lead to role behavior conrming the identities to which such affect
is linked, intense negative affect will lead to role behavior denying the identities.
The latter expectation may hold only for negative sentiments and not for acute
emotional outbursts. With regard to the latter, Heises (1979) theorization and re-
search evidence as well as Burkes (1996, 2004) indicates that persons try very hard
to reafrm existing identities when these are threatened by disconrming events,
and both expect identity change when reconciliation is too difcult. Resolving
this apparent contradiction may require distinguishing varying levels of affective
intensity, the lower the level the more likely the reafrmation response. Alterna-
tively, the expectation linking negative affect to higher salience but also to role
behavior denying the identities may only occur in the context of other, affectively
positive identities that are also highly salient and are threatened by conrming
responses to the negative identities.
20
A further complication: if identities are
consistently and pathologically negative, conrmation of these negative identities
may well be sought (Pinel & Swann, 2000; Swann, 1983; Swann & Read, 1981).
The reciprocal of the affective commonality-commitment relation will likely
hold.
Greater commitment will lead to greater affective commonality. If positive affect generated by
role relationships increases commitment to those relationships, so too is higher commitment
likely to generate positive affect,
the argument being that the higher costs of foregoing particular role relationships
will generate positive feelings toward those with whom one is implicated in those
role relationships, that interdependence breeds positive evaluation.
21
And
Integrating Emotion into Identity Theory 13
If intense affect increases the salience of an identity, so will high identity salience heighten the
affectively positive response to role partners in relations in which that highly salient identity
nds expression.
A further set of propositions implicating commitment, identity salience, role
performance and affect can be asserted:
The higher persons commitment to a network of role relationships, the more will role
performances be in line with network expectations.
Performances meeting network expectations can be expected to earn persons positive affect
(respect, liking) from network members, and therefore self esteem for themselves.
Performances failing to meet expectations will produce negative affect in both others (anger,
disappointment) and self (embarrassment, guilt), lowering the performers self-esteem.
Generally, the greater the discrepancy involved in failing to meet expectations, the greater the
affective response, at least when commitments are high.
Failures of others to meet expectations that in turn prevent persons from meeting their own
performance expectations will intensify affective responses to those others.
In keeping with the earlier discussion of the signal functions of emotion, intense affect growing
out of role performances are likely to convert into self-attributions.
Much of the preceding treatment of affect-identity theory linkages has had as
its implicit focus the longer-term, more stable affect termed sentiments. I turn
now to a focus on the more immediate, more volatile acute emotional responses
experienced in ongoing social interaction. That consideration can begin with
the earlier observation that role relationships are context and occasion for
experiencing emotion. Identities on the basis of which persons enter and maintain
role relationships are motivational: they impel behavior that either expresses
or denies them, the former if they carry positive affective meaning for the person,
the latter if their affective meaning is negative (see Stryker & Macke, 1985 for
development of this argument).
However, ones behavior alone can never afrm or deny an identity; others
with whom one interacts in role relationships must respond in ways that accept
the afrmation or denial. This suggests that much emotion is generated in the
interplay of identity claims and others responses to the claims.
Emotional outbursts will occur when role-partners behave in ways that contradict the identity
claims made, whether these take the forms of behaviors that followfroman identity or behaviors
evidencing the absence of the identity.
In keeping with the symbolic interactionist frame, the social process portrayed
can be internalized and conducted within the person. That this internalized
14 SHELDON STRYKER
process results in intense emotional response is the basis for Higgins(1987,
1989) theorizing and is evidenced in his experimental results.
Suggested as well is that:
Structural or interactional barriers to the enactment of highly positive identities or to behaviors
denying highly salient negative identities will also generate intense emotional responses.
The denial of opportunities to play out roles expressive of those identities is, in
effect, a denial of them, calling their validity into question. Insofar as self is orga-
nized around those identities, we can expect highly emotional reactions. Similarly,
When the validity of performances expressive of salient positive identities is denied by responses
of others, immediate highly charged emotional responses can be expected.
With regard to negative identities, equivalent processes can be expected when opportunities to
behave in ways denying those identities are closed off or when the validity of behaviors denying
the negative identities is rejected.
22
Identity afrmation/denial processes occur in role relationships. In identity theory
terms, this reintroduces the concept of commitment. Persons require participation
in role relationships if they are to enact identities, for such performances
require others as co-participants and as veriers of the identity implications of
performances. If identities have a motivational thrust and require acting in the
context of role relationships, there are strong grounds for expecting effects of
commitments on the occurrence of emotional responses, and the reciprocal effects
of emotional responses on commitment.
The greater the commitment to networks in which these processes occur, the more intense the
emotional responses to these processes
since there is more at stake when commitment is high and since high commitment
makes it more likely that the signals will be transmitted both to and from self.
The obverse can also be expected:
The more intense the emotional responses to identity afrmation or denial, the greater will be
its impact on commitment,
both because others reading intense responses will increase or decrease (depending
on the valence of the emotional response) their interaction with the person and
because the person reading his/her own emotional responses will be led to seek
or avoid further interaction with those others. We can also surmise that
The more an emotional response engendered in a role relationship is experienced as out of the
ordinary, spontaneous, uncontrollable, the greater its effect on commitment.
Underlying this expectation is the presumption that the greater the surprise
contained in signals to self read from emotional responses, the greater the
credence given those signals.
Integrating Emotion into Identity Theory 15
Since identity theory posits a causal relationship of commitment on identity
salience,
23
Whatever effects the experience and expression of emotion have on commitment can also be
expected to be seen on the relative salience of identities entering self.
We can expect direct effects of emotion on identity salience and direct effects of identity salience
on the experience and expression of emotion
and these effects may be different in direction from those operating indirectly
through commitment. For example, a strong, negative emotional experience
generated in role relationships can be expected to curtail interaction and so
decrease commitment; and the decreased commitment can be expected to
decrease the salience of the identities invoked in the relationship. But a strong,
negative emotional experience can be expected to alter salience directly, since
that experience will constitute a message to self, with the nature of the change a
consequence of the content of the message. Under some circumstances, what is
signaled is likely to be that a highly salient identity impelled behavior that led to
an unhappy rebuff; that lesson should diminish the salience of the identity. Under
other circumstances, the signal could be a function of the surprise quotient of
the emotional experience, and the message that the identity had more meaning
and import than previously realized could increase the salience of the identity.
Should this occur, an expected consequence is that the range of situations taken to
be identity relevant will expand, and more opportunities for the enactment of the
identity will be available. Thoits (1994) research as well as Simons (1992) on
the impact of identity-relevant events on psychological reactions is also relevant
to the expectation that the salience of identities will impact emotional reactions.
More specically, these researchers suggest that, relative to their effects on less
salient identities, events threatening highly salient identities will engender stronger
negative emotions and events reinforcing such identities will engender stronger
positive emotions.
24
As earlier noted, Affect Control Theory (Heise, 1979) as well as Burkes
control theory of self processes (Burke, 2004) suggest that emotional shocks to
the self system generate actions to restore pre-existing affective balance, but that
More intense emotional experience as well as persistent discrepant emotional experiences
are less likely to be accommodated by homeostatic processes, thus are more likely to
change the self by altering the relative salience of identities contained in self (see Kiecolt &
Mabry, 2000).
As earlier argued, such responses can also be expected to alter commitments.
Here, we can add that they are likely to have greater direct impact on salience
than on commitment, since fewer constraints likely operate on the former (Serpe,
1987; Stryker & Serpe, 1987).
16 SHELDON STRYKER
Since spontaneity, uncontrollability, and out-of-the-ordinariness of emotional
outbursts can be expected to intensify the impact of such outbursts on commitment,
and given the theoretically expected and empirically demonstrated impact of
commitment on identity salience,
Emotional outbursts can be expected to intensify the impact of either positive or negative
emotional experience on identity salience. In particular, non-normative emotional outbursts
will do so
since these are prone to be experienced as spontaneous and out-of-the ordinary,
as well as signals informing persons of their true selves. And, again, the
experiential characteristics of emotions will carry a message to self, will testify to
the import of the generating role-relationships, and so will intensify the meaning
read from the message with respect to self.
Finally,
The higher the salience of the identity invoked in a situation, the more emotional potential
will be involved, i.e. the more likely are messages drawn in that situation to have an experienced
affective quality, and any emotional reaction linked to that identity will be intensied.
Should the expression of a highly salient identity be blocked structurally or
interactionally, a highly charged negative emotional response can be expected.
25
With respect to successful expression of highly salient identities, the situation
maybe more complicated. Certainly, successful expression of highly salient
identities can be expected to be pleasurable, but it is perhaps from the surprise
element in interactional experience that the intensity of positive emotion derives.
CONCLUSION: THE UNFINISHED AGENDA
I began this essay with an agenda, seeking to integrate affect into identity theory by
asking if and how affect or emotion impacts the basic identity theoretic variables
of commitment, identity salience, and role performance. While I believe that the
essay indicates reasonably well why affect can be expected to matter and howit can
be expected to matter, the utility of the theoretical effort expended will necessarily
be decided by whatever research may be stimulated by the theorizing (indeed, if
any such research is, in fact, undertaken). In short, then, the most pressing issue
is to translate the abstractions of theory into the specics of research based on the
theoretical claims in this chapter.
That fundamental gap in sociological and social psychological work aside, I am
well aware that issues in theorizing relations of affect and identity theory variables
have not all been resolved. It likely will not have gone unnoticed that I have had
little to say about the meanings of specic emotional responses (love, hate, guilt,
Integrating Emotion into Identity Theory 17
embarrassment, joy, anger, etc.) and how these may relate to concepts of identity
theory. The reason for this is simple: I can, at least at this time, offer no particular
insights into how these meanings either inform or are informed by the theory. It is
worth noting, however, that existing accounts of how the ways in which specic
emotional responses are generated, whether from the point of view of cognitive
psychology (e.g. Higgins, 1987, 1989) or from the point of view of sociological
social psychology (Goffman, 1959, 1967; Hochschild, 1979, 1983; Riezler, 1943;
cheff, 1988, 2003; Shott, 1979) are compatible with the theorization presented.
So, too, with respect to issues that have been more central to the present effort:
the putative distinction between sentiment and emotion; the dimensionality of
affective responses, in particular acute emotional outbursts, the relationships of
affective responses to commitment, identity salience, and role performance both
as cause and as consequence. I believe a promising beginning has been made;
I am under no illusion that more has been accomplished.
NOTES
1. I use the terms affect and emotion interchangeably in this chapter.
2. See the journal Self and Identity, initially appearing in 2002, for representative work
from psychological perspectives.
3. See the journal Symbolic Interaction since its inception for work reecting this
perspective. Much, certainly not all, of the work on the topic of self and emotions appearing
in Social Psychology Quarterly also reects this perspective; see in particular the special
issue of this journal on sentiments, affect, and emotion, Vol. 52, 1989.
4. As Thoits has remarked (2003), an avalanche of studies of emotional socialization
on the job followed Hochschilds ground-breaking work. See Thoits (2003) for further
leads into this literature, as well as into the literature on emotion management.
5. Identity and identity theory mean many things in the social science literature (Stryker,
2000a). The meaning of these terms in this chapter is limited specically to the set of
ideas developed especially in my own work as well as that Peter J. Burke (2004; Stryker
& Burke, 2000).
6. This volume has been republished in 2002 under the same title by Blackburn Press,
Caldwell, NJ. It contains a new Foreword entitled Symbolic Interactionism Then and
Now (Stryker, 2002).
7. Identity theory has sometimes been read as requiring a single well-dened ordering
of the salience of identities (see, e.g. Amato & Marsiglio, 2002). This seems to be true
particularly for persons pursuing implications of the idea that the sheer number of roles (and
identities) persons hold implies the availability of important behavioral resources affecting
their well-being (e.g. Thoits, 2003). Some comments seemappropriate. First, identity theory
was developed as a theory of role-choice, seeking to explain such choice by variations in
identity salience; identity accumulation theory has different objectives, in Thoits case to
account for the effects of identities on persons psychological health or well-being, in Amato
and Marsiglias case to deal with consequences of identities that are mutually reinforcing in
18 SHELDON STRYKER
their implications for behavior and in that sense seemto imply the existence of a higher order
identity. Second, identity theory does not require that identities vary in salience; some may
well be equivalently salient, whether or not they have reinforcing behavioral implications.
I have discussed relevant matters in Stryker (1989) and in Stryker (2000a).
8. David Heise notes that Ekmans work showing how rened an instrument the human
face is for emotional expression suggests that evolution provided for the elaboration and
not the atrophy of emotions in humans (personal communication).
9. To avoid excluding the work of persons working from a symbolic interactionism
frame from the previous discussion of the general literature on emotions, theoretical and
empirical work on emotions that belongs in the current discussion was introduced there.
That noted, I seek to minimize redundancy at this point.
10. While once concern with the sentiments dominated sociologists concern with
emotion, Thoits (personal communication) suggests that more recently sentiments receive
less attention than they deserve.
11. Given the existence of prior selves and/or of multiple identities within self, internal
dynamics of the self allows an internal conversation that can take the place of others in
(re)shaping self. However, it is dubious that such external conversation can sustain a new
self without external support.
12. This is certainly not to denigrate the role of discourse in making society possible,
only to assert the correlative role of emotional experience in doing so.
13. These ideas are not novel. They are anticipated by Hochschild (1983); more by
Heise (1979, 1999); and perhaps most of all by Thoits (1985, 2003) and Gordon (1989).
14. That the objective character of the world (the situation) often impacts behavior
either directly or indirectly though affecting the subjective responses of denitions and
interpretation of situations cannot be denied. This observation is the basis for what has
been called a structural symbolic interactionism (Stryker, 1980; Stryker & Vryan, 2003).
15. As Peggy Thoits has reminded me (personal communication).
16. The characterizations of emotional experience offered are not intended in all
or-nothing terms. That is, emotions are experienced in varying degrees as being out of
the ordinary, spontaneous, uncontrollable, etc. That emotions may be socially constructed
does not deny they may be experienced in the ways indicated; indeed, these ways may
well be aspects of the constructions.
17. Not all routinized affect or sentiment has this taken for granted quality. Some highly
charged ritual, religious or otherwise, is apparently not experienced in this way, but as out
of the ordinary, spontaneous, and as not under control (Laubach, 2002). When this is the
case, we can perhaps expect that such affect will have consequences closely akin to those
of acute affect. Interestingly, as Turner (1976) asserts and as Hochschilds (1983) work
illustrates, responses that are non-normative (and that appear to the person as spontaneous)
can serve to signal to persons their true selves.
18. The process entailed here may well be involved in the evolution of social movements
from spontaneous crowds.
19. While less likely, social relationships creating negative affect can lead to higher
identity salience and commitment. Peggy Thoits (personal communication) suggests that
conict ridden family relationships illustrate this observation. David Heise (personal
communication) reminded me of the literature showing that leaders sometimes generate
negative affect in groups through scapegoating or witchhunting in order to increase
group cohesion.
Integrating Emotion into Identity Theory 19
20. This hypothesis was being examined in the context of a genital herpes self-help
group by my student, Elizabeth Craft before her untimely death from a congenital heart
disease. The underlying ideas were in good part Elizabeths. James D. Lee completed part
of Elizabeths research; see Lee and Craft (2002). Peggy Thoits (personal communication)
suggests that the idea contained in this hypothesis may be useful for labeling theory more
generally.
21. This may simply be a reformulation of Heiders (1958) balance principle.
22. Thoits (personal communication) observed that the idea that people react emotion-
ally to the absence of opportunity to deny a negative identity may be useful for labeling
theory more generally: people get upset because they are not allowed to deny a negative
identity with which they are labeled, and that upset probably reinforces their negative
identity in the eyes of others.
23. The theory also recognizes the reciprocal causal effects of identity salience on
commitment (Serpe, 1987; Stryker & Serpe, 1987) but argues on theoretical as well as
empirical grounds that the effects of commitment on salience will be greater than vice versa.
24. These expectations restate suggestions offered by Peggy Thoits (personal
communication). Her suggestions are also relevant to subsequent expectations presented.
25. Thoits (1994, 2003) notes that a conception of humans as psychological activists
permits the observation that they may actively intervene to avoid, reduce or otherwise
mitigate the negative affective effects of events threatening salient identities. Clearly, the
possibility of such intervention calls for serious attention in further efforts to relate affect
to identity theory.
ACKNOWLEDGMENTS
This chapter has evolved from work initially done in 19861987 while a Fellow
of the Center for Advanced Studies in the Behavioral Sciences and presented in
the paper, The Interplay of Affect and Identity, at the ASA meetings in 1987.
Support of the John D. and Catherine T. MacArthur Foundation for that Fellowship
is gratefully acknowledged. For reasons that need not be recounted, that paper was
never published, but over the years people have asked about it and cited it. These
indicators were sufcient to motivate a revision, up-dating and extension. Peggy
Thoits reread the original paper at my request, and helped enormously to allow me
to see what in it was still useful and what required further development. I cannot
thank her enough for this act of personal and professional courtesy.
REFERENCES
Amato, M., & Marsiglio, W. (2002). Self-structure, identity and commitment: Promise keepers Godly
Man project. Symbolic Interaction, 25, 4165.
Arnold, M. B. (1960). Emotion and personality. Psychological Aspects (Vol. 1). New York: Columbia
University Press.
20 SHELDON STRYKER
Averill, J. R. (1980). A constructionist view of emotion. In: J. R. Plutchik & H. Kellerman
(Eds), Emotion: Theory, Research and Experience (Vol. 1, pp. 305339). New York:
Academic.
Berscheid, E. (1983). Emotion. In: H. H. Kelley (Ed.), Close Relationships. New York: W. H.
Freeman.
Blumer, H. (1951). Collective behavior. In: A. M. Lee (Ed.), The Principles of Sociology (pp. 167222).
New York: Barnes & Noble.
Blumer, H. (1969). Symbolic interactionism: Perspective and method. Englewood Cliffs, NJ:
Prentice-Hall.
Burke, P. J. (1996). Social identities and psychosocial stress. In: H. B. Kaplan (Ed.), Psychosocial
Stress: Perspectives on Structure, Theory, Life Course, and Methods (pp. 141174). San Diego,
CA: Academic Press.
Burke, P. J. (1997). An identity model for network exchange. American Sociological Review, 62,
134150.
Burke, P. J. (1999). Identity processes and social stress. American Sociological Review, 56,
836849.
Burke, P. J. (2004). Identities and social structure: The Cooley-Mead Award address. Social Psychology
Quarterly, 67, 515.
Burke, P. J., & Reitzes, D. C. (1981). The link between identity and role performance. Social
Psychology Quarterly, 44, 8392.
Burke, P. J., & Stets, J. E. (1999). Trust and commitment through self-verication. Social Psychology
Quarterly, 62, 347366.
Burke, P. J., & Tully, J. C. (1977). The measurement of role/identity. Social Forces, 55, 8197.
Cahill, S. E. (1999). Emotional capital and professional socialization: The case of mortuary science
students (and me). Social Psychology Quarterly, 62, 101116.
Cast, A. D., & Burke, P. J. (2002). A theory of self-esteem. Social Forces, 80, 10411068.
Clark, C. (1990). Emotion and micropolitics in everyday life: Some patterns and paradoxes of place.
In: T. D. Kemper (Ed.), Research Agendas in the Sociology of Emotions (pp. 305333). Albany,
NY: State University of New York Press.
Clark, C. (1997). Misery and company: Sympathy in everyday life. Chicago: University of Chicago
Press.
Cooley, C. H. (1902). Human nature and social order. New York: Scribner.
Darwin, C. (1872). The expression of emotion in man and animals. London: Murray.
Denzin, N. K. (1984). On understanding emotion. San Francisco: Jossey-Bass.
Denzin, N. K. (1990). On understanding emotion: The interpretive-cultural agenda. In: T. D. Kemper
(Ed.), Research Agendas in the Sociology of Emotions (pp. 85116). Albany, NY: State
University of New York Press.
Ekman, P. (Ed.) (1982). Emotion in the Human Face (2nd ed.). Cambridge: Cambridge University Press.
Francis, L. E. (1997). Ideology and interpersonal emotion management: Redening identity in two
support groups. Social Psychology Quarterly, 60, 153171.
Francis, L. E., Monahan, K., & Berger, C. (1999). A laughing matter: The use of humor in medical
interactions. Motivation and Emotion, 23, 155174.
Freud, S. (1948). Inhibitions, symptoms, and anxiety. London: Hogart Press; Institute of Psycho-
analysis.
Gerth, H. H., &Mills, C. W. (1953). Character and social structure. London: Routledge &Kegan Paul.
Goffman, E. (1959). The presentation of self in everyday life. Garden City, NY: Doubleday.
Goffman, E. (1967). Interaction ritual. Chicago: Aldine.
Integrating Emotion into Identity Theory 21
Gordon, S. L. (1981). The sociology of sentiments and emotions. In: M. Rosenberg & R. H.
Turner (Eds), Social Psychology: Sociological Perspectives (pp. 562592). New York:
Basic Books.
Gordon, S. L. (1989). The socialization of childrens emotions: Emotional culture, exposure,
and competence. In: C. Saarni & P. Harris (Eds), Childrens Understanding of Emotions
(pp. 319349). New York, NY: Cambridge University Press.
Gross, E., & Stone, G. P. (1964). Embarrassment and the analysis of role requirements. American
Journal of Sociology, 70, 115.
Hafferty, F. W. (1988). Cadaver stories and the emotional socialization of medical students. Journal
of Health and Social Behavior, 29, 344356.
Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.
Heise, D. R. (1979). Understanding events: Affect and the construction of social action. New York,
NY: Cambridge University Press.
Heise, D. R. (1999). Controlling affective experience interpersonally. Social Psychology Quarterly,
62, 416.
Higgins, E. T. (1987). Self-discrepancy: A theory relating self and affect. Psychological Review, 94,
319340.
Higgins, E. T. (1989). Self-discrepancy theory: What patters of self-beliefs cause people to suffer?
In: L. Berkowitz (Ed.), Advances in Experimental Social Psychology (Vol. 22, pp. 93136).
Orlando, FL: Academic Press.
Hochschild, A. R. (1979). Emotion work, feeling rules, and social structure. American Journal of
Sociology, 85, 551575.
Hochschild, A. R. (1983). The managed heart. Berkeley, CA: University of California Press.
Homans, G. (1950). The human group. New York: Harcourt Brace Jovanovich.
Hume, D. (1748). Philosophical essays concerning human understanding. London: A. Millar.
Illouz, E. (1997). Consuming the romantic utopia: Love and the cultural contradictions of capitalism.
Berkeley, CA: University of California Press.
Izard, C. E. (1969). The emotions and emotion constructs in personality and culture research. In:
R. B. Cattell (Ed.), Handbook of Modern Personality Theory. Chicago: Aldine.
Kemper, T. D. (1987). How many emotions are there? Wedding the social and the autonomic
components. American Journal of Sociology, 87, 336362.
Kiecolt, K. J. (1994). Stress and the decision to change oneself: Atheoretical model. Social Psychology
Quarterly, 57, 4763.
Kiecolt, K. J., & Mabry, J. B. (2000). Agency in young adulthood: Intentional self-change among
college students. In: T. J. Owens (Ed.), Advances in Life-Course Research, Self, Identity and
Social Movements (Vol. 5, pp. 181205). Greenwich, CT: JAI Press.
Klinnert, M. D., Campos, J. J., Sorce, J. F., Emde, R. N., & Svejda, M. (1983). Social referencing in
infancy. In: R. Plutchik & H. Kellerman (Eds), Emotion: Theory, Research, and Experience
(Vol. 2, pp. 5786). New York: Academic Press.
Laubach, M. (2002). Self possessed: The sociology of altered states of consciousness. Unpublished
Doctoral Dissertation. Bloomington, IN: Indiana University.
Lazarus, R. S. (1992). Progress on a cognitive-motivational-relational theory of emotion. American
Psychologist, 48, 819834.
Lee, J. D., & Craft, E. A. (2002). Protecting ones self from a stigmatized disease . . . Once one has it.
Deviant Behavior, 23, 267299.
Lindesmith, A. R., & Strauss, A. L. (1949). Social psychology. New York: Dryden.
Lively, K. J. (2000). Reciprocal emotional management. Work and Occupations, 27, 3263.
22 SHELDON STRYKER
Lively, K. J. (2002). Client contact and emotional labor. Work and occupations, 29, 198
225.
Markus, H., & Zajonc, R. B. (1985). The cognitive perspective in social psychology. In: G. Lindzey &
E. Aronson (Eds), Handbook of Social Psychology (3rd ed., Vol. I, pp. 137230). New York:
Random House.
McGuire, W. J. (1973). Perceptual selectivity and the self-concept. Division 8 Presidential Address
presented at the meeting of the American Psychological Association, Montreal.
McGuire, W. J., & Padawer-Singer, A. (1976). Trait salience in the spontaneous self-concept. Journal
of Personality and Social Psychology, 33, 743754.
Mead, G. H. (1934). Mind, self and society. Chicago, IL: University of Chicago Press.
Pinel, E. C., & Swann, W. B., Jr., (2000). Finding the self through others. In: S. Stryker, T. J. Owens
& R. W. White (Eds), Self, Identity and Social Movements (pp. 132152). Minneapolis, MN:
University of Minnesota.
Plutchik, R. (1980). Emotion: A psychoevolutionary synthesis. New York: Harper & Row.
Pollak, L. H., & Thoits, P. A. (1989). Processes of emotional socialization. Social Psychology
Quarterly, 52, 2234.
Riezler, K. (1943). Comment on the social psychology of shame. American Journal of Sociology, 48,
457465.
Rosaldo, M. Z. (1980). Knowledge and passion. New York: Cambridge.
Rosaldo, M. Z. (1984). Toward an anthropology of self and feeling. In: R. Schweder & R. A. Levine
(Eds), Culture Theory: Essays on Mind, Self, and Emotion (pp. 137157). Cambridge, MA:
Cambridge University Press.
Scheff, T. (1988). Shame and conformity: The deference-emotion system. American Sociological
Review, 53, 395406.
Scheff, T. (2000). Shame and the social bond. Sociological Theory, 18, 8498.
Scheff, T. (2003). Shame in self and society. Symbolic Interaction, 26, 239262.
Schwartz, M., & Stryker, S. (1970). Deviance, selves and others. Washington, DC: American
Sociological Association.
Serpe, R. T. (1987). Stability and change in self: A structural symbolic interactionist explanation.
Social Psychology Quarterly, 50, 4455.
Serpe, R. T., & Stryker, S. (1987). The Construction of self and the reconstruction of social
relationships. In: E. J. Lawler & B. Markovsky (Eds), Advances in Group Processes (Vol. 4,
pp. 4166). Greenwich, CT: JAI Press.
Shibutani, T. (1961). Society and personality: An interactionist approach to social psychology.
Englewood Cliffs, NJ: Prentice-Hall.
Shott, S. (1979). Emotion and social life: A symbolic interactionist analysis. American Journal of
Sociology, 84, 13171334.
Simmons, R. (2001). Comfort with the self. In: T. J. Owens, S. Stryker &N. Goodman (Eds), Extending
Self-Esteem Theory and Research: Sociological and Psychological Currents (pp. 198222).
New York: Cambridge.
Smith, A. (1759). The theory of moral sentiments. Indianapolis: Liberty Press.
Smith-Lovin, L., & Heise, D. R. (1988). Analyzing social interaction: Advances in affect control
theory. New York: Gordon & Breach.
Stryker, S. (1968). Identity salience and role performance: The relevance of symbolic interaction
theory for family research. Journal of Marriage and the Family, 30, 558564.
Stryker, S. (1980). Symbolic interactionism: A social structural version. Menlo Park, CA: Ben-
jamin/Cummings.
Integrating Emotion into Identity Theory 23
Stryker, S. (1989). Further developments in identity theory: Singularity vs. multiplicity of self. In:
J. Berger, M. Zelditch Jr. & Bo Anderson (Eds), Sociological Theories in Progress: New
Formulations (pp. 3557). Newbury Park, CA: Sage.
Stryker, S. (1994). Freedom and constraint in social and personal life: Toward resolving the paradox
of self. In: G. M. Platt & C. Gordon (Eds), Self, Collective Behavior and Society: Essays
Honoring the Contributions of Ralph H. Turner (pp. 119138). Greenwich, CT: JAI Press.
Stryker, S. (2000a). Identity competition: Key to differential social movement participation? In:
S. Stryker, T. J. Owens & R. W. White (Eds), Self, Identity, and Social Movements (pp. 2140).
Minneapolis: University of Minnesota Press.
Stryker, S. (2000b). Identity theory (pp. 38). In: E. F. Borgatta & R. J. V. Montgomery (Eds),
Encyclopedia of Sociology (2nd ed., Vol. 2, pp. 12531258). New York, NY: Macmillan.
Stryker, S. (2001). Traditional symbolic interactionism, role theory, and structural symbolic inter-
actionism: The road to identity theory. In: J. Turner (Ed.), Handbook of Sociological Theory
(pp. 211232). New York, NY: Kluwer/Plenum.
Stryker, S. (2002). Symbolic interactionism: Asocial structural version. Caldwell, NY: BlackburnPress.
Stryker, S., & Burke, P. J. (2000). The past, present and future of an identity theory. Social Psychology
Quarterly, 63, 284297.
Stryker, S., & Serpe, R. T. (1982). Commitment, identity salience, and role behavior: Theory and
research example. In: W. Ickes & E. S. Knowles (Eds), Personality, Roles, and Social Behavior
(pp. 199218). New York: Springer-Verlag.
Stryker, S., & Serpe, R. T. (1994). Identity salience and psychological centrality: Equivalent,
overlapping or complementary concepts? Social Psychology Quarterly, 57, 1635.
Stryker, S., & Statham, A. (1985). Symbolic interaction and role theory. In: G. Lindzey & E. Aronson
(Eds), Handbook of Social Psychology (3rd ed., Vol. I, pp. 311378). NewYork: RandomHouse.
Stryker, S., & Vryan, K. D. (2003). The symbolic interactionist frame. In: J. D. DeLamater (Ed.),
Handbook of Social Psychology (pp. 328). New York: Kluwer/Plenum.
Swann, W. B. Jr. (1983). Self-verication: Bringing social reality into harmony with the self. In:
J. Suls & A. G. Greenwood (Eds), Social Psychological Perspectives on the Self (pp. 233266).
Hillsdale NY: Lawrence Erlbaum.
Swann, W. B., Jr., & Read, S. J. (1981). Self-verication processes: How we sustain our self-
conceptions. Journal of Experimental and Social Psychology, 17, 351372.
Taylor, V. (2000). Emotions and identity in womens self-help movements. In: S. Stryker, T. Owens
& R. W. White (Eds), Self, Identity and Social Movements (pp. 271299). Minneapolis, MN:
University of Minnesota Press.
Thoits, P. A. (1985). Self-labeling processes in mental illness: The role of emotional deviance.
American Journal of Sociology, 92, 221249.
Thoits, P. A. (1996). Managing the emotions of others. Symbolic Interaction, 19, 85110.
Thoits, P. A. (2003). Emotion norms, emotion work, and social order. In: A. S. R. Manstead, N. H.
Frijda & A. H. Fischer (Eds), Feelings and Emotions: The Amsterdam Symposium. New York,
NY: Cambridge University Press.
Tomkins, S. S. (1963). Affect, imagery, consciousness: The negative affects. New York: Tavis-
tock/Routledge.
Turner, R. H. (1976). The real self: From institution to impulse. American Journal of Sociology, 81,
9861016.
Weinrich, J. D. (1980). Toward a sociobiological theory of emotions. In: R. Plutchik & H. Kellerman
(Eds), Emotion: Theory, Research, and Experience (Vol. 1, pp. 11338). New York: Academic
Press.
IDENTITIES, EVENTS, AND MOODS
Peter J. Burke
INTRODUCTION
Identity verication is the ongoing process of controlling perceptions of self-
relevant meanings in a situation so that they correspond to the meanings held in the
identity standard that denes who one is in the situation. Identity control theory
posits that when a disturbance to this process occurs leading to a lack of such
correspondence, a persons identities are not veried. As a result, they engage in
behavior that serves to counteract the disturbance and change meanings and re-
sources in the situation so that ones reected appraisals or perceived self-relevant
meanings once again match the meanings held in ones identity standard (Burke,
1991, 1996; Stets & Burke, 1996, 2003). Accompanying this cognitive-behavioral
process, there is an affective response to the discrepancy between perceptions and
standard (Burke, 1991, 1996). Prior work has shown that when the discrepancy is
large or is increasing, negative emotions result; and, when the discrepancy is small
or decreasing, positive affect results (Burke & Stets, 1999; Cast & Burke, 2002;
Ellestad & Stets, 1998; Stets, 2003; Stets & Tsushima, 2001).
For example, Burke and Stets (1999) show that when the spousal identity is
conrmed by ones partner, there is an increase of feelings of love for the partner
(as well as trust and commitment), while the lack of conrmation leads to increased
levels of distress. Cast and Burke (2002) show that when the spousal identity of
newly married couples is not conrmed there are decreased levels of self-worth and
self-efcacy, and that when the lack of conrmation goes on for an extended period
of time, there are further decreases in these self-feelings. Stets (2003) examines
injustice as a discrepancy between an identity standard that expects a certain level
Theory and Research on Human Emotions
Advances in Group Processes, Volume 21, 2549
Copyright 2004 by Elsevier Ltd.
All rights of reproduction in any form reserved
ISSN: 0882-6145/doi:10.1016/S0882-6145(04)21002-5
25
26 PETER J. BURKE
of payoff and a perception that the level of payoff is less than is set in the standard.
She shows the feelings of anger, resentfulness, and disgust that result from this
discrepancy. The lack of conrmation does not have to result from an under-
evaluation of the self by others. Burke and Harrod (2002) show that the same
sorts of distress occur when there is an over-evaluation of the self; people feel best
when others evaluate them in the same way that they evaluate themselves. It is
clear, therefore, that the lack of identity verication has affective consequences.
However, these studies generally are over a longer time-span where the lack of
verication is ongoing. In contrast to longer-term, ongoing verication problems,
identity-disconrming events that happen periodically but generally do not persist
have not been studied.
1
Outside the area of identity research, however, much of the discussion of sources
of distress does focus on the impact of daily life-events (e.g. Brown, 1974; Cochran
& Hammen, 1985; Dohrenwend, 1973; Thoits, 1978). Events are things that
happen to people, such as being red from a job or getting married. They are
also things that people confront at a point in time such as being told that ones
grandmother has been put in intensive care, or that there is a new medication that
will help a particular condition. However, the stressful consequences of such life-
events have not been uniformly conrmed; sometime negative events increased
distress, sometimes positive events increased distress (Burke, 1996). Which
events produce distress, and which events do not produce distress often has been
problematic.
The link between life-events and distress was claried by the works of Thoits
(1991) and Burke (1991) both of whom noted that much of the inconsistency in
research examining the stress induced by life-events can removed by focusing
on what have been called identity-relevant life-events. Thoits argument was
that events that threaten or disrupt salient or important identities should produce
distress. Events that do not do this, or events that enhance important identities
should not produce distress (Brown &McGill, 1989; Burke, 1991, 1996; Hammen,
Marks, DeMayo & Mayol, 1985; Hammen, Marks, Mayol & DeMayo, 1985).
The key to understanding one major source of distress thus lies in noting how
both events and ongoing interactions can disrupt the process of identity verication.
With ongoing interactions as the source of the identity disruption, however, the
effects generally appear to be persistent and relatively long-term, as interactants
take time to work out mutually verifying relationships in a changing environment
(Burke & Stets, 1999). These long-term effects produce not only increases in
distress but decreases in feelings of self-worth and self-efcacy (Cast & Burke,
2002). Life-events, on the other hand, tend to be single occurrences rather than
on-going processes, and consequently the distressful outcomes are expected to be
relatively short-lived.
Identities, Events, and Moods 27
In the present paper, I am concerned with the more immediate disturbance of
moods resulting fromthe disruption of identity processes by events. By examining
these abrupt affective consequences, we gain a better understanding of the role of
identity processes in the emotional life of individuals. Both shorter- and longer-
term consequences are expected within the framework of identity control theory,
but the consequences of the shorter term, event driven processes may themselves be
a function of the longer-term problems in identity verication. That is, events may
have different consequences if they occur to someone who already has ongoing
identity verication problems as opposed to occurring to someone who does not
have such problems; or if they occur to someone with lower vs. higher feelings of
self-worth.
MOODS
Moods are affective responses that are seen to differ fromemotions in three primary
ways, though not all researchers agree on all points (Frijda, 1993). Moods are
usually seen to be longer in duration than emotions, to have lower intensity, and to
be more diffuse and global (Ekman, 1994). Frijda (1993) suggests that the last of
these is, perhaps, the most important and the most agreed upon difference between
moods and emotions.
2
Emotions are seen to be about something, being angry at
someone or happy about something, whereas moods often have no orientation to
a target or object. Having no target, however, is not the same thing as having no
particular cause. Moods may be caused by some event, yet the feelings are diffuse
and unfocused (Frijda, 1994).
Two basic moods have received attention and operate somewhat independently.
One of these reects the positive-negative or tense-calm dimension of feeling
(which I will refer to as unease/distress), while the other reects the tired-energetic
dimension of feeling (which I will refer to as activity/arousal) (Thayer, 1996;
Watson & Tellegen, 1985).
3
The rst of these mood dimensions is often taken as
providing a cue in a self-regulatory system (Morris, 1992) with good (positive)
moods indicating a satisfactory state of affairs and bad (negative) moods indicating
some sort of discrepancy in the system. Morris further suggests that the cue is
more specic than simply pointing out a problem. He suggests that the problem
is one of potentially insufcient resources needed to meet ones goals (Morris,
1992). This makes the mood especially relevant to identity processes since the
meanings and expectations held in our role or group identity standards are goals
that are obtained when identities are veried, that is when situated self-relevant
perceptions are brought to match the identity standards. Achieving goals is what
identities do (Burke, 1991, 2003).
28 PETER J. BURKE
The second mood dimension seems more a function of the natural biological
rhythms of the day (as well as sugar intake, exercise, and drugs such as caffeine)
(Morris, 1992). While these two dimensions are somewhat independent, they are
not unrelated. Thayer (1996) suggests that small amounts of tension (the rst
dimension of mood) will increase the level of energy, but greater distress will
ultimately reduce the level of energy. At the same time, increasing the level of
energy often has a tension-reducing consequence.
IDENTITY DISRUPTING EVENTS
In the present paper, I am interested in the changes in mood that occur in re-
sponse to identity disrupting events, especially as these may be modied by levels
of self-worth and longer-term identity disruption. The model suggests that when
an identity-disrupting event occurs, people will have an increased level of un-
ease/distress that signals a problem needing attention and provides a motivation to
change the situation to reduce the problem. At the same time, it is suggested that
when an identity is already under threat from lack of self-verication, the effects
of further disruption from events may have an even bigger consequence for the
level of unease/distress. At the same time, it is also well known that feelings of
self-worth act as a buffer to the effects of distressful events (Baumeister, 1998;
Blaine & Crocker, 1993; Campbell et al., 1990; Cast & Burke, 2002). These ideas
lead to the following hypotheses:
Hypothesis 1. Identity disrupting events increase the level of unease/distress.
This basic hypothesis is modied by the hypothesized moderating effects of the
level of verication of other important identities and the level of self-worth of the
individual.
Hypothesis 1a. The degree to which other identities are already not conrmed
magnies the effect of identity disrupting events on the level of unease/distress.
Hypothesis 1b. The level of self-worth diminishes the effect of identity dis-
rupting events on the level of unease/distress.
Additionally, as suggested above, unease/distress, like depression, act to reduce
the level of activity/arousal in a person (Koehn, 2001; Stouffer Calderon,
2001; Williamson & Shaffer, 2000). This effect, along with the suggested mod-
erating effects of self-worth and identity discrepancy, is given in Hypothesis 2.
In Hypothesis 3, I also examine the direct effect of identity-disrupting events
on activity/arousal along with the moderating effects of self-worth and
identity discrepancy.
Identities, Events, and Moods 29
Hypothesis 2. The level of unease/distress directly decreases the level of activ-
ity/arousal.
As with Hypothesis 1, this hypothesis is modied by the expected moderating
effects of the level of verication of other identities and the level of self-esteem
of the individual.
Hypothesis 2a. The degree to which other identities are already not conrmed
magnies the effect of unease/distress on the level of activity/arousal.
Hypothesis 2b. The level of self-worth diminishes the effect of unease/distress
on the level of activity/arousal.
In addition to the effects of unease/distress on activity/arousal, I hypothesize
that the identity disconrming events themselves also inuence the level of
activity/arousal as they inuence the level of unease/distress.
Hypothesis 3. Identity disrupting events decrease the level of activity/arousal.
Again, this basic hypothesis is modied by the expected moderating effects of
the level of verication of other identities and the level of self-esteem of the
individual.
Hypothesis 3a. The degree to which other identities are already not conrmed
magnies the effect of identity disrupting events on the level of activity/arousal.
Hypothesis 3b. The level of self-worth diminishes the effect of identity dis-
rupting events on the level of activity/arousal.
In addition to these three main hypotheses, I explore the continued effects of
identity disrupting events on mood by examining these effects on the day after
an event, and two days after the event. To the extent that the affective outcomes
of identity disrupting events continue for the next day or two, I hypothesize that
the level of self-worth and the extent to which ones identity is already not being
conrmed will also moderate these effects.
METHODS
Sample
I examine the effects of identitydisruptiononmoodusingthree waves of data froma
longitudinal study investigating marital dynamics in the rst two years of marriage
(Tallman et al., 1998). Each data collection period included a 90-minute face-to-
face interview, a 15-minute videotaping of a conversation focused on solving an
30 PETER J. BURKE
area of disagreement, and four consecutive one-week daily diaries kept by each
respondent. The present analyses are based on information from two sources:
29,291 daily diary entries over the three time-periods provided information about
the events and moods, while the face-to-face interviews provide information about
the levels of self-worth and the degree to which the spousal identity was or was
not veried in the marriage.
The sample was drawn from marriage registration records in 1991 and 1992 in
two mid-size communities in Washington State. Of the 1,295 couples registered
to marry, 574 met the criteria for involvement (both were over the age of 18, were
marrying for the rst time, and had no children). These couples were contacted
and asked to participate; 286 completed all data collection processes in the rst
period. There was a 15% attrition rate from the rst data collection period to the
second period and an additional 4.2% attrition rate from the second to the third
period of data collection. Couples who dropped out of the study after the rst or
second round were more likely to be young (p < 0.05), less educated (p < 0.05),
and of a lower socioeconomic status (p < 0.05).
4
Measures
Not all events are disrupting to the identity process. Some events mark the conr-
mation of identities as we achieve goals for which we set our sights, or important
relationships are solidied.
5
To measure identity-disrupting events I have used the
daily diary data to select a set of events the occurrence of which corresponds to the
lack of verication of an identity. In most cases these events are directly relevant
to the self, in other cases they are events that occur to someone else but that affect
the relationship between the self and the other. These events include receiving
bad news regarding health,
6
having problems and hassles on the job,
7
having dif-
culties with friends or neighbors,
8
having problems with people in business or
government,
9
and other important disruptive events, most of which involve family
problems of one sort or another.
10
In each case, these events are disruptions or
interruptions of normal identity verication process.
Several other variables were created to indicate whether moods were observed
on the day of an event (coded 1 if yes and 0 if no), on the day after an event
occurred (coded 1 if yes and 0 if no), or two days after an event occurred (1 if yes
and 0 if no). Finally, if an event occurred to ones spouse (spouse event) rather
than to the self, that was coded as well (1 if yes and 0 if no). This was included on
the assumption that in a close relationship, and event to ones partner is likely to
inuence ones own identity because its verication is tied up in the verication
of the partner.
Identities, Events, and Moods 31
Table 1. Principle Components Factor Analysis of Unease/Distress Scale Items.
Item Loading
Calm
a
0.65
Contented
a
0.72
Comfortable
a
0.64
Uneasy 0.80
Worried 0.78
Uptight 0.84
Tense 0.80
Relaxed
a
0.69
Bothered 0.83
Distressed 0.79
Omega reliability 0.96
a
Reverse coded.
Moods were measured daily from self-ratings on a series of 20 self-feelings.
Each item was rated to indicate how you feel right now on a scale that ranged
from not at all (0) to very much (4) on a ve-point scale. Factor analysis of the
items conrmed the two dimensions that have appeared in the literature: distress
and arousal (Thayer, 1996). The measure of unease/distress is made up of 10 items
measuring background feelings of, on the one hand, being bothered and uptight
vs. on the other hand, feelings of calm and contentment. These items are presented
in Table 1, along with the loadings from a factor analysis of just these 10 items.
The omega reliability for the set of items is 0.96. The items were standardized and
summed to make the scale that was used. Certain items, as indicated in Table 1,
were reverse coded before summing. A high score on the scale indicates higher
levels of unease/distress.
The measure of activity/arousal is made up of eight items measuring background
feelings of being lively and energetic vs. feelings of being tired or drowsy. These
items are presented in Table 2, along with the loadings of a factor analysis of
just these 8 items. The omega reliability for the set of items is 0.95. The items
were standardized and summed to make the scale that was used. Certain items,
as indicated in Table 2, were reverse coded before summing. A high score on the
scale indicates higher levels of activity/arousal.
11
Spouse identity discrepancy represents the degree to which a person fails to
verify their spousal identity. It is measured following the procedures used by Swann
et al. (1994, 1992), who examined the extent to which an individuals self-view of
what it means to be a spouse was congruent with their spouses view of who they
were as a spouse. Self-verication occurs when self-views are conrmed by the
views that their spouse holds for them.
32 PETER J. BURKE
Table 2. Principle Components Factor Analysis of Activity/Arousal Scale
Items.
Item Loading
Active 0.85
Vigorous 0.79
Lively 0.85
Tired
a
0.78
Drowsy
a
0.77
Energetic 0.85
Alert 0.66
Sleepy
a
0.79
Omega reliability 0.95
a
Reverse coded.
For this measure, respondents rated ten spousal role activities in terms of the
extent to which they felt they themselves should engage in the activity (their spouse
identity standard), and the degree to which they thought their spouse should engage
in the activity. Since these data were collected for each spouse in the marriage, it is
possible to determine whether there is a correspondence between what persons feel
they should do in the spousal role and what their partners feel they should do in the
spousal role.
12
Spousal role activities that I examine include three areas that are
important components of the spousal role: instrumental, expressive, and economic.
An example item for the instrumental area is Being responsible for cleaning the
house. For the expressive area an example item is Maintaining contact with
parents and in-laws or other members of the family. Finally, an example item
for the economic area is Providing income for the family before the children are
born. Response categories for all the items ranged from not doing that activity
in the household to doing all of that activity in the household (coded 04). The
full set of items is given in Table 3.
Identity discrepancy is operationalized as the amount of disagreement between
ones self-rating in each of the spousal activities and the partners rating of the
self in each of these activities. The absolute difference between the two scores is
calculated. Given the response categories, a maximum disagreement of four in an
area would arise when the respondent reported he or she should perform all of an
activity and the partner reported that the respondent should perform none of the
activity (or vice versa). The disagreement scores were averaged across the 11 areas
with a theoretical range of 0 (perfect agreement) to 4 (maximum disagreement).
Self-worth is one of the two components of self-esteem, the other being self-
efcacy (Gecas & Schwalbe, 1983). Cast and Burke (2002) proposed the measure
Identities, Events, and Moods 33
Table 3. Spouse Identity Items.
Item: To what extent should you (your spouse) be responsible for . . .
. . . cleaning the house?
. . . preparing and serving meals?
. . . washing, ironing and mending the clothes?
. . . home repair?
. . . yard work?
. . . taking care of the bills and accounts?
. . . shopping for groceries?
. . . maintaining contact with parents and in-laws or other members of the family?
. . . providing the family income before children are born?
. . . providing the family income after children are born?
Table 4. Self-Worth Scale Items.
Item Loading
I feel I am a person of worth, at least on an equal basis with others. 0.68
I feel that I have a number of good qualities. 0.67
I feel I do not have much to be proud of.
a
0.47
I take a positive attitude toward myself. 0.76
On the whole, I am satised with myself. 0.73
I wish I could have more respect for myself.
a
0.55
At times, I think I am not good at all.
a
0.59
Omega reliability 0.88
a
Reverse coded.
used here. It uses the seven items of the Rosenberg self-esteem scale (1979) that
tap into the self-worth (the other three relate more to self-efcacy). The items form
a single dimension and have an omega reliability of 0.88. The items were stan-
dardized, given a common orientation and summed, with higher scores reecting
higher levels of self-worth. The items are given in Table 4.
Finally, male is a binary variable, coded one if the respondent is male and zero
otherwise.
Analyses
Because the data included multiple observations of moods collected on individuals
over time, a two-level hierarchical regression analysis was conducted (Raudenbush
&Bryk, 2002). The basic idea is that a regression of mood levels on events (present
34 PETER J. BURKE
or absent) are run for each individual. This is the lowest level regression. The
regression coefcients in this lowest level are then modeled to be a function of
characteristics of the individual. This is the second level regression and shows
the moderating effects of the individual characteristics of longer-term identity
verication problems (discrepancy) and self-worth on the lowest level effects.
Year and sex were included at the higher-level in order to test whether the results
changed over time or by sex. The two-level model that was estimated is given
in the following equations. The level one model for unease/distress is given in
Eq. (1), while the level two model for this outcome is given in the set of equations
labeled (2).
Distress =
0
+
1
(Event) +
2
(DayAfter) +
3
(TwoAfter)
+
4
(SpouseEvent) +r (1)

0
=
00
+
01
(Discrepancy) +
02
(Worth) +
03
(Sex)
+
04
(Year2) +
05
(Year3) +u
0

1
=
10
+
11
(Discrepancy) +
12
(Worth) +
13
(Sex)
+
14
(Year2) +
15
(Year3) +u
1

2
=
20
+
21
(Discrepancy) +
22
(Worth) +
23
(Sex)
+
24
(Year2) +
25
(Year3) +u
2

3
=
30
+
31
(Discrepancy) +
32
(Worth) +
33
(Sex)
+
34
(Year2) +
35
(Year3) +u
3

4
=
40
+
41
(Discrepancy) +
42
(Worth) +
43
(Sex)
+
44
(Year2) +
45
(Year3) +u
4
(2)
This model indicates in Eq. (1) that the level of unease/distress a person reports
is a function of whether or not an event has occurred on that day, the day before,
two days before, or to the spouse. The beta coefcients indicate the magnitude of
the effects of the events on the mood reported. In addition, the level two model in
Eq. (2) indicates that the magnitude of each of these effects of events on moods
(the beta coefcients in Eq. (1)) may be moderated by characteristics of the person
being modeled. These characteristics are the level of self-worth of the person, and
the level of spousal identity discrepancy of the person, as well as the persons
sex. Finally, dummy variables for the possible changes in the effects over time are
included. All of these effects in the level two model (the gammas) are things that
change the betas in the level one equation.
The model for the level of activity/arousal that is reported is slightly different
than the model for unease/distress because I allow for the possibility of a direct
Identities, Events, and Moods 35
effect of distress on arousal. For this reason the level one model is given in Eq. (3)
with the changed level two equations given in set of equations labeled (4).
Arousal =
0
+
1
(Event) +
2
(DayAfter) +
3
(TwoAfter)
+
4
(SpouseEvent) +
5
(Distress) +r (3)

0
=
00
+
01
(Discrepancy) +
02
(Worth) +
03
(Sex)
+
04
(Year2) +
05
(Year3) +u
0

1
=
10
+
11
(Discrepancy) +
12
(Worth) +
13
(Sex)
+
14
(Year2) +
15
(Year3) +u
1

2
=
20
+
21
(Discrepancy) +
22
(Worth) +
23
(Sex)
+
24
(Year2) +
25
(Year3) +u
2

3
=
30
+
31
(Discrepancy) +
32
(Worth) +
33
(Sex)
+
34
(Year2) +
35
(Year3) +u
3

4
=
40
+
41
(Discrepancy) +
42
(Worth) +
43
(Sex)
+
44
(Year2) +
45
(Year3) +u
4

5
=
50
+
51
(Discrepancy) +
52
(Worth) +
53
(Sex)
+
54
(Year2) +
55
(Year3) +u
5
(4)
Hypothesis 1 is tested by the magnitude of the beta coefcients in Eq. (1), while
Hypotheses 1a and 1b are tested by the appropriate gamma coefcients in Eq. (2).
Hypothesis 2 is tested by the magnitude of the
5
coefcient in Eq. (3), while
Hypotheses 2a and 2b are tested by the
51
and
52
coefcients in Eq. (4). Finally,
Hypotheses 3, 3a, and 3b are tested by the other beta and gamma coefcients in
Eqs (3) and (4).
RESULTS
Moods Across the Day and Week
Before examining the results showing the effects of interruptions in identity ver-
ication on mood, I begin with an overview of the daily and weekly variations
in mood that were noted by respondents. Because respondents lled out the scale
at varying times of the day and responded as they were feeling at the time, they
present a picture of the ups and downs across the day and across the week. An
initial analysis of the time ordered data by hour over the four weeks shows a strong
autocorrelation between observations separated by 24 hours, i.e. how one feels at
36 PETER J. BURKE
Fig. 1. Distribution of Standardized Unease/Distress Levels across the Day for Day of
Event, Day After Event, Two Days After Event and Days with No Events.
any time of the day correlates strongly with how they feel at that time of every day.
This is true for both levels of unease/distress as well as levels of activity/arousal.
13
These results suggest that there are strong daily cycles of both unease/distress and
activity/arousal. Figure 1 shows average daily levels of unease/distress for four
groups of respondents: those who suffered an event within the last day, those
who suffered an event the day before, those who suffered an event two days before,
and those who did not suffer an event within that time-frame. Figure 2 shows the
average level of activity/arousal over the day for the same groups of persons.
We see in Fig. 1 that the level of unease/distress in response to an event is much
higher than the level when no event has occurred. Additionally, we see elevated
levels of unease/distress one and even two days later, though the two-day later
score is often not very different from the level of unease/distress when no event
has occurred. In addition, we see a daily rhythmto the level of unease/distress with
very high levels occurring early in the morning, between 5 and 6 am. A second
though smaller peak occurs early in the afternoon and a third peak occurs in the
early evening hours between 6 and 7 pm. We see a similar rhythm to the level
of unease/distress a day after the event though the overall levels are diminished.
Indeed, even on days when no event has occurred, there is a similar rhythm to
the levels of unease/distress with people feeling more unease/distress in the early
Identities, Events, and Moods 37
Fig. 2. Distribution of Standardized Activity/Arousal Levels across the Day for Day of
Event, Day After Event, Two Days After Event and Days with No Events.
morning hours and then afternoon and early evening, with diminished levels later
in the morning and later evening.
In Fig. 2 we see a different cyclic pattern for levels of activity/arousal. Overall,
there are two peaks of activity/arousal: in the middle morning and in the middle
afternoon. Noon, and Midnight are times of lowered activity. We also see some
evidence of lowered activity/arousal for people on days they have suffered an event.
This pattern of activity/arousal follows closely that suggested by Thayer (1996) as
the cycle of energy.
Turning now to the question of weekly cycles, Figs 3 and 4 show the average
levels of unease/distress and activity/arousal respectively for each day of the week.
Looking rst at unease/distress, we see that in the absence of an event, people have
the highest levels of unease/distress on Wednesday, the hump day of the week, and
they have the lowest levels of unease/distress on the weekends; results that are not
surprising. If an event has occurred, however, the pattern is somewhat different.
Mondays and Thursdays appear to be the worst days, while Wednesday and the
weekends appear to be the best days.
14
With respect to the level of activity/arousal, we see little change over the week,
though Friday and Saturday show somewhat elevated levels of activity/arousal.
Again, we see the reduced levels of activity/arousal in response to the occurrence
of an event.
38 PETER J. BURKE
Fig. 3. Distribution of Standardized Unease/Distress Levels across the Week for Day of
Event, Day After Event, Two Days After Event and Days with No Events.
Fig. 4. Distribution of Standardized Activity/Arousal Levels across the Week for Day of
Event, Day After Event, Two Days After Event and Days with No Events.
Identities, Events, and Moods 39
In these preliminary results, we see strong evidence of daily and weekly cycles
in both dimensions of mood indicating both biological rhythms as well as social
rhythms, the latter showing up especially with the weekly ups and downs of the
unease/distress levels. We also see some preliminary evidence of the elevation in
distress and depression of activity levels when identity-disconrming events occur,
even a day or two after the event. I turn nowto the analysis of the effects of identity
disrupting events on affective outcomes and tests of the hypotheses.
Identity Disruption and Mood
Before looking at the results for the hierarchical regression, Tables 5 and 6 present
the means, variances and standard deviations of the measures used in the study.
Among the level two variables, we note in Table 5 a negative correlation between
identity discrepancy and self-worth indicating that persons who are having trouble
verifying their spouse identity tend to have lower feelings of self-worth. Table 6
shows that feelings of unease/distress are associated with events (and continue in
the days after an event), and lowered levels of activity/arousal are also associated
with the occurrence of events (and continue in the days after an event). I also note
the small negative correlation between unease/distress and activity/arousal.
The results of the hierarchical regression are reported in Table 7 for un-
ease/distress, testing the rst hypotheses, and in Table 8 for activity/arousal, testing
the second hypotheses. Looking rst at Table 7 and the effects for the intercept
(level 1), we see that the baseline level of unease/distress (
00
) is signicantly
belowthe overall mean level of zero. Thus, on days when no event occurs (and it is
not the day after or two days after an event), people are feeling less unease/distress
than average. But, to the extent that the spousal identity is not being conrmed,
Table 5. Means, Standard Deviations and Correlations for Level-Two Variables
(N = 1443).
Self-Worth Discrepancy Male Year 2 Year 3
Self-worth 1.00
Discrepancy 0.15
*
1.00
Male 0.01 0.00 1.00
Year 2 0.01 0.02 0.00 1.00
Year 3 0.04 0.02 0.00 0.43
*
1.00
Mean 0.36 0.00 0.50 0.32 0.28
Std. Dev. 0.20 0.70 0.50 0.47 0.45

p 0.05.
40 PETER J. BURKE
Table 6. Means, Standard Deviations and Correlations for Level-One Variables
(N = 29,291).
Event Day After Two Days Spouse Unease/ Activity/
Event After Event Distress Arousal
Event 1.00
Day after event 0.11
*
1.00
Two days after 0.09
*
0.13
*
1.00
Spouse event 0.28
*
0.06
*
0.03
*
1.00
Unease/distress 0.18
*
0.05
*
0.02
*
0.09
*
1.00
Activity/arousal 0.06
*
0.04
*
0.04
*
0.04
*
0.08
*
1.00
Mean 0.09 0.08 0.07 0.09 0.01 0.02
Std. Dev. 0.29 0.27 0.26 0.29 0.78 0.79

p 0.05.
the average level of unease/distress is signicantly increased. On the other hand,
to the extent that the person has higher self-worth, their level of unease/distress is
signicantly decreased. These results also show that males have a higher level of
unease/distress than females, and there is a slightly higher level of distress in the
second year.
Turning now to the effects of an event, we see that the average level of
unease/distress is increased signicantly on the day an event occurs (
10
) thus
conrming our Hypothesis 1. This effect is magnied to the extent that the spousal
identity of the person is not being conrmed. Increases in the amount of discrep-
ancy (lack of verication) bring about increases in the level of unease/distress
caused by an event. This moderating effect conrms Hypothesis 1a. Finally, we
see there is not support for Hypothesis 1b concerning the moderating effects
of self-worth. Apparently persons with high as well as low levels of self-worth
equally feel the impact of an event.
Although I did not have hypotheses for the remaining results, these exploratory
analyses show that feelings of unease/distress continue into the day after an event
has occurred. Two days after an event, the average person is over the mood change
unless they have lower feelings of self-worth, in which case they do continue to
have feelings of unease/distress.
Looking at the effects of an event that happens to the spouse, Table 7 shows that
people do feel heightened levels of unease/distress when their spouse receives an
identity-disconrming event independent of any events that may have happened to
themselves. And, this effect is signicantly increased if there are already problems
with verication of the spousal identity. Interestingly, women feel the effects of a
spousal event to a greater extent than do men which may reect womens greater
sensitivity to relationships and greater interdependence compared to men.
Identities, Events, and Moods 41
Table 7. Hierarchical Linear Model Results for Unease/Distress.
Level One Effect Level Two Effect Coefcient
Intercept (
0
) Intercept (
00
) 0.13
Discrepancy 0.14
Self-worth 0.22
Male 0.12
Year 2 0.08
Year 3
Event (
1
) Intercept (
10
) 0.50
Discrepancy 0.23
Self-worth
Male
Year 2
Year 3
Day after event (
2
) Intercept (
20
) 0.09
Discrepancy
Self-worth
Male
Year 2
Year 3
Two days after event (
3
) Intercept (
30
)
Discrepancy
Self-worth 0.07
Male
Year 2
Year 3
Spouse event (
4
) Intercept (
40
) 0.13
Discrepancy 0.18
Self-worth
Male 0.06
Year 2
Year 3
: p > 0.05.
I turn now to the remaining hypothesis concerning the consequences of
unease/distress and identity disruption on the activity/arousal mood. The results
of this analysis are presented in Table 8. We see that overall persons who are
experiencing ongoing problems with spousal identity verication (discrepancy)
report signicantly lower levels of activity/arousal, and we also note the higher
level of activity/arousal reported by men over that reported by women.
With respect to Hypothesis 2, we see that unease/distress reduces the level of
activity/arousal (
40
) as hypothesized. Additionally, in support of Hypothesis 2a,
42 PETER J. BURKE
Table 8. Hierarchical Linear Model Results for Activity/Arousal.
Level One Effect Level Two Effect Coefcient
Intercept (
0
) Intercept (
00
) 0.07
Discrepancy 0.26
Self-worth
Male 0.16
Year 2
Year 3
Event (
1
) Intercept (
10
) 0.03
Discrepancy
Self-worth
Male
Year 2
Year 3
Day after event (
2
) Intercept (
20
)
Discrepancy
Self-worth
Male
Year 2
Year 3
Two days after event (
3
) Intercept (
30
)
Discrepancy
Self-worth
Male
Year 2
Year 3
Spouse event (
4
) Intercept (
40
)
Discrepancy
Self-worth
Male
Year 2
Year 3
Unease/distress (
5
) Intercept (
40
) 0.10
Discrepancy
Self-worth 0.05
Male
Year 2 0.06
Year 3
: p > 0.05.
Identities, Events, and Moods 43
we see that the effect of unease/distress is moderated by level of self-worth. Persons
with higher levels of self-worth do not suffer the same effects on activity/ arousal
from distress that persons with lower levels of self-worth do. On the other hand,
contrary to Hypothesis 2b, there are no moderating effects of identity discrepancy.
Finally, with respect to Hypothesis 3, we do see a direct effect of an event in
reducing levels of activity/arousal independent of the current level of reported
distress. This effect is not moderated by either self-worth or spousal identity veri-
cation problems so that Hypotheses 3a and 3b are not supported.
DISCUSSION
According to identity control theory, the lack of identity verication is distressful
(Burke, 1991, 1996). While a number of studies have tested this for ongoing
relationships (e.g. Burke & Stets, 1999; Stets, 1997; Stets & Burke, 1996), no
studies have examinedthe process as it inuences individuals ona day-to-daybasis.
The study of moods however, affords us the opportunity to examine this process.
According to Morris (1992), moods exist for the sake of signaling states of the
self in terms of the physical, psychological, and social resources available to meet
perceived environmental demands. According to identity control theory, identities
manipulate signs and symbols that indirectly control active and potential resources
in the situation through the process of self-verication (Burke, 2003; Freese &
Burke, 1994). Within identity control theory, therefore, the lack of verication
comes about because active or potential resources are not brought to the levels
indicated in ones identity standard as the reference value. Moods thus indicate the
lack of identity verication when resources are not available to meet the reference
level of the identity standard.
In the present paper I examine the impact of identity disconrming life-events on
the moods that people report. The twoprimarydimensions of mood, unease/distress
and activity/arousal, were measured daily for a period of four weeks at three time
points, each separated by a year. At the same time it was noted whether an identity-
disconrming event had occurred on each day, the day before or two days before.
Fromidentity control theory, it was hypothesized that these identity-disconrming
events would increase unease/distress and decrease the level of activity/arousal.
And, because we already know that persons who are having trouble verifying an
identity over a period of time are susceptible to more distress, while people who
have higher levels of self-worth are less susceptible to distress, it was hypothesized
that these conditions would moderate the consequences of the daily events.
Overall, the result show that persons with problems in verifying their spousal
identity show higher levels of unease/distress and lower levels of activity/arousal.
44 PETER J. BURKE
Additionally, persons with higher levels of self-worth experience lower levels of
unease/distress, but levels of activity/arousal are not changed.
Turning to the specic hypotheses we see that in general they were conrmed.
People who experienced an identity-disconrming life-event on a particular day
reported higher levels of unease/distress, and this mood persisted into the next day,
though at a somewhat reduced level. And, for persons who had problems verifying
their spousal identity, these effects were strengthened. I did not nd, however,
that these effects were diminished for persons with higher levels of self-worth as
had been hypothesized. These same effects were noted when the event occurred
not to the self, but to the spouse. Because the spouse is important for verifying
ones own spousal identity, an event that disturbed the spouses identity was
apparently felt as threatening to ones own identity, and even more so when
the verication of the spousal identity was itself already problematic or when
the respondent was female and therefore likely to be more interdependent and
relationship oriented.
With respect to the second dimension of mood, i.e. activity/arousal, the effects
were somewhat weaker, though generally consistent with the hypothesis. However,
the effect did not persist until the next day, nor was it moderated by self-worth or
problems with verifying the spousal identity. Thus, while the events themselves
had only a small depressing effect of activity/arousal, there was a larger indirect
effect of the event through its effect on the level of unease/distress. The level of
self-worth moderated this latter, indirect effect, with persons feeling the effect less
if they had higher levels of self-worth.
The overall picture is thus one that conrms the predictions of identity control
theory about the consequences of even temporary dislocations of the identity pro-
cesses in negatively altering a persons mood for a period of time, usually not more
than a day or two. Although in separate analyses not reported, I did not nd an
added effect for another event that occurs on the day following an event (that is
the second event had no more effect than the rst event), according to the results
presented, successive events are expected to continue the heightened level of the
distressed mood. This continued effect of the second or even third event would be
on top of the already noted persisting effect on the day after an event.
Identity disconrmation by daily events does alter the moods reported by people.
This outcome provides additional evidence for the functional analysis of mood
suggested by Morris (1992). Both identity control theory and Morris functional
theory of moods converge on the notion of resources. They may differ, however,
in that Morris discusses only the lack of resources to handle needs, while identity
control theory suggests that problems also arise when resources are at levels
that are higher than they should be (Burke & Harrod, 2002). Future research needs
to explore this possibility for the effects of events that disconrm identities by
providing too much of a good thing.
Identities, Events, and Moods 45
One additional area of future research relating identity verication and moods
lies in the already noted apparent correspondence between the two primary
dimensions of mood (unease/distress and activity/arousal) and the two common
classes of psychiatric disorder (anxiety and depression). Higgins has explored
the relationship between self-verication and the anxiety/depression distinction
(Higgins, 1987). In doing this, he has distinguished between what he calls the
ought-self and the ideal-self. The rst of this consists of those aspects of ones self
or an identity that are expectations held by others about what one ought to do in
a particular role, for example. The second consists of those aspects of an identity
that are ideals or wishes about the way we want to be. Higgins has suggested,
and shown, that the lack of verication of the ought-self leads to symptoms
of distress, while the lack of verication of the ideal-self leads to depressed
symptoms (cf. Higgins, 1987, 1989; Higgins et al., 1987, 1994). Marcussen and
Large (Marcussen & Large, 2003) have applied this idea to identity control theory.
In the present research we have no way of distinguishing between events that
disturb identity verication of an ought variety from one that disturb identity
verication of an ideal variety. However, future research could perhaps do this
and examine whether these two types of identity disruption each differentially
affect the two dimensions of mood (unease/distress and activity/arousal) in the
manner that would be expected. Should that be the case, one could then begin to
theorize more fruitfully about the nature of the resources that are under the control
of each system.
NOTES
1. There is a third, more micro, level that may be analyzed. This level examines the
effects of the more minor disturbances to identity verication that occur in the give-and-
take of ongoing interaction. Some of the work of Gottman (1982, 1987, 1993), though not
cast in identity terms, may be viewed in this way.
2. Concerning the three levels discussed earlier (long-term, ongoing problems of identity
verication, short-term, occasional identity disturbing events, and generally minor distur-
bances felt in the micro-processes of normal interaction), moods may be appropriate to the
middle level and emotions to the more micro-level. However, there is not an unambiguous
line between levels, so the effects are likely to blend from one level to another.
3. The conceptual similarity between these two dimensions and the two most common
classes of psychiatric disorders (anxiety and depression) can be noted (Sapolsky, 2003).
4. A fuller description of the data and data collection process is available elsewhere
(Tallman et al., 1998).
5. Examples of reported events include statements such as, My girlfriend who has had
a heart transplant went in for her usual yearly check-up. This is her second year. All is good.
Doctors are happy! I organized house and boxes all morning, Grandma surprised me with
a visit and brought fruit, cookies, and neat platter to match our dishes, then helped me with
my work, and I got three interviews this week for a new job.
46 PETER J. BURKE
6. Examples of reported events include statements such as My cousin received good
and bad news. She found out that there is a medication that will help her, but wont be
good for her six-week-old fetus, A friend at work has a daughter who lives in St. Annes.
Her health is very bad. She is in the hospital now. I dont know the outcome, and (My)
mother-in-law is having health problems.
7. Examples of reported events include statements such as I lost my job, we ran out of
coupon books to sell. I would have lost my job on the 31st anyway, Close friend didnt
get job she was hoping for near where we live, Erik feels that he is being harassed on the
job. His boss is always picking on him, and My manager keeps telling me to do stuff I
know will be wrong. And then when it does end up being wrong I get yelled at and insulted
and some times physically threatened.
8. Examples of reported events include statements such as, Two of my friends want to
get married and theyve only known Each other for about two weeks, My friend who is
19 just discovered she is pregnant for the third time and she is planning to have an abortion,
which I feel is very wrong, and I found out one of my best friends said some things that
werent very nice.
9. Examples of reported events include statements such as Tried to return a shower gift
at Victorias Secret and we could only exchange or get store credit, even with the sales tag,
Randy and I went to the mall to buy videotape. We were in a hurry and the store was really
busy. The salespeople were slow and ignored us, and I tried to gure out our income tax
for this year. It looks like we will end up owing. I hate owing taxes to the government. I
havent itemized yet so it could be better than I think. I still think we will probably owe
money.
10. Examples of reported events include statements such as, My parents arent com-
municating at all. My mom is angry with my dad for things ve years ago and my dad has
completely forgotten them, My husband and I argued over money he spent that was partly
mine and I became angry because we dont have much spending money, and My brother
backed out of this tournament and, considering he bowls better than I do, I was concerned
about being able to live up to his average.
11. Two items, aroused and passive, that had been included in the daily diaries had
uniqueness values greater than 0.7 and were excluded. The item aroused may have had
sexual connotations for some persons but not others, while the passive item may have
had connotations of (lack of) control for some but not others.
12. Ideally, we would want to measure the individuals reected appraisals, i.e. their
perceptions of their spouses expectations for them, but we do not have this measure. I use
the spouses actual views as a proxy for the perceptions.
13. There is one signicant autocorrelation of 0.33 with a lag of 24 hours for un-
ease/distress. For activity/arousal, there are signicant autocorrelations of about 0.60 cen-
tering on lags of 1 hour and 24 hours.
14. Additional analyses not reported show that this pattern is not the result simply of
events being more likely on days there is higher levels of unease/distress.
ACKNOWLEDGMENTS
The research reported in this paper is based on data from a longitudinal study of
rst-married couples, Socialization into Marital Roles, funded by NIMH grant
Identities, Events, and Moods 47
MH46828, under the direction of Irving Tallman, Peter J. Burke, and Viktor Gecas.
I would like to thank Jan E. Stets for her comments on an earlier version.
REFERENCES
Baumeister, R. F. (1998). The self. In: D. T. Gilbert & S. T. Fiske (Eds), The Handbook of Social
Psychology (4th ed., Vol. 2, pp. 680740). Boston, MA: McGraw-Hill.
Blaine, B., & Crocker, J. (1993). Self-esteem and self-serving biases in reactions to positive and
negative events: An integrative view. In: R. F. Baumeister (Ed.), Self-Esteem: The Puzzle of
Low Self-Regard (pp. 5585). New York, NY: Plenum Press.
Brown, G. W. (1974). Meaning, Measurement, and Stress of Life Events. In: B. S. Dohrenwend &B. P.
Dohrenwend (Eds), Stressful Life Events: Their Nature and Effects (pp. 217244). New York:
Wiley.
Brown, J. D., & McGill, K. L. (1989). The cost of good fortune: When positive life events produce
negative health consequences. Journal of Personality and Social Psychology, 57, 11031110.
Burke, P. J. (1991). Identity processes and social stress. American Sociological Review, 56, 836
849.
Burke, P. J. (1996). Social identities and psychosocial stress. In: H. B. Kaplan (Ed.), Psychosocial
Stress: Perspectives on Structure, Theory, Life Course, and Methods (pp. 141174). Orlando,
FL: Academic Press.
Burke, P. J. (2003). Identities and social structure. Address given at the Cooley-Mead Award Ceremony,
American Sociological Association Section on Social Psychology, Atlanta.
Burke, P. J., & Harrod, M. M. (2002). To good to be believed? Presented at the annual meeting of the
American Sociological Association, Chicago.
Burke, P. J., & Stets, J. E. (1999). Trust and commitment through self-verication. Social Psychology
Quarterly, 62, 347366.
Campbell, J. D., Chew, B., & Scratchley, L. S. (1990). Cognitive and emotional reactions to daily
events: The effects of self-esteem and self-complexity. Journal of Personality, 59, 473505.
Cast, A. D., & Burke, P. J. (2002). A theory of self-esteem. Social Forces, 80, 10411068.
Cochran, S. D., & Hammen, C. L. (1985). Perceptions of stressful life events and depression: A test of
attributional models. Journal of Personality and Social Psychology, 48, 15621571.
Dohrenwend, B. S. (1973). Life events as stressors: A methodological inquiry. Journal of Health and
Social Behavior, 167175.
Ekman, P. (1994). Moods, emotions, and traits. In: P. Ekman & R. J. Davidson (Eds), The Nature of
Emotion: Fundamental Questions (pp. 5658). New York: Oxford University Press.
Ellestad, J., & Stets, J. E. (1998). Jealousy and parenting: Predicting emotions from identity theory.
Sociological Perspectives, 41, 639668.
Freese, L., & Burke, P. J. (1994). Persons, identities, and social interaction. In: B. Markovsky, K.
Heimer & J. OBrien (Eds), Advances in Group Processes (Vol. 11, pp. 124). Greenwich, CT:
JAI Press.
Frijda, N. H. (1993). Moods, emotion episodes, and emotions. In: M. Lewis & J. M. Haviland (Eds),
Handbook of Emotions (pp. 381404). New York: Guilford.
Frijda, N. H. (1994). Varieties of affect: Emotions and episodes, moods, and sentiments. In: P. Ekman &
R. J. Davidson (Eds), The Nature of Emotion: Fundamental Questions (pp. 6067). New York:
Oxford University Press.
48 PETER J. BURKE
Gecas, V., &Schwalbe, M. L. (1983). Beyond the looking-glass self: Social structure and efcacy-based
self-esteem. Social Psychology Quarterly, 46, 7788.
Gottman, J. M. (1982). Emotional responsiveness in marital conversations. Journal of Communication,
32, 108120.
Gottman, J. M. (1987). The sequential analysis of family interaction. In: T. Jacob et al. (Eds), Family In-
teraction and Psychopathology: Theories, Methods, and Findings, Applied Clinical Psychology
(pp. 453478). New York, NY: Plenum Press.
Gottman, J. M. (1993). Studying emotion in social interaction. In: M. Lewis & J. M. Haviland (Eds),
Handbook of Emotions (pp. 475487). New York, NY: Guilford Press.
Hammen, C., Marks, T., DeMayo, R., & Mayol, A. (1985). Self-schemas and risk for depression: A
prospective study. Journal of Personality & Social Psychology, 49, 11471159.
Hammen, C., Marks, T., Mayol, A., & DeMayo, R. (1985). Depressive self-schemas, life stress, and
vulnerability to depression. Journal of Abnormal Psychology, 94, 308319.
Higgins, E. T. (1987). Self-discrepancy: A theory relating self and affect. Psychological Review, 94,
319340.
Higgins, E. T. (1989). Self-discrepancy theory: What patterns of self-beliefs cause people to suffer?
In: L. Berkowitz (Ed.), Advances in Experimental Social Psychology (Vol. 22, pp. 93136).
New York: Academic Press.
Higgins, E. T., Klein, R. L., & Strauman, T. J. (1987). Self discrepancies: Distinguishing among self-
states, self-state conicts, and emotional vulnerabilities. In: K. Yardley & T. Honess (Eds), Self
and Identity: Psychosocial Perspectives (pp. 173186). New York: Wiley.
Higgins, E. T., Roney, C. J. R., Crowe, E., & Hymes, C. (1994). Ideal vs. ought predilections for
approach and avoidance Distinct self-regulatory systems. Journal of Personality and Social
Psychology, 66, 276286.
Koehn, P. M. (2001). Physical activity and depression in adolescents: An extension of the biopsychoso-
cial model. Dissertation Abstracts International: Section B: The Sciences and Engineering, 61,
4989.
Marcussen, K., & Large, M. D. (2003). Using identity discrepancy theory to predict psychological
distress. In: P. J. Burke, T. J. Owens, R. T. Serpe & P. A. Thoits (Eds), Advances in Identity
Theory and Research. New York: Kluwer/Plenum.
Morris, W. N. (1992). A functional analysis of the role of mood in affective systems. In: M. S. Clark
(Ed.), Emotions (pp. 256293). Newbury Park, CA: Sage.
Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data analysis
methods. Thousand Oaks: Sage.
Rosenberg, M. (1979). Conceiving the self. New York: Basic Books.
Sapolsky, R. (2003). Taming stress. Scientic American, 289, 8795.
Stets, J. E. (1997). Status and identity in marital interaction. Social Psychology Quarterly, 60, 185217.
Stets, J. E. (2003). Justice, emotion, and identity theory. In: P. J. Burke, T. J. Owens, P. A. Thoits
& R. Serpe (Eds), Advances in Identity Theory and Research (pp. 105122). New York:
Kluwer/Plenum.
Stets, J. E., & Burke, P. J. (1996). Gender, control, and interaction. Social Psychology Quarterly, 59,
193220.
Stets, J. E., & Burke, P. J. (2003). Self-verication, control, and aggression. Under review.
Stets, J. E., & Tsushima, T. (2001). Negative emotion and coping responses within identity control
theory. Social Psychology Quarterly, 64, 283295.
Stouffer Calderon, K. (2001). Making the connection between depression and activity levels among
the oldest-old: A measure of life satisfaction. Activities, Adaptation and Aging, 25, 5973.
Identities, Events, and Moods 49
Swann, W. B., Jr., de la Ronde, C., & Hixon, J. G. (1994). Authenticity and positivity strivings in
marriage and courtship. Journal of Personality and Social Psychology, 66, 857869.
Swann, W. B., Jr., Hixon, J. G., & de la Ronde, C. (1992). Embracing the bitter truth: Negative
self-concepts and marital commitment. Psychological Science, 3, 118121.
Thayer, R. E. (1996). The origin of everyday moods. New York: Oxford University Press.
Thoits, P. A. (1978). Life events, social integration, and psychological distress. Dissertation Abstracts
International, 39, 3869.
Thoits, P. A. (1991). On merging identity theory and stress research. Social Psychology Quarterly, 54,
101112.
Watson, D., & Tellegen, A. (1985). Toward a consensual structure of mood. Psychological Bulletin,
98, 219235.
Williamson, G. M., & Shaffer, D. R. (2000). The activity restriction model of depressed affect:
Antecedents and consequences of restricted normal activities.
EMOTIONS IN IDENTITY THEORY:
THE EFFECT OF STATUS
Jan E. Stets
ABSTRACT
This research continues to advance the role of emotion in identity theory
by examining how the external social structure inuences internal identity
processes to produce negative emotions. According to identity control theory,
negative arousal emerges when one experiences identity feedback that is
non-verifying, persistent, and from a source who is familiar compared to
unfamiliar to one. While other research has not denitively supported these
relationships (Stets, 2003, 2005), the current research examines whether
the identity theory hypotheses are conditioned upon ones status in the
social structure. Using the diffuse status characteristic of gender where
the status of male is high and the status of female is low, I investigate the
role of status (both as the recipient and source of non-verifying identity
feedback), persistence, and familiarity in producing negative emotions. The
data are based on a laboratory experiment that simulates a work situation
and invokes the worker identity. Workers of high or low status are the
recipients of identity non-verication that is persistent or non-persistent
and that is from a familiar or unfamiliar other. Managers of high or
low status and who are familiar or unfamiliar with the workers are the
source of persistent or non-persistent identity non-verication. The results
reveal that the status of actors both as the recipient and source of identity
non-verication are signicant for negative emotions, suggesting that status
Theory and Research on Human Emotions
Advances in Group Processes, Volume 21, 5176
Copyright 2004 by Elsevier Ltd.
All rights of reproduction in any form reserved
ISSN: 0882-6145/doi:10.1016/S0882-6145(04)21003-7
51
52 JAN E. STETS
effects need to be incorporated into the theoretical development of emotions in
identity theory.
EMOTIONS IN IDENTITY THEORY:
THE EFFECT OF STATUS
Given the request by Stryker and Burke (2000) that researchers make more explicit
how emotions t into the symbolic interaction framework of identity theory (IT),
developing theory and research in Identity Control Theory (ICT) has begun to take
up this challenge (Stets, 2003, 2005; Stets &Tsushima, 2001).
1
Broadly speaking,
identity theorists maintain that positive emotion results from meeting ones
identity expectations and negative emotion results from not meeting ones identity
expectations (Stets, 2003, 2005). For example, Burke and his colleagues argue that
continuous identity verication (congruence between perceived meanings of the
self in the situation and meanings held in the identity standard) registers positive
emotion, and a lack of identity verication (incongruence between self-in-situation
meanings and identity standard meanings) that cannot be handled automatically
within the self-regulatory system registers negative emotion (Bartels, 1997; Burke
& Stets, 1999; Cast & Burke, 2002). Similarly, Stryker (1987) maintains that
identities that are adequately played out will generate positive feelings, while
identities that are inadequately played out will generate negative feelings. The
latter signals that self-in-situation meanings are not supported by others. Thus, the
lack of support from others as to who one is registers negative feelings for Stryker
in the same way that a lack of identity verication registers negative feelings for
Burke. The current research develops this theory of emotion further.
The role of emotion in identity theory is closely aligned to affect control theory
(ACT), another variant of symbolic interactionism (Heise, 1979; Smith-Lovin
& Heise, 1988). Like IT, in ACT, an emotion signals the extent to which events
conrm or disconrm ones identity in a situation. Specically, when an event in a
situation does not meet fundamental sentiments (this corresponds to identity stan-
dard meanings in ICT) because situational meanings labeled transient impressions
(analogous to self-in-situation meanings in ICT) differ from the fundamental
sentiments, a deection (incongruence in ICT terms) has occurred. The greater
the deection, the higher the level of emotional arousal. In response, persons
will create new events that restore situational meanings to fundamental, identity
standard meanings. One difference between ACT and IT is that affect control
theorists maintain that a deection can generate positive emotion when transient
impressions are more positive than the fundamental sentiments of an identity.
Identity theorists assume that negative emotions emerge from all incongruence.
Emotions in Identity Theory: The Effect of Status 53
Other social psychological theories are similar to IT and ACT in that emotions
are aroused on the basis of the degree of congruity between what one expects
and what one obtains in a situation (Turner, 1999). For example, in expectations
states theory, positive emotion results when expectations associated with the
distribution of status in a situation are met, for example, deference is given to
high status people (Ridgeway & Johnson, 1990). When status expectations are
not met, negative emotion occurs. A similar assumption is made in Kempers
(1990) theory and Collins (1990) theory on power and emotion in interaction.
When expectations associated with power and status positions are met, positive
emotion results, and unmet expectations result in negative emotion. In exchange
theory, repeated exchange agreements generate an emotional buzz between actors
in the form of satisfaction while repeated exchange disagreements generate
dissatisfaction (Lawler & Yoon, 1996). In distributive justice theory, an evaluation
of congruence between ones actual outcomes and ones expected outcomes in a
situation leads to positive emotion, while the evaluation of incongruence leads to
negative emotion (Hegtvedt & Markovsky, 1995).
2
In the current research, I further develop the role of emotion in identity theory
by investigating how ones status in the social structure effects emotions. This is
important for two reasons. First, in examining how the external structure impinges
on internal identity processes, I move toward integrating the two parts of identity
theory: one that emphasizes social structural sources of self processes, and the
other which focuses on the internal dynamics of self processes (Stryker & Burke,
2000). Second, by addressing status, I bring ideas from expectation states theory
(Ridgeway & Walker, 1995) and status characteristics theory (Wagner & Berger,
1993) into identity theory, thereby relating ideas across theoretical research
programs (Stryker & Burke, 2000).
Of theoretical concern in this research is whether high status persons compared
to low status persons experience more intense emotions when: (1) there is a lack
of identity verication; (2) the lack of identity verication is persistent; (3) the
lack of identity verication comes from a source that is an important other; and
(4) the lack of identity verication comes from a low status, important other. The
thesis is that because high status persons are more likely than low status persons
to experience identity verication (Cast et al., 1999; Stets & Harrod, 2004), they
will be more likely to report negative emotional arousal when identity verication
is not forthcoming. This negative emotional arousal will be heightened when the
identity disconrmation is frequent, and the source of the disconrmation is a low
status other. In the latter, low status persons will be more likely than high status
persons to be perceived as lacking the authority to disrupt the verication process.
An experiment that simulates a work situation and invokes the worker identity is
carried out. Male and female participants (high and low status workers) carry out
54 JAN E. STETS
some simple tasks, and they learn that another subject (a male or female (high or
lowstatus) manager) in the situation will give thempoints/feedback following their
task performance. Participants are given points that reect either: (1) what they
would expect to get in the worker role (verication); (2) more than what they would
expect to get (lack of verication in a positive direction); or (3) less than what they
would (lack of verication in a negative direction). The points are administered
at prearranged intervals throughout participants task performance (persistent
identity verication) or only once after task completion (non-persistent identity
verication). Before participants begin their work, they either have the opportunity
to get to knowthe other person who will be involved in the study with them(famil-
iarity (importance) condition), or they are not given this opportunity (unfamiliar
(unimportance) condition). Self-reports on participants feelings are obtained
throughout the study.
THEORY
The Identity Process
An identity is the categorization of the self as an occupant of a role; an actor
incorporates into the self all the meanings and expectations associated with the
role and its performance (Stets & Burke, 2000). These are actors role identities.
Persons act within the context of social structure in which they and others are
labeled as an occupant of roles in society (Stryker, 1980). Emotion in identity
theory is related to how effectively ones role identity is played out.
Stryker (1987) indicates that inadequate role performance leads to negative
emotion and adequate role performance produces positive emotion.
3
Identities
that lead to negative feelings should be less likely to played out in situations and
move down in the salience hierarchy of identities, while identities that generate
positive feelings should be played out more often in situations and move up
in the salience hierarchy. Since salient identities are more likely to be invoked
in situations (Stryker, 1980), identities that generate positive emotion are more
likely to be activated in situations. In general, salient identities are threatened to
the extent that there is a lack of support for these identities from others because
of poor role performance. Actors experience negative emotions as a result of
diminished support, thereby challenging the maintenance of their identities.
Burke (1991, 1996) has a similar view on the role of emotion in the identity
process. Focusing on the internal dynamics of the self, Burke (1991) maintains
that when an identity is activated in a situation, a feedback loop is established
(see Fig. 1). The identity standard denes what it means to be who one is in a
Emotions in Identity Theory: The Effect of Status 55
Fig. 1. Identity Control Model.
situation. The input contains the meanings of how one sees itself in a situation.
These self-meanings are often derived from how others see the self, labeled the
reected appraisals. The comparator compares the meanings from the input
and the meanings from the standard and registers the difference or error between
them. The result of this comparison is the output. This is behavior that acts upon
the environment either to maintain congruence between the input and standard
meanings by continuing to act, uninterrupted, or by acting differently when a large
error is registered, in order to change the input meanings to match the meanings
in the standard.
As shown in Fig. 1, emotion (like the output/behavior) results from the
comparison between input meanings and identity standard meanings. Continuous
congruence between input and identity standard meanings what is known as
self-verication registers positive emotion. Incongruence that cannot be handled
automatically by the system a lack of self-verication registers negative
emotion. The role of emotion given these internal, identity control processes
parallels the role of emotion in Strykers discussion of identities, role performance,
and social support. Both Stryker and Burke associate negative emotions with
unmet identity expectations; how others perceive the self differs from ones own
self-perceptions. Positive emotions are equated with identity expectations being
met. I make a similar assumption in this study.
56 JAN E. STETS
Emotions vary in terms of their strength or intensity. Stryker (1987) argues that
the strength of an emotion is a function of how important (not how salient) an
identity is in ones salience hierarchy with more important identities producing
a stronger emotion.
4
Burke (1991) similarly argues that stronger emotions result
from the disruption of more salient and committed identities. However, Burke
also hypothesizes that greater emotional distress will be related to how frequent
the identity process is interrupted, and whether the source of the interruption is a
signicant other. I test this in this research.
Frequent interruptions in the identity verication process, or what I label
persistent interruptions, are expected to lead to more intense negative arousal than
infrequent or occasional interruptions.
5
In the ICT model, an interruption consti-
tutes a break in the smooth owing cybernetic feedback loop that characterizes the
identity process. The interruption may be a function of the feedback loop breaking
on the output side of the identity model where ones behavior does not generate
feedback from others that who one is in the situation approximates ones identity
standard meanings. This is labeled a Type 1a interruption (Burke, 1991, 1996).
Alternatively, the interruption may be a function of the feedback loop breaking
on the input side of the identity model where self-in-situation meanings are not
perceived as congruent with identity standard meanings, even when the behavior
is effective. This is a Type 1b interruption (Burke, 1991, 1996). In this study, inter-
ruption is likely to be due to a Type 1a interruption where ones behavior does not
inuence how others treat the individual. Others may either ignore ones behavior,
or they may impose meaning onto the person, independent of the persons behavior
and intentions behind those behaviors. The more frequent the interruptions, the
more the continuously operating identity processes is prevented, and the more
negative the emotions that ensue.
6
Interruptions from a signicant other, or what I identify as a familiar other,
should also lead to more intense negative arousal than an interruption from an
unfamiliar other. According to Burke, signicant others are those with whom one
has built up a set of mutually veried expectations. Patterns of interaction support
each others identity in situations. The meanings that have been built up form
a tightly organized process. If this process gets interrupted, it should produce
more distress than the interactive patterns that emerge (and that are unlikely
to be as supportive or patterned) with non-signicant others. This is consistent
with Mandlers (1982) assumption that the interruption of more highly organized
processes will lead to higher levels of autonomic arousal.
Extrapolating from this, a familiar other will be perceived as someone who will
verify ones expectations. Consistent with this is the assumption that a familiar
other is safe and unlikely to harm us compared to an unfamiliar other (Berscheid
& Reis, 1998). Indeed, evidence on the mere exposure effect reveals that simply
Emotions in Identity Theory: The Effect of Status 57
providing another with some exposure to a stimulus is enough to generate liking
for that stimulus (Zajonc, 2000). Thus, negative arousal should be more likely
to emerge when a familiar other violates the implicit expectation of security and
predictability and verication of expectations compared to an unfamiliar other.
Using the same data that is used in the current study, recent research tested the
ICT predictions regarding emotional arousal and the lack of verication, as well
as the frequency of the verication interruption and its source (Stets, 2005). The
results showed that identity non-verication did inuence negative arousal, but
more for identity non-verication in a negative direction than identity verication
in a positive direction. Surprisingly, frequent or persistent non-verication
dampened peoples emotional responses rather than strengthening them. A
strong emotional response signals a discrepancy between situational meanings
and meanings in the standard, while a weaker emotional response suggests an
increasing correspondence between situational and standard meanings. Thus, the
results suggested that people are adjusting their identity standards to the feedback
that they are receiving in the situation.
Stets (2005) further found that interruptions from an important or familiar other
intensied negative emotions in highly specied conditions: for non-verication
in a positive direction that was not persistent. Becoming familiar with another
for a short period of time (a proxy for an important other) was sufcient in
formulating the expectation that the other would conrm ones expectations, and
when that did not occur, negative emotions emerged. However, the effect occurred
when the feedback from another was more positive than ones own standard,
and the feedback occurred once. In general, Stets ndings revealed that ones
emotional responses to non-verifying identity contexts as predicted by ICT may
be conditional, dependent upon the context and other factors being present.
The Current Study
In the present research, I incorporate social structural features into the identity
control process and examine ones status in the social structure, specically ones
gender.
7
Following expectations states theory, the status of males is regarded as
high in the social structure while the status of females is regarded as low(Ridgeway,
1993). I investigate ones gender both in terms of being the recipient and source of
identity outcomes. In other words, I examine gender as a status characteristic both
fromthe perspective of a man or womans identity being veried (or not), and from
the standpoint of a man or woman providing feedback that veries (or not) the
identity of another in a situation. By studying the status process as it interacts with
the identity control process, we can see how different theoretical programs (here,
58 JAN E. STETS
ICTand EST) can be related, as shown in other recent work (Cast et al., 1999; Stets,
1997; Stets & Burke, 1996). In general, I am investigating whether the hypotheses
outlined in ICT are conditioned upon ones status in the social structure.
Gender is a diffuse status characteristic. Diffuse status characteristics carry
general expectations for competence as well as cultural associations of specic
skills (Ridgeway, 1993; Ridgeway & Walker, 1995). The former refers more to
the status of gender, that is, the performance expectations associated with being
male or female in the social structure that are imported into a task situation as
in men being directive and the leader in the group and women being deferential
and the follower. The latter refers more to the role of gender in society, that is,
the meaning of being male or female in society that is expressed, for example,
being masculine (dominance and autonomy) or feminine (submissiveness and
afliation) (Stets & Burke, 1996). The above distinction is a subtle yet important
one in understanding how gender as a diffuse status characteristic is hypothesized
to relate to identity verication and emotions as outlined below.
When gender as status becomes activated, as it is in most interactions, it
invokes the belief that men are competent and valuable and women are less
competent and not as valuable (Ridgeway, 1993; Ridgeway &Smith-Lovin, 1999).
Essentially, men are seen as more legitimate occupants of high status positions
than equivalent women (Ridgeway et al., 1994). While women can work at
modifying gender expectations in order to achieve more inuence, the standard by
which they can prove their competence is much stricter than it is for men; they must
work harder than men to show that they are capable (Foddy & Smithson, 1989;
Foschi, 1989).
Recent research reveals that those with high status are more likely than those
with low status to experience identity-verication (Stets & Harrod, 2004). This,
along with earlier research (Cast et al., 1999) that showed that higher status spouses
are more likely than lower status spouses to inuence their partners view of them
(the higher status spouses), indicates that ones position in the social structure
inuences the success by which conrmation about who one is can be achieved.
Further, the greater identity-verication that is accomplished by those higher in
the social structure supports the idea that high status persons are inuential and
competent actors and should be held in high regard, while the difculty in low
status persons achieving identity-verication reafrms their second-class status
in the hierarchy.
If high status persons are more likely than low status persons to enjoy identity-
verication, when identity-verication is not forthcoming in a situation, they may
be more likely to experience negative arousal. Having become accustomed to
achieving identity-verication, perhaps even seeing it as a right that is owed
to them, high status persons may be more reactive than low status person to
Emotions in Identity Theory: The Effect of Status 59
non-verication. The idea here is that a sense of entitlement is operating for
high status males. On the basis of who they are (a valued and rewarded group
in society), men react negatively to any unexpected outcomes; females do not
respond as negatively because they come to expect less (Major, 1987; Steil, 1994).
Indeed, research shows that women not only have a lower sense of entitlement
than men for monetary outcomes such as pay for work (Major, 1987), but they
also have a lower sense of entitlement for non-monetary outcomes such as bearing
a disproportionate amount of responsibility for child care and domestic work
(Steil, 1994).
Given the above, the following two hypotheses are offered for status, emotion,
and identity non-verication in a positive and negative direction:
H1. The lack of identity verication in a positive direction will produce more
negative emotions for men than women.
H2. The lack of identity verication in a negative direction will produce more
negative emotions for men than women.
Since ICT predicts that frequent or persistent interruptions in the identity veri-
cation process are more likely to lead to stronger negative arousal than infrequent
or occasional interruptions, I anticipate that this prediction will hold for high
status persons compared to low status persons. The more that the ongoing identity
process is interrupted, the more that high status persons will experience distress
since it departs from the smooth owing identity feedback process that they have
come to expect. Therefore, the next set of hypotheses include the following:
H3. Persistent lack of identity verication in a positive direction will produce
more negative emotions for men than women.
H4. Persistent lack of identity verication in a negative direction will produce
more negative emotions for men than women.
According to ICT, interruption of the identity verication process is more dis-
tressing when the source of the interruption is a signicant other (operationalized
here as a familiar other). With signicant others, relationships are forged that
are based on mutually veried expectations. When identities are not veried
in these relationships, it can generate distress. EST assumptions regarding the
relationship between peoples characteristics and the status value that these
characteristics imply in terms of shaping a hierarchy of power, prestige, and
inuence in interaction primarily deals with individuals who have no history
of prior interaction. When we begin to relax this constraint, we nd that status
does not always inuence affective responses in the manner predicted from EST
(Stets, 1997; Stets & Burke, 1996). Since most of our daily interactions are
60 JAN E. STETS
with people who we know rather than strangers, this research examines how the
nature of ones relationship with another interacts with the diffuse status char-
acteristic of gender to produce particular affective responses within the identity
control process.
As mentioned above, gender carries cultural associations of specic skills, and
this is more the role aspect of gender as a diffuse status characteristic. With
respect to forging and maintaining relationships, the stereotypical belief for the
genders is that women compared to men are more concerned and skilled at doing
relationships including developing and sustaining relationships as reected in
the communal vs. agentic distinction in the literature (Eagly & Karau, 2002). A
communal orientation involves a concern for others welfare and involves such
behaviors as nurturance, interpersonal sensitivity, helpfulness, and sympathy. An
agentic orientation involves assertion, control, and condence, and it includes
such behaviors as dominance, independence, and forcefulness. Because of the
beliefs about womens greater skills in relationships, I anticipate that women will
respond more negatively than men to identity non-verication from a familiar
other. Thus, my next two hypotheses are:
H5. The lack of identity verication in a positive direction in which the
source is a familiar other will produce more negative emotions for women than
men.
H6. The lack of identity verication in a negative direction in which the
source is a familiar other will produce more negative emotions for women than
men.
Finally, in an extension of the effect of familiarity on emotional arousal, I
investigate the gender of the familiar other. Since, according to EST, those with
high status in the social structure have more authority in an interaction because
they are judged as more competent and worthy than those with low status, women
should respond more negatively when the familiar other is of low status than when
the person is of high status. Identity non-verication from a low status, familiar
other may be seen as illegitimate compared to non-verication from a high status,
familiar other. The nal two hypotheses including the following:
H7. The lack of identity verication in a positive direction in which the source
is a low status, familiar other will produce more negative emotions for women
than men.
H8. The lack of identity verication in a negative direction in which the source
is a low status, familiar other will produce more negative emotions for women
than men.
Emotions in Identity Theory: The Effect of Status 61
METHOD
Subjects and Design
This study takes place in a laboratory that simulates a work situation. Male and
female participants are volunteers recruited from classes at a large northwestern
university. They are paid $10 for their participation in the study. During the
study, they carry out three simple clerical tasks. Before they begin the tasks, they
are given the opportunity to get to know another male or female who will be
involved in the study with them (familiarity condition), or they are not given this
opportunity (unfamiliar condition). Following each task, the participants are given
points/feedback for their work (more points than expected, the expected number
of points, or fewer points than expected) after each task (persistence), or after they
complete all three tasks (non-persistence). After each task or at the end of all three
tasks, participants are asked a series of questions, among them their emotional
reactions to the points/feedback they received following their performance. This
is a 3 2 2 2 2 experimental design (feedback, gender of the participant,
persistence, familiarity, and gender of the manager/confederate). Each condition
has 5 persons and 42 cells have an extra person. A total of 282 participants
are obtained.
Procedure
Two persons (a male or female participant and a male or female confederate) are
ushered into a roomby a supervisor. The participant does not knowthat the other is
a confederate. The supervisor tells themthat an advertising agency, HIGHLIGHTS,
has been asked by one of the major automobile manufacturing companies to run
an ad campaign for a new car that will enter the market in the near future. The ad
agency is seeking feedback to assess their ad campaign strategies. The individuals
will perform three clerical tasks that are the tasks that HIGHLIGHTS employees
will carry out for the ad campaign. Following this, they are administered a brief
background survey.
After lling out the questionnaire, the individuals either get acquainted with
each other for 10 minutes, or they are not given this opportunity.
8
If given the
opportunity to get to know the other, the supervisor leaves the room, and the
10-minute conversation is videotaped. After 10 minutes, the supervisor re-enters
the room and administers a survey to each person regarding their reactions to
the other. Among other things, the survey contains a liking scale (Rubin, 1973)
and a trust scale (Larzelere & Huston, 1980). If participants are not given the
62 JAN E. STETS
opportunity to get to know the other, the survey is administered after they ll out
the background questionnaire. They are asked to ll out the survey based on their
rst impression of the other.
9
Following this, the supervisor randomly assigns the participant to be the
worker for the study and the other (confederate) to be the manager.
10
The
participant is then taken to a second room to watch a ve-minute instructional
video that demonstrates the three clerical tasks that he or she will perform.
11
The
participant in the video receives 100 points following the managers evaluation
that the participant did average work. This sets the identity standard/expectation
that average work means 100 points. The participant is told that while he or she is
watching the video, the supervisor will instruct the other on how to be a manager
for the study in another room. Since the other is the confederate, no instruction
really occurs. After the participant nishes viewing the video, work on the tasks
begin.
The manager begins by reviewing with the participant the three simple clerical
tasks that are the kinds of tasks that HIGHLIGHTS employees will be carrying
out.
12
The manager indicates that after task completion, the manager will give
the participant 100 points for doing average work, 150 points for doing above
average work, and 50 points for doing below average work. To insure that the
participant understands this reward schedule, the participant is asked to ll out
a brief survey following the managers explanation.
13
This procedure sets the
expectation/standard of 100 points for average work for the participant consistent
with what is seen in the video, and it reiterates that the manager will determine
the points the participant gets. The latter is important since a state of injustice
involves perceiving another as responsible for ones outcomes.
In all conditions, the participants receive feedback that they have done average
work. Thus, they should expect 100 points. In the lack of verication in a positive
direction, participants learn that the manager awarded them 150 points for their
work. In the lack of verication in a negative direction, participants receive
feedback that they were awarded 50 points for their work.
Depending upon whether one is assigned to the non-persistent or persistent
condition, the feedback/points is administered at the end of all three tasks or
after each task. Evaluation of ones performance in each condition is based on
weighing the participants work (in ounces) on a scale, and then comparing this
with a bogus evaluation sheet posted on a wall in the room that identies (in
ounces) what constitutes below average, average, and above average work. If
one is assigned to the non-persistent condition, ones emotional reactions to the
managers feedback is obtained at the end of all three tasks, otherwise they are
obtained after each task (persistent condition). Following this, the participants are
debriefed. In total, each person spends about an hour in the lab experiment.
Emotions in Identity Theory: The Effect of Status 63
Manipulation Checks
Following receipt of their reward, participants were asked a series of questions.
To see if the standard of 100 points means average work had been adopted,
participants were asked, After the manager told you how you performed on
these tasks, how many points did you expect the manager to give you? Response
categories included 50 points, 100 points, or 150 points. Since the managers
always told the participants that they did average work, participants should
respond that they expected 100 points. The results showed that the mean response
for this question was 100.89. This was not signicantly different from the mean
value of 100 (t = 0.71, p = 0.48).
To determine whether participants gave the manager responsibility for the
rewards they received, following receipt of their feedback participants were
asked, Who determined how many points you got for these tasks? Almost all
of the participants (96%) said the manager was responsible for the reward(s) they
received.
Measures
Identity non-verication in the worker role is coded 1 if participants receive
feedback of 150 points when they expect 100 points (non-verication in a positive
direction), 0 if participants are given 100 points (identity verication), and 1 if
participants receive 50 points when they expected 100 points (non-verication in
a negative direction). The participants gender is coded 0 for males (high status)
and 1 for females (low status). Persistence is coded 0 for non-persistence and 1 for
persistence. Familiarity is coded 0 for unfamiliar (no opportunity to get to know
the other) and 1 for familiar (opportunity to get to know the other). The managers
gender is coded 0 for males (high status) and 1 for females (low status).
To measure negative emotional reactions, participants are asked to circle a
number that corresponds to the various ways they feel after getting the points for
their task(s). Specically, participants are asked on a scale from 0 (Didnt feel the
emotion at all) to 10 (Intensely felt the emotion) how much they feel fear, guilt,
anger, resentment, disgust, and sadness.
14
Although these emotions are distinctly
different, they have in common an evaluative orientation or hedonic tone that is
negative. Additionally, identity theory does not make predictions about specic
emotions, only about their directionality. Therefore, the items are factor analyzed.
They formed and single underlying dimension. The items are summed with a
high score representing more negative emotions. The omega reliability for this
sale is 0.92.
64 JAN E. STETS
Analysis
To properly test each set of hypotheses presented earlier, I rst divide the sample
into those who identity was disconrmed in a positive direction, and those whose
identity was disconrmed in a negative direction. I then examine how gender
(status) interacts with persistence, familiarity, and the managers gender (status)
to produce the negative emotions that are theoretically expected for each situation.
Ordinary least squares estimates are obtained.
15
RESULTS
Non-Verication in a Positive Direction
Table 1 provides a series of models, each of which separately tests the hypotheses
of identity non-verication in a positive direction on the outcome of negative
emotions. In Model 1, we see that while the effect of status is in the expected
direction with men reacting more negatively than women to non-verication in
a positive direction, the effect does not reach statistical signicance. Therefore,
Hypothesis 1 is not supported. Model 2 adds the effect of persistence and the
interaction effect of gender and persistence to test Hypothesis 3. Surprisingly,
this hypothesis is also not conrmed. A persistent lack of identity verication
in a positive direction does not produce more negative emotions for men than
women. Given that the interaction term is not signicant, this is dropped from all
further equations. Model 3 shows what immediately happens when the interaction
effect is omitted; the effect of persistence emerges. However, rather than persistent
non-verication strengthening reports of negative emotions, it weakens them. In
Model 4, this effect is maintained with the inclusion of familiarity and gender
familiarity to test Hypothesis 5. While the effect is in the expected direction, it is
not signicant, thus Hypothesis 5 is also disconrmed. When the source of the non-
verication is a familiar other, women do not react signicantly more negatively
than men. Consequently, this interaction effect is excluded from all remaining
equations. InModel 5, I ndthat persistence continues tohave a signicant negative
effect for both males and females when the interaction term is omitted.
Model 6 extends Model 5 by investigating whether the signicant effect of per-
sistence that is found in Models 3, 4, and 5, is contingent upon whether the source
of the disconrming feedback is a familiar other. Because there is no hypothesis
for this relationship, it is exploratory. The results in Model 6 reveal a signicant
interaction effect. The effect of persistence is contingent upon whether the source
is a familiar or unfamiliar person. Further, while the main effect of familiarity
E
m
o
t
i
o
n
s
i
n
I
d
e
n
t
i
t
y
T
h
e
o
r
y
:
T
h
e
E
f
f
e
c
t
o
f
S
t
a
t
u
s
6
5
Table 1. OLS Standardized Estimates of Negative Emotions for Identity Non-Verication in a Positive Direction
(N = 94).
Independent Variables Dependent Variables: Negative Emotions
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8
Gender 0.10 0.03 0.10 0.20 0.10 0.10 0.20 0.20
Persistence 0.14 0.20
*
0.21
*
0.20
*
0.01 0.01 0.01
Gender Persistence 0.12
Familiarity 0.04 0.14 0.35
*
0.45
*
0.45
*
Gender Familiarity 0.17
Persistence Familiarity 0.38
*
0.38
*
0.38
*
Mgr. gender 0.02 0.07
Gender Mgr. gender 0.03 0.18
Familiarity Mgr. gender 0.34 0.17
Gender Familiarity Mgr. gender 0.25
R
2
0.01 0.06 0.05 0.08 0.07 0.12
*
0.16
*
0.14
*

p < 0.05.
66 JAN E. STETS
Table 2. Means for Negative Emotions by Persistence and Familiarity.
Familiarity Persistence
No Yes
Unfamiliar 0.40 0.38
Familiar 0.01 0.48
emerges, this effect is continent upon whether the non-verication is persistent.
Table 2 provides more detailed information on the persistence familiarity
interaction effect. An examination of the means shows that participants report less
negative emotions when a familiar other persistently provides feedback that is
more positive than they expect. An unfamiliar other does not produce this effect.
Thus, the ICT prediction that familiarity increases negative emotions only occurs
under the condition of the positive non-verication occurring once compared to
more than once.
Model 7 tests Hypothesis 7 regarding the status of the source of the non-
verifying feedback by including in the equation the effects of gender, familiarity,
and the managers gender as well as their corresponding two-way interactions and
the three-way interaction term. The results in Model 7 do not support Hypothesis
7 since the three-way interaction is not signicant. The persistence familiarity
interaction remains and continues to be signicant when the three-way interaction
is dropped in Model 8.
In general, the results in Table 1 reveal that when persons receive non-verifying
feedback that exceeds their identity standard, the recipients status as well as
the status of the source of the non-verifying feedback does not inuence the
negative feelings that emerge. What is most likely to produce negative emotions
is non-verifying feedback from a familiar other who provides this feedback once
compared to more than once.
Non-Verication in a Negative Direction
Table 3 provides the results for identity feedback that is disconrming in a negative
direction. Model 1 tests Hypothesis 2 and the results support the hypothesis.
The lack of identity verication in a negative direction does produce more negative
emotions for high status males compared to low status females. Model 2 presents
the results for Hypothesis 4 on the relationship between status, persistence,
and non-verication. Like the ndings for identity non-verication in a positive
direction, the status persistence interaction effect is not signicant for identity
E
m
o
t
i
o
n
s
i
n
I
d
e
n
t
i
t
y
T
h
e
o
r
y
:
T
h
e
E
f
f
e
c
t
o
f
S
t
a
t
u
s
6
7
Table 3. OLS Standardized Estimates of Negative Emotions for Identity Non-Verication in a Negative Direction
(N = 94).
Independent Variables Dependent Variables: Negative Emotions
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7
Gender 0.20
*
0.11 0.20
*
0.40
*
0.40
*
0.44
*
0.35
*
Persistence 0.17 0.25
*
0.25
*
0.15 0.26
*
0.26
*
Gender Persistence 0.15
Familiarity 0.31
*
0.20 0.12 0.03
Gender Familiarity 0.33
*
0.33
*
0.47
*
0.32
*
Persistence Familiarity 0.17
Mgr. gender 0.09 0.18
Gender Mgr. gender 0.07 0.08
Familiarity Mgr. gender 0.30 0.44
*
Gender Familiarity Mgr. gender 0.23
R
2
0.04
*
0.11
*
0.10
*
0.16
*
0.17
*
0.25
*
0.24
*

p < 0.05.
68 JAN E. STETS
Table 4. Means for Negative Emotions by Gender and Familiarity.
Gender Familiarity
Unfamiliar Familiar
Male 0.86 0.40
Female 0.25 0.39
disconrmation in a negative direction. Interestingly, when the interaction term is
dropped from the equation as shown in Model 3, the main effects of gender and
persistence appear. Men and those who are disconrmed once (compared to more
than once) are more likely to report negative emotions.
In Model 4, the effects of gender and familiarity are examined. While there
are signicant main effects for gender (men are more negative) and familiarity
(persons are more negative when the source of the non-verication is an unfamiliar
person), it is the interaction effect that is important. Acomparison of the means for
negative emotions by gender when the source is familiar and unfamiliar is shown
in Table 4. The results show that men respond more negatively to identity non-
verication when the disruption is from an unfamiliar other while women react
more negatively to identity disruption froma familiar other. These ndings support
Hypothesis 6.
Like identity non-verication in a positive direction, I explored whether the
signicant effect of persistence on identity non-verication in a negative direction
(Models 3 and 4) might be understood within the context of familiarity. The
results in Table 5 show that the persistence familiarity is not signicant. Thus,
people report more negative emotions when a familiar other disconrms their
identity once (compared to more than once), but only for identity disconrmation
in a positive direction.
Model 6 includes a test of Hypothesis 8, that is, whether women will report
more negative emotions than men for identity verication in a negative direction
in which the source is a low status, familiar other. The three-way interaction of
gender familiarity the managers gender is not signicant, thus disconrming
Table 5. Means for Negative Emotions by Familiarity and Mgr. Gender.
Familiarity Mgr. Gender
Male Female
Unfamiliar 0.46 0.66
Familiar 0.68 0.12
Emotions in Identity Theory: The Effect of Status 69
this hypothesis. When this interaction term is dropped from the equation in
Model 7, the familiarity managers gender interaction termbecomes signicant.
The means of this interaction are presented in Table 5. The results indicate that
participants react more negatively to identity non-verication in a negative direc-
tion when the source is a high status, familiar other. When the source is a lowstatus
other, participants react more negatively when that person is an unfamiliar other.
In comparing the R-square values in Table 1 and Table 3, the models do better
in explaining disconrmation in a negative direction than disconrmation in a
positive direction, however, not always in the manner predicted by ICT. The
effects of status on the identity non-verication process are not as consistent
as anticipated, but when they emerge, they are more likely to predict negative
emotions for identity disconrmation in a negative direction.
DISCUSSION
Because high status people are better able to verify their identities, the current
study examined whether high status persons compared to low status persons
experienced more negative emotions when their identities are not conrmed.
In general, I found that the role of status provides more insights into identity
disconrmation in a negative direction than identity disconrmation in a positive
direction. The two types of disconrmation are somewhat different (Stets, 2003).
The feedback in either direction is non-verifying; however, when it is positive, the
information may enhance the self since it is more favorable than what is expected,
while when it is negative, it does not enhance the self because it is more negative
than what is expected. While the multiple meanings of positive disconrmation are
inconsistent with each other (non-verifying but enhancing), the multiple meanings
of negative disconrmation are consistent with one another (non-verifying
and non-enhancing), and the consistency in the latter meanings may register a
stronger message of disconrmation. Status effects may thus emerge when the
non-verication carries unambiguous meanings.
The effects of the status of the recipient of identity disconrmation as well as
the status of the source of identity disconrmation both appear to be consequential
for emotional reactions. As the recipient of identity disconrmation, high and
low status persons reacted negatively to identity non-verication in a negative
direction based upon how familiar the source was to them. High status (male)
individuals reacted more negatively when the source of disconrming feedback
was an unfamiliar other; low status (female) individuals reacted more negatively
when the source was a familiar other. There is evidence that women are more
likely than men to experience negative feelings such as anger in the context of
70 JAN E. STETS
close relationships (Kring, 2000). In this study, familiarity is only a proxy of more
intimate relationships. While talking for a brief period of time is not the same
as having a history of interaction, it nevertheless had an impact on participants
feelings. For example, participants did report greater attraction (for example,
increased liking) for familiar others than unfamiliar others.
As mentioned earlier, gender as a diffuse status characteristic has a role
aspect to it, that is, it carries cultural associations of certain performances. Given
the distinction between men behaving more agency-oriented and women behaving
more communal-oriented (Eagly & Karau, 2002), high and low status actors
appear to respond most negatively to identity disruptions by those who theyve
grown accustomed to interacting with: for men, it is people to whom they are
not connected, where self-other boundaries are delineated, and where there is
differentiation with others; for women, it is people to whom they are connected
to, where self-other boundaries are not as distinct, and where there is integration
with others. Both men and women may see the disconrming feedback from those
whom they feel most comfortable with as a serious breach of what is expected,
that is, support for the self and identity conrmation.
An unexpected nding was how the source of the non-verifying feedback (in a
negative direction) inuenced participants negative emotions when the sources
status and familiarity were considered. When the source of the disruption was
a high status (male), participants reacted more negatively when the high status
person was a familiar other compared to an unfamiliar other. However, when
the source was a low status (female), participants reacted more negatively when
the low status person was an unfamiliar other compared to a familiar other. How
might these unexpected results be interpreted?
As pointed out many years ago by Brown (1965), we expect relations with high
status persons to be non-intimate and distant while those with low status persons
are more intimate and close. This is reected, for example, in how actors formally
address high status persons compared to the informal address directed at low
status persons. In this study, when a high status person becomes an acquaintance,
or when a low status person is distant, this information by itself may be distressful
because it is unexpected or non-normative. Indeed, as discussed above, high
status persons (men) are less likely than low status persons to forge relationships,
and low status persons (women) are less likely than high status persons to remain
aloof. Given the co-occurrence of these seemingly conictual features, one may
react negatively. If, in addition, persons receive non-verifying feedback in a
negative direction, this will further heighten their negative response. Thus, there
may be multiple bases of negative emotions: the identity disconrmation, and the
conicting messages that the source of the identity disconrmation may send.
ICT has dealt with the former, but we need to include the latter.
Emotions in Identity Theory: The Effect of Status 71
Overall, in bringing status into the analysis, we see how ICT can develop as a
theory. First, the ICT prediction that persons will react more negatively to identity
disruptions from signicant others (here, familiar others) may hold for low status
persons more than high status persons. Compared to high status persons, low
status persons may be more relationship-oriented, and thus they may react more
negatively to identity disruptions from others to whom they have developed a
bond. Further research is needed to see if the pattern that emerged for gender
(that is, women reacting more negatively than men to a familiar other) also
occurs for other status characteristics such as age, race, education, occupation,
and income.
Second, while the sources familiarity may generate a strong affective response,
the sources status may also contribute to this response. If a source is familiar to
one but is also of high status, this is an unexpected occurrence because we expect
relations with high status people to be distant. Alternatively, if the source is a
stranger or unknown to one but is also of low status, this also is an unanticipated
event since we expect relations with low status people to be familiar and friendly.
These inconsistent, albeit conicting messages about the source may inuence
negative arousal. At the same time, the recipient is experiencing non-verifying
feedback from the source which is also producing negative arousal. Thus, the
source may send multiple messages to the recipient: (1) non-verifying identity
feedback; and (2) the message that he or she is behaving in a fashion that is
discordant with his or her status. Since each, on their own, can produce negative
arousal, when they are sent simultaneously in a situation, they may build upon one
another thereby strengthening ones negative response in the situation. Therefore,
in understanding affective responses in situations, identity control theorists need
to investigate the meanings that lie behind identity non-verication as well as the
meanings that senders of identity non-verication signal by their behavior.
NOTES
1. As Stryker and Burke (2000) point out, identity theory has evolved into two different
but related directions. The rst, represented in the work of Stryker and his colleagues
(Stryker, 1980; Stryker & Serpe, 1982), focuses on how social structures affect the self,
and in turn, social behavior. The second, taken up in the work of Burke and his associates
(Burke, 1991) and labeled Identity Control Theory (ICT), addresses the internal operations
of the self as they inuence social behavior.
2. The importance of the congruency notion is shared not only by sociological social
psychologists, but also psychological social psychologists as revealed in Swanns (1990)
self-verication theory and Higgins (1989) self-discrepancy theory.
3. This is consistent with the work of Cooley ([1909] 1962) and Shott (1979) who
view negative emotion as emerging when others do not accept ones self-image that has
72 JAN E. STETS
been built up, and positive emotion as emerging when normative, moral conduct receives
approval from others.
4. This is similar to McCall and Simmons (1978) idea that negative emotions are more
likely to emerge when more prominent (important) identities are threatened.
5. Burke borrows this idea from Mandlers (1982) interruption theory. According to
Mandler, distress is felt when organized action is interrupted. Looked at another way, an
expectancy has not been conrmed. The distress that is instigated by the interruption signals
that something is wrong. An actor responds by attempting to adapt to the interruption.
The more repeated the interruption, the more the person is unable to initiate and sustain
organized activity, and the more distress there will be.
6. In an indirect test of this, Cast and Burke (2002) nd that the persistent lack of
self-verication over time leads to an increasing loss of self-esteem, a self-feeling.
7. In this study, there was not sufcient variation on other status characteristics such as
race, education, or occupation to include them in the analysis.
8. In the get acquainted condition, a sheet of 12 questions is provided as a guide to get
to know each other. Questions include such things as What is your favorite activity?,
What qualities do you think are most important for a good friendship?, and Describe
an experience from your life that made you feel happy; what was so special about
that experience?
9. Compared to the unfamiliar condition, participants in the familiar condition reported
more liking for the other (t = 8.96, p < 0.01), saw the other as a potential friend
(t = 4.65, p < 0.01), as a person to invite to a party (t = 5.39, p < 0.01), as someone
they would be happy to introduce to their friends (t = 6.06, p < 0.05), and a person they
would be happy to work with on a job (t = 5.70, p < 0.01). Participants in the familiar
condition were not more trusting of the other than in the unfamiliar condition. This is not
surprising since trust takes time to develop in a relationship.
10. Participants choose a number between 1 and 10. The confederate always chooses
rst and chooses the number 3. After the participant chooses a number, the supervisor
tells the two that the folded papers in the box in front of them contains numbers that range
from 1 to 10. In reality, all folded papers contain the number 7. Based on whatever number
the participant chooses, the supervisor then does a quick calculation in his head. The
supervisor explains that whoever chose furthest from the number (if the participant chose
1 or 2) or closest to the number (if the participant chose 4 thru 10) that is picked in the box
will be the worker and the other will be the manager. Thus, the participant is always the
worker and the confederate is always the manager.
11. Eight different instructional videos are made to allow for the male or female
participant in the video to carry out the tasks under the condition of non-persistence with
a male manager, non-persistence with a female manager, persistence with a male manager,
and persistence with a female manager. Depending upon what condition the participant is
randomly assigned to ahead of time determines the video s/he will view. For example, if
the participant is male and is randomly assigned to the persistence condition with a male
confederate, then he will viewthe video of another male receiving persistent rewards froma
male manager.
12. The three tasks are simple and include: (1) alphabetizing promotional letters on
the new car by the selected familys last name (listed at the top of the promotional letter);
(2) writing each familys name and address from the top of the promotional letter onto
to a HIGHLIGHTS mailing envelope and clipping a return envelope that is addressed
to HIGHLIGHTS to the promotional letter and mailing envelope; and (3) taking a new
Emotions in Identity Theory: The Effect of Status 73
stack of promotional letters, doing task two again, but keeping each promotional packet
in alphabetical order as participants go along. Each task is carried out for six minutes and
participants are videotaped throughout their participation. Extensive pre-testing revealed
that six minutes is long enough for participants to develop an attitude toward the tasks.
13. If the manager sees that the participant is answering incorrectly on the survey, the
manager reviews the reward schedule with the participant. Additionally, if the participant
does not identify the manager as the person who will determine how many points she or he
gets after task completion, the manager reviews the administration of points, emphasizing
that it is the manager who will determine how many points the participant gets.
14. Previous researchers have measured emotions using this single-item scale (Driskell
& Webster, 1997; Hegtvedt, 1990; Lovaglia, 1997; Lovaglia & Houser, 1996; Ridgeway
& Diekema, 1989; Shelly, 2001; Sprecher, 1986).
15. An alternative analysis of estimating the emotions associated with each of identity
conditions is to use the entire sample and include the identity conditions as dummy
variables. However, this is inappropriate because participants are included in the analysis
for which there are no theoretical predictions as to how they would feel given the emotion
that is being examined. For example, in estimating the effects of persistence and familiarity
on negative emotions for those whose identity is disconrmed in a negative direction, the
analysis would include individuals whose identities are conrmed. There are no hypotheses
for negative emotions when individuals identities are conrmed. Further, there are no
predictions as to how persistence and familiarity would be related to negative emotions
when persons identities are conrmed. Thus, the inclusion of these individuals in the
negative emotions equation would be contaminating the results.
ACKNOWLEDGMENTS
This research was supported by National Science Foundation Grant SES 9904215.
REFERENCES
Bartels, D. J. (1997). An examination of the primary emotions of anger and sadness in marriage within
the context of identity theory. Unpublished MA. Pullman, WA: Washington State University.
Berscheid, E., & Reis, H. T. (1998). Attraction and close relationships. In: D. T. Gilbert, S. T. Fiske
& G. Lindzey (Eds), Handbook of Social Psychology (pp. 193281). Boston: McGraw-Hill.
Brown, R. (1965). Social psychology. New York: Basic Books.
Burke, P. J. (1991). Identity processes and social stress. American Sociological Review, 56, 836849.
Burke, P. J. (1996). Social identities and psychosocial stress. In: H. B. Kaplan (Ed.), Psychosocial
Stress: Perspectives on Structure, Theory, Life Course, and Methods (pp. 141174). Orlando,
FL: Academic Press.
Burke, P. J., & Stets, J. E. (1999). Trust and commitment through self-verication. Social Psychology
Quarterly, 62, 347366.
Cast, A. D., & Burke, P. J. (2002). A theory of self-esteem. Social Forces, 80, 10411068.
Cast, A. D., Stets, J. E., & Burke, P. J. (1999). Does the self conform to the views of others? Social
Psychology Quarterly, 62, 6882.
74 JAN E. STETS
Collins, R. (1990). Stratication, emotional energy, and the transient emotions. In: T. D. Kemper (Ed.),
Research Agendas in the Sociology of Emotions (pp. 2757). New York: State University of
New York Press.
Cooley, C. H. ([1909] 1962). Social organization. New York: Scribner.
Driskell, J. E., &Webster, M. J. (1997). Status and sentiment in task groups. In: J. Szmatka, J. Skvoretz
& J. Berger (Eds), Status, Network, and Structure: Theory Development in Group Processes
(pp. 179200). Stanford: Stanford University Press.
Eagly, A. H., & Karau, S. J. (2002). Role congruity theory of prejudice toward female leaders.
Psychological Review, 109, 573598.
Foddy, M., & Smithson, M. (1989). Fuzzy sets and double standards: Modeling the process of
ability inference. In: J. Berger, J. M. Zelditch & B. Anderson (Eds), Sociological Theories in
Progress: New Formulations (pp. 7399). Sage.
Foschi, M. (1989). Status characteristics, standards, and attributions. In: J. Berger, J. M. Zelditch & B.
Anderson (Eds), Sociological Theories in Progress: New Formulations (pp. 5872). Newbury
Park, CA: Sage.
Hegtvedt, K. A. (1990). The effects of relationship structure on emotional responses to inequity.
Social Psychology Quarterly, 53, 214228.
Hegtvedt, K. A., & Markovsky, B. (1995). Justice and injustice. In: K. S. Cook, G. A. Fine & J. S.
House (Eds), Sociological Perspectives on Social Psychology (pp. 257280). Boston: Allyn
& Bacon.
Heise, D. R. (1979). Understanding events: Affect and the construction of social action. Cambridge:
Cambridge University Press.
Higgins, E. T. (1989). Self-discrepancy theory: What patterns of self-beliefs cause people to suffer?
In: L. Berkowitz (Ed.), Advances in Experimental Social Psychology (Vol. 22, pp. 93136).
New York: Academic Press.
Kemper, T. D. (1990). Social relations and emotions: A structural approach. In: T. D. Kemper (Ed.),
Research Agendas in the Sociology of Emotions (pp. 207237). Albany: State University of
New York Press.
Kring, A. M. (2000). Gender and anger. In: A. H. Fischer (Ed.), Gender and Emotion: Social
Psychological Perspectives (pp. 211231). New York: Cambridge University Press.
Larzelere, R. E., & Huston, T. L. (1980). The dyadic trust scale. Journal of Marriage and the Family,
42, 595604.
Lawler, E. J., & Yoon, J. (1996). Commitment in exchange relations: Test of a theory of relational
cohesion. American Sociological Review, 61, 89108.
Lovaglia, M. J. (1997). Status, emotion, and structural power. In: J. Szmatka, J. Skvoretz & J. Berger
(Eds), Status, Network, and Structure (pp. 159178). Stanford: Stanford University Press.
Lovaglia, M. J., & Houser, J. A. (1996). Emotional reactions and status in groups. American
Sociological Review, 61, 867883.
Major, B. (1987). Gender, justice, and the psychology of entitlement. In: P. Shaver & C. Hendrick
(Eds), Sex and Gender (pp. 124148). Newbury Park: Sage.
Mandler, G. (1982). Stress and thought processes. In: L. Goldberger & S. Breznitz (Eds), Handbook
of Stress: Theoretical and Clinical Aspects (pp. 88104). New York: Free Press.
McCall, G. J., & Simmons, J. L. (1978). Identities and interactions. New York: Free Press.
Ridgeway, C. L. (1993). Gender, status, and the social psychology of expectations. In: P. England
(Ed.), Theory on Gender/Feminism on Theory (pp. 175198). New York: Aldine.
Ridgeway, C. L., & Diekema, D. (1989). Dominance and collective hierarchy formation in male and
female task groups. American Sociological Review, 54, 7993.
Emotions in Identity Theory: The Effect of Status 75
Ridgeway, C. L., & Johnson, C. (1990). What is the relationship between socioemotional behavior
and status in task groups? American Journal of Sociology, 95, 11891212.
Ridgeway, C. L., Johnson, C., & Diekema, D. (1994). External status, legitimacy, and compliance in
male and female groups. Social Forces, 72, 10511077.
Ridgeway, C. L., & Smith-Lovin, L. (1999). Gender and interaction. In: J. S. Chafetz (Ed.), Handbook
of the Sociology of Gender (pp. 247274). New York: Kluwer Academic/Plenum.
Ridgeway, C. L., & Walker, H. A. (1995). Status structures. In: K. S. Cook, G. A. Fine & J. S. House
(Eds), Sociological Perspectives on Social Psychology. Boston: Allyn & Bacon.
Rubin, Z. (1973). Liking and loving: An invitation to social psychology. New York: Holt Rinehart and
Winston.
Shelly, R. K. (2001). How performance expectations arise from sentiments. Social Psychology
Quarterly, 64, 7287.
Shott, S. (1979). Emotion and social life: A symbolic interactionist analysis. American Journal of
Sociology, 84, 13171334.
Smith-Lovin, L., & Heise, D. R. (1988). Analyzing social interaction: Advances in affect control
theory. New York: Gordon and Breach Science Publishers.
Sprecher, S. (1986). The relation between inequity and emotions in close relationships. Social
Psychology Quarterly, 49, 309321.
Steil, J. M. (1994). Equality and entitlement in marriage: Benets and barriers. In: M. J. Lerner &
G. Mikula (Eds), Entitlement and the Affectional Bond (pp. 229258). New York: Plenum
Press.
Stets, J. E. (1997). Status and identity in marital interaction. Social Psychology Quarterly, 60,
185217.
Stets, J. E. (2003). Justice, emotion, and identity theory. In: P. J. Burke, T. J. Owens, R. T. Serpe &
P. A. Thoits (Eds), Advances in Identity Theory and Research (pp. 105122). New York:
Kluwer Academic/Plenum.
Stets, J. E. (2005). Examining emotions in identity theory. Social Psychology Quarterly, 68.
Stets, J. E., & Burke, P. J. (1996). Gender, control, and interaction. Social Psychology Quarterly, 59,
193220.
Stets, J. E., & Burke, P. J. (2000). Identity theory and social identity theory. Social Psychology
Quarterly, 63, 224237.
Stets, J. E., & Harrod, M. M. (2004). Verication across multiple identities: The role of status. Social
Psychology Quarterly, 67, 155171.
Stets, J. E., & Tsushima, T. (2001). Negative emotion and coping responses within identity control
theory. Social Psychology Quarterly, 64, 283295.
Stryker, S. (1980). Symbolic interactionism: A social structural version. Menlo Park, CA: Benjamin
Cummings.
Stryker, S. (1987). The interplay of affect and identity: Exploring the relationships of social structure,
social interaction, self, and emotion. Chicago: American Sociological Association.
Stryker, S., & Burke, P. J. (2000). The past, present, and future of an identity theory. Social Psychology
Quarterly, 63, 284297.
Stryker, S., & Serpe, R. T. (1982). Commitment, identity salience, and role behavior: A theory and
research example. In: W. Ickes & E. S. Knowles (Eds), Personality, Roles, and Social Behavior
(pp. 199218). New York: Springer-Verlag.
Swann, W. B., Jr. (1990). To be adored or to be known?: The interplay of self-enhancement and
self-verication. In: E. T. Higgins & R. M. Sorrentino (Eds), Handbook of Motivation and
Cognition (pp. 408450). New York: Guilford.
76 JAN E. STETS
Turner, J. H. (1999). Toward a general sociological theory of emotions. Journal for the Theory of
Social Behavior, 29, 132162.
Wagner, D. G., & Berger, J. (1993). Status characteristics theory: The growth of a program. In:
J. Berger, M. Zelditch, Jr., et al. (Eds), Theoretical Research Programs: Studies in the Growth
of Theory (pp. 2363). Stanford: Stanford University Press.
Zajonc, R. B. (2000). Feeling and thinking: Closing the debate over the independence of affect. In:
J. P. Forgas (Ed.), Feeling and Thinking: The Role of Affect in Social Cognition (pp. 3158).
New York: Cambridge University Press.
PHYSIOLOGICAL MEASURES OF
THEORETICAL CONCEPTS: SOME
IDEAS FOR LINKING DEFLECTION
AND EMOTION TO PHYSICAL
RESPONSES DURING INTERACTION
Dawn T. Robinson, Christabel L. Rogalin
and Lynn Smith-Lovin
ABSTRACT
After a vigorous debate in the late 1970s, the sociology of emotion put aside
most discussion of whether or not the physiological arousal associated with
emotion labels is differentiated. Since this early period, scholars have made
great progress on two fronts. First, theories about the interrelationship
of identity, action and emotion have specied a family of new concepts
related to emotion. Second, a large corpus of research on the physiological
correlates of emotional experience emerged. In this chapter, we review
the well-developed control theories of identity and emotion, and focus
on the key concepts that might relate to different physiological states.
We then review the general classes of physiological measures, discussing
their reliability, intrusiveness and other features that might determine their
usefulness for tracking responses to social interaction. We then offer a
highly provisional mapping of physiological measures onto the concepts
Theory and Research on Human Emotions
Advances in Group Processes, Volume 21, 77115
Copyright 2004 by Elsevier Ltd.
All rights of reproduction in any form reserved
ISSN: 0882-6145/doi:10.1016/S0882-6145(04)21004-9
77
78 DAWN T. ROBINSON ET AL.
that they might potentially measure, given past research about how these
physiological processes relate to environmental stimuli. While any linkage
between concepts and measures must be speculative at this point, we
hope that this review will serve as a stimulus to theoretically guided research
that begins to assess the validity of these new measures for sociological use.
INTRODUCTION
There are as many denitions of emotion as there are theoretical views of its
origins and consequences. This apparent disarray is not as troubling as it might
sound. Instead, it represents the fact that such concepts are only scientically
useful in the context of a theoretical statement that embeds themin a larger process.
Nonetheless, reviews of the sociology of emotion both early (Thoits, 1989) and late
(Turner, 2000) note that virtually all denitions of emotion contain two elements
some element of physiological arousal and an interpretation of that arousal using
contextual cues and cultural knowledge.
Much of the early debate in the subeld centered on which of these two
elements was dominant. Those called positivists (e.g. Kemper, 1978) argued
that social interaction led to predictable, distinctive, differentiated physiological
states; cultural labels were then attached to the multiplicity of social situations and
mixes of physiological arousal that were meaningful (and frequently encountered)
within a given culture (Kemper, 1987). Social constructionists (e.g. Gordon, 1990;
Hochschild, 1983) gave more weight to the socio-cultural nexus that gives us
the words, behavioral cues, and narrative scripts to dene emotions. While these
scholars ranged from those who implied a real emotion that is then managed
by actors to comply with social norms (Hochschild, 1983, 1990) to those with the
more radical position that even the autonomic responses entailed by an emotion
are specied by our cultural understanding of it (Gordon, 1990), they shared a
distaste for any serious reliance on the biological substrate for our understanding
of the important social dynamics that produced and were motivated by emotion.
Turner (2000) reviewed 20 authors (ranging from Darwin in 1872 to Turner
himself in 1996) who posited between 3 and 10 primary emotions that they be-
lieved to be primary, basic and universal across cultural settings. While the search
for a set of primary emotions centered in psychology and evolutionary biology,
most sociologists and anthropologists interested in understanding emotion were
adamant about emotions social and cultural rather than physiological roots.
Kemper (1978, 1990) and Turner (1996, 1999a, b, 2000) were notable exceptions
to this general position among social scientists. This often rancorous debate
1
gave
way to a more productive period of synthesis and development in the late 1980s
and 1990s. Marked roughly by Kempers (1987) article in the American Journal
Physiology of Deection and Emotion 79
of Sociology, sociologists reached a point of relative consensus that the social
process that led to emotional experiences was a more important focus than the
issue of whether there were one or four or ten primary, universal emotions that
had a physiologically distinct prole. Theory development proceeded with little
attention to the biological substrate with which social processes were intertwined.
Not surprisingly, research on the physiology of emotion did not remain
dormant during this period. A wealth of new evidence has been generated by
new measurement technologies. While much of the attention in recent years has
been focused on the cognitive neuroscience of emotion (for example, Lane &
Nadel, 2000), several substantial research programs developed a much fuller
understanding of the autonomic nervous system and related endocrine physiology.
Since these developments have proceeded in parallel, with relatively little
attention by sociologists (Turner, 2000, being a notable exception) to the progress
of physiological research, we suggest that it is time to reassess whether or not
physiological analogues to emotional experience might be useful for research
on sociological theories of emotion. In this chapter we will review briey the
general types of physiological responses that might be related to socio-emotional
experience. We will concentrate in particular on those classes of physiological
response that appear to have some usefulness in monitoring responses to
ongoing social interaction. Because of this constraint, we will largely ignore
the voluminous new literature on direct measurement of central nervous system
activity through positron emission tomography (PET) and functional magnetic
resonance imaging (fMRI) technologies, although there is evidence that these
patterns may be reliably linked to specic emotional experience (LeDoux, 1996).
While technology is developing that would allow real time assessments of brain
activity during social interaction, that technology is unlikely to become available
to social scientists in the near future. Instead, we will concentrate on measures
that could be implemented in current laboratory settings. We try to concentrate
on measures that have some demonstrated reliability across gender, culture and
other major social divides, noting limitations of this sort when they are known.
Arguably, the rst step in discovering whether or not a measurement technology
is useful for ones research is a pre-existing conceptual understanding of what
one is trying to measure. Therefore, before we delve into the often contradictory
research on physiological responses and emotional experience, we briey review
the sociological theories that might most productively employ such measurement
technology. Since such measures are most feasible inside the laboratory, we
will concentrate our summary of concepts in need of measurement within the
structural symbolic interactionist theoretical tradition. These researchers have
used experimental studies to test and elaborate their theories, and have potential
for making use of physiological measurement. They also offer a wide array of
80 DAWN T. ROBINSON ET AL.
affective concepts within the context of fairly well-dened theoretical structures,
so that we might expect the interplay between theoretical prediction and mea-
surement over the course of a series of studies that would be necessary to rmly
establish the validity of any physiological measure for a theoretical concept.
CONCEPTS IN THE STRUCTURAL SYMBOLIC
INTERACTIONIST THEORIES OF EMOTION
The two major theories of emotion in the more quantitative, structural branch
of symbolic interaction are affect control theory (Heise, 1979; MacKinnon, 1994;
Smith-Lovin &Heise, 1988) and identity control theory (Burke, 1991, 1996; Burke
& Reitzes, 1991). As variants of symbolic interactionism, both emphasize the
central role of identity meanings and the ability of social interactions to sustain
them in producing emotional responses. Both theories share a common basis in
William T. Powers (1973) cybernetic control model of perception. We therefore
begin with the concepts that they share in common, before proceeding to the
somewhat differing predictions that they make about emotional response.
As control systems, both theories have as their central mechanism the com-
parison between a reference signal (which is constant for all purposes within the
context of a given social interaction) and input fromthe situational context in which
the actor is embedded. The reference signal in both theories is a set of meanings. In
affect control theory, these meanings are fundamental sentiments that are acquired
through past interactions, exposure to cultural materials and other sources. The
meanings have both cognitive and affective components, which are inseparably
evoked when self, other interactants and actions are labeled in the denition of a
situation (MacKinnon, 1994). The affective meanings fall on three dimensions that,
based on the work of Charles Osgood and his colleagues (Osgood, May & Miron,
1975; Osgood, Suci &Tannenbaum, 1957), are presumed to be universal and appli-
cable to all concepts that actors use to dene the situation. The three dimensions are
evaluation (goodness vs. badness), potency (powerfulness vs. powerlessness), and
activation (lively vs. quiet). Identity control theory presumes a somewhat less cul-
tural, more personalized set of self-identity meanings that are arranged in a salience
hierarchy within a stable self structure (Burke, 1991; Stryker, 1980). The meanings
can vary in the number and type of dimensions that are relevant, making the assess-
ment of the reference signal a more individualized, domain-specic enterprise.
In both theories, actors operate to maintain the meanings that act as their
reference levels within a situation. Perceived events within the situation are
compared to the reference meanings to assess the current state of the system with
regard to meaning maintenance. In affect control theory, the actor processes an
Physiology of Deection and Emotion 81
event of the general form the actor does behavior to the object person within this
setting to produce a set of situated meanings for the actor, behavior, object-person
and setting, which are then compared to the fundamental sentiments associated
with those event elements. This comparison process results in an assessment of
deection, which is mathematically dened as the sum of the squared differences
between the situated meanings and the fundamental sentiments for each event
element on each dimension (Heise, 1979). An equivalent way of describing
deection is the distance in the multiple-dimensional evaluation potency
activity space that the event elements have been disturbed by the perceived event.
In identity control theory, the reference level is an identity standard (a stored set
of self-meanings) that serves as a guide for behavioral outputs by the actor. The
perceptual input is how people see themselves in the situation, a result of reected
appraisals by others and of self-perception within the setting. A comparator
compares the perceptual input with the identity standard and assesses the amount
of discrepancy between the two. While identity control theory is not formulated
in mathematical terms, the discrepancy is clearly a difference between the two
theoretical elements the identity standard and the perceptions from the situation.
While they differ in the location of the reference signal (all event elements
vs. self-identity), the dimensions on which they are compared (three stable
dimensions vs. more individuated, domain specic structure), and the form of
the comparison (mathematically specied vs. a general sense of difference), both
theories share the same central control principle. They assume that people operate
so as to conrm the meanings that they hold for the situation. Under normal
circumstances, people act so as to conrm their identity meanings given the social
resources (e.g. network alters and behavioral options) that the situation offers.
When events fail to support meanings, new actions (or, if action is not possible,
cognitive relabelings) will occur to restore those meanings.
Burke (1991, 1996) has been most explicit about the psychological, and
presumably physiological, experience of discrepancy between self-meaning
reference levels and situational perceptions. He relies heavily on Mandler (1982)
in predicting that autonomic activity (including distress and anxiety) will result
whenever some organized act or thought process is interrupted. Such interruptions
may occur when demands are made on individuals that tax or exceed the resources
that those individuals have for managing them, or they may be the result of social
factors outside of the individual (e.g. other actors with competing ideas about
the meaning of the individuals role-identities). Interruption happens whenever
expectancies are disconrmed or when lines of action are prevented from comple-
tion. Since, under identity control theory, people expect their identity meanings
to be conrmed by those around them and engage in lines of action consistent
with those meanings, clearly discrepancy constitutes an interruption and leads to
82 DAWN T. ROBINSON ET AL.
stress. Among other things, this response leads the actor to re-orient attention to
the identity that is being threatened.
In spite of its precise mathematical denition, affect control theory has been
less explicit about the direct experience of deection. Perhaps because it has a
more elaborated model of emotion (see Averett &Heise, 1987; MacKinnon, 1994;
Smith-Lovin, 1990; and discussion below), most affect control scholars have
regarded deection as a theoretical construct without a direct experiential com-
ponent. However, it is clear from the discussions of deection in the context of
the theoretical model and from the few studies that have directly linked deection
to perceived likelihood (e.g. Heise & MacKinnon, 1987) that this disconrmation
of the affective meanings implied by ones denition of a situation would lead
to a sense of disequilibrium, surprise (or even shock in its extreme instances),
and stress. Autonomic arousal should result in the sense that the control system
explicitly predicts some response to this state of disconrmation. The response
can be either: (a) action to create new events that, when processed, restore situated
meanings to their fundamental prole; or (b) a rethinking of events such that
the new labels do not generate so much meaning disturbance. But in either case,
action by the individual confronted by deection is required.
While the twocontrol theories agree that some type of disequilibriumandarousal
(which we might label stress) will be produced when deection/discrepancy is
high, they have rather different positions on the specic emotions that will be ex-
perienced. As noted above, Burke (1991, 1996) is clear that the valence of emotion
produced by discrepancy is negative. He follows Stryker (1987) in suggesting that
failure to meet normative expectations for performance in salient role-identities
will result in negative emotion. Burke labels the discrepancy-produced emo-
tions as anxiety and distress. While psychologists might see these two emotions
as relatively distinct (because distress is backward looking and anxiety refers
to worry about future events), they share the character of being negative and
activating. Indeed, their motivational role leading to attention and action within the
theory argues explicitly for some type of autonomic arousal as well as displeasure.
Similarly, Stets (2003) combines identity control theory and justice/equity theory
to hypothesize that over-reward (reected appraisals that are higher than those
expected for conrmation of identity meanings) will lead to guilt and fear. When
she nds that over-reward produces more positive feelings, she modies the
theoretical ideas somewhat (Stets, 2003, pp. 118119). She suggests that in cases
of moderate upward discrepancy, the self-enhancing impact of the over-reward
might weaken the effect of the non-verication. In negative discrepancies, we
effectively have two negatives: the non-verication and the punishment or under
reward that constitutes negative treatment from the interaction partner. In cases of
positive discrepancy, the non-verication of identity is counterbalanced somewhat
Physiology of Deection and Emotion 83
from the positive feedback of the treatment of others the fact that one is being
treated well, and the just world belief that good things come to good people.
Therefore, she argues it is easier to adjust the self-identity meanings upward in a
self-enhancing way; the fact that she nds that repeated over reward leads to lower
levels of emotion, while repeated under reward leads to higher levels of emotion
is a primary impetus for this theoretical revision.
While the effects of positive discrepancy may be in theoretical ux, Stryker
(1987), Burke (1991, 1996) and Stets (2003) all agree that conrming identity
meanings by producing role performances that are expected and approved by
others will result in positive emotion. When reected appraisals from the envi-
ronment produce no discrepancy, people are conrming their identity meanings
and no restorative action is required. These positive effects will strongest when
the identities are salient in the self hierarchy and when interactions are with alters
who are important and intimate in the individuals social network.
Affect control theorys predictions are quite a bit more complex, because of its
mathematical model (Averett & Heise, 1987; Heise & Calhan, 1995; MacKinnon,
1994; Robinson & Smith-Lovin, 1992, 1999; Smith-Lovin, 1990). Emotions
accompany deections and signal to an actor how his or her experience did
or did not conrm expectations based on the current denition of the situation
(Smith-Lovin, 1990; Smith-Lovin &Heise, 1988). The type of emotional response
is determined by both the size and direction of the deection as well as by the
transient impression produced by the situation, all of which are calculated on
the three underlying dimensions of affective meaning evaluation, potency
and activity. In other words, the two components of emotions in the model
are: (a) the situated meanings brought about by the interaction; and (b) the
distance and direction of movement between those meanings and the fundamental
meanings culturally associated with the role-identity of the focal actor. Thus,
when deections are small (i.e. when events are conrming), an actors emotional
response is largely determined by his or her identity (i.e. the fundamental
evaluation, potency and activity meanings of that cultural symbol). However,
when events are disconrming, the nature of this deection heavily determines the
character of the emotional response. Considering only the evaluation component
of emotion, the equation predicting the emotional response in affect control theory
can be reduced to:
E = 2T I or T +(T I)
where E is the evaluation of the emotion, T is the transient evaluation produced
by the social interaction, and I is the evaluation of the fundamental self-identity
(Averett & Heise, 1987). Thus, we see that emotional response will be inuenced
both by the impressions created by the current situation (T) and by the deection
84 DAWN T. ROBINSON ET AL.
of ones situated identity meanings away from fundamental self-identity (T I)
(Averett & Heise, 1987; Smith-Lovin, 1990). When social events fail to conrm
fundamental identity, the nature of ones transient feelings in the current setting
(T) is very important in predicting the evaluative content of ones emotional
response. If we take a negative transient impression and subtract from it a positive
self-identity (I), a quite negative emotion will be the result. If a person with a nega-
tive self-identity is deected to a temporarily positive self-image, the emotion will
be positive (T will be high, and the subtraction of a negative I will add to this pos-
itivity). The emotion is positive both because of the pleasant transient impression,
and because it is so far above the normally negative fundamental self-image.
On the other hand, when social events conrm the sentiments of ones
fundamental identity, or
T = I
Then,
E = 2T I = 2I I = I
In other words, when one is not deected, the valence of ones emotional
experiences mirror the valence of ones fundamental identity. The empirically
estimated equations in the mathematical model differ somewhat for the other two
dimensions (see Averett & Heise, 1987, for a full discussion) but the basic form
of the emotions predictions is similar. Rather than assuming that discrepancies
have the same emotional impact on everyone, affect control theory stresses the
importance of identity in determining emotion. And, rather than assuming that
discrepancy only produces positively and negatively valenced emotions (either
satisfaction with current role performance or distress/anxiety about failure to
maintain identity standards), the mathematical model underlying affect control
theory predicts specic three-dimensional proles (on the evaluation, potency and
activity dimensions) for the emotions that will be produced by a given identity
embedded in an event dened by another actors role-identity, an action that one
of them takes, and perhaps a socially dened setting in which the action occurs.
RELATED THEORIES IN PSYCHOLOGY
Despite our focus on sociological theories relating identity, action and emotion,
we will mention two closely related psychological theories that suggest similar
emotion productions. These psychological traditions might also benet from the
physiological measures of the emotion that we attempt to map conceptually in the
next sections, since they share some close conceptual links with the sociological
Physiology of Deection and Emotion 85
approaches. The two theories with the most obvious connections are Carvers and
Scheiers (1981, 1990, 2000) self-regulation theory and Higgins (1987, 1989)
self-discrepancy theory.
Carver and Scheier, like Heise and Burke, drewexplicitly on WilliamT. Powers
(1973) perceptual control model in developing their ideas about how intentions,
actions and emotions were related. Their theory is focused on how people ac-
complish goals, and has much in common with an earlier treatment of cybernetic
systems in Miller, Galanter and Pribam (1960). The core idea is that actors
have hierarchically organized sets of goals. Goals at one level imply behavioral
programs at lower levels to accomplish those goals (if one has the goal of winning
a Nobel Prize, one would implement a lower-level goal of earning a Ph.D.,
which would also imply lots of lower routines about going to class, meeting with
professors, conducting research, etc.). The system is a control model because
the goals act as a reference level for monitoring the environment and assessing
whether or not progress is being made toward the goal. Actors act to minimize
discrepancies, which mean to accomplish goals.
In addition to the negative (discrepancy reducing) feedback system, Carver
and Scheier also write about positive (discrepancy enlarging) feedback systems
with reference levels that the actor is motivated to move away from.
2
They
relate these two types of systems to physiological research differentiating the
approach/engagement behavioral activation system and the avoidance/withdrawal
behavioral inhibition system (Carver & Scheier, 2000, pp. 4647). There is
evidence that these systems are independent and located in different neurological
domains. Their predictions about emotional responses are in some ways highly par-
allel to the affect control theory model. In particular, there is a distinction between
the general sense of expectancy (discrepancy or deection), and the emotions
produced: they
. . . suggest that the result of the comparison process at the heart of this loop (the error signal
generated by the comparator) is manifest phenomenologically in two forms: one is a hazy and
nonverbal sense of expectancy condence or doubt; the other is affect, feeling a sense of
positiveness or negativeness (Carver & Scheier, 2000, p. 51).
There are some important differences, however. In self-regulation theory, actors
experience emotion not as a direct result of discrepancy between goal and
environment, but as a result of the rate at which they are progressing toward the
goal (or, in the case of positive feedback loops, avoiding the undesired outcome).
In effect, they posit a second-order loop that measures whether or not progress
in reducing or enlarging the discrepancy is adequate. The rate of discrepancy
change becomes the reference level against which the current state is judged. If the
rate of discrepancy change is lower than anticipated, people experience negative
86 DAWN T. ROBINSON ET AL.
emotions; if the rate of discrepancy change is as rapid (or more rapid) than
expected, positive emotions are experienced. Emotions are somewhat different in
character in discrepancy reducing (negative feedback) and discrepancy enlarging
(positive feedback) loops. Fast progress in discrepancy reduction (moving toward
a goal) produces elation, while inadequate progress produces depression. Rapid
progress in a discrepancy enlarging an avoidance loop results in relief, while poor
progress produces anxiety. In each case, moving in the undesirable direction leads
to a negative evaluation emotion, but the potency and activity of the emotion
vary for slow progress toward goals, potency and activity are low (depression
is low evaluation, low potency and low activity), while for slow progress away
from disliked states the emotion is low potency and high activity. When moving
in desirable directions, emotions are also differentiated on the non-evaluative
dimensions. Doing well toward goals leads to high potency and activity, while
doing well at avoidance produces lower activity (Carver & Scheier, 1990).
While Carver and Scheier deal more with goals than with self-identities, one
direct descendent of this approach, Higgins (1987, 1989) self-discrepancy theory,
deals directly with different representations of the self. Higgins distinguished
between multiple selves the actual self, the ideal self (a personal standard) and
the ought self (a standard derived from responsibilities to others). Here, the two
major comparisons between the actual and ideal and between the actual and
ought create potential discrepancies that lead to negative emotions. Again, these
emotional responses are different on the potency and activity dimension. Failing
to achieve congruence between the actual and the ideal self leads to depression,
a low potency, low activity emotion. Failing to achieve congruence between the
actual and the ought self leads to anxiety, a high activity, low potency emotion.
Clearly, these predictions are closely related to the approach and avoidance
systems that are the core of the self-regulation model people want to approach
the ideal self and to avoid failing in their responsibilities to others, the ought
self. Higgins arguments differ from Carvers and Scheiers in dealing only with
the negative emotions, and in concentrating on the more social aspects of fear of
punishment by others in not fullling the ought self.
ASSESSING THESE THEORIES
We have presented the dominant control theories in the sociology of identity and
emotion and discussed some related theories in psychology. Now we turn our
attention to a search for physiological analogues of the main concepts invoked by
these theories. There are at least four aspects of affective/emotional response that
Physiology of Deection and Emotion 87
we would need to measure to assess these theories and their predictions. We would
need a measure of the generalized stress/surprise/disequilibrium that corresponds
to deection and/or discrepancy. We would need a measure of positively evaluated
emotion the response to conrming identities in identity control theory and the
positive pole of emotion evaluation in affect control theory. We need a measure of
negatively evaluated emotion the response to disconrming identity meanings
in identity control theory and the negative pole of emotion evaluation in affect
control theory. And, for a more complete assessment of emotion for affect control
predictions, we would need measurements of potency and activation in emotional
response as well.
This theoretical outline may sound like a rather unrealistic goal, given the
current state of our knowledge about the physiology of emotion and the difcult
of measuring it in socially-meaningful experimental situations. We offer two
grounds for hope in our endeavor. First, the voluminous literature on variants of
primary emotions, while it varies a great deal in its details, almost always includes
a few key elements (see Turner 1999a, b, 2000 for a more complete discussion).
Scholars acknowledge that satisfaction/happiness, aversion/fear, assertion/anger
and disappointment/sadness combine with some sense of intensity or activation
to produce basic, primary emotions. This common ground can be mapped onto
a three-dimensional representation (see MacKinnon & Keating, 1989; Morgan
& Heise, 1988) that supports the three-dimensional conception of emotion
in affect control theory (with the activation operating more to discriminate
positive emotions, and activity and potency combining to discriminate negative
emotions). Happiness contrasts with fear, anger and sadness to form the evaluative
dimension, fear contrasts with anger to form the potency dimension, and the level
of intensity forms the activity dimension (contrasting states like irritation with
fury, gratication with elation, and dispirited concern with terror and anguish).
If this roughly three-dimensional representation is so central to the experience
and observation of emotion across cultures and situations, it is more likely that
we will be able to nd reliable physiological analogues to these components of
emotional experience. Our second point of optimism is more theoretical than
empirical in nature. Having two theories that present well formulated models for
when we expect to nd various emotional responses greatly increases our chances
of matching measures to concepts. We will return again in our conclusion to the
route that we believe will be most effective in homing in on an understanding of
the potential physiological measures and how they relate to theoretical concepts;
sufce it to say at this point that we would not even begin this journey without
a clear theoretical understanding of how we expect identity, action and emotion
to interrelate.
88 DAWN T. ROBINSON ET AL.
THE PHYSIOLOGY OF EMOTION
AND (MAYBE) DEFLECTION
There has been an explosion of recent research on the neurophysiology of emotion.
Much of this more recent work focuses on activity in the central nervous system
through the use of positron emission tomography (PET) functional magnetic
resonance imaging (fMRI) techniques. These studies are giving us insight into
processes such as the lateralization of brain activity that will undoubtedly lead
to more rened theories about the relationship between the brain and social
experience. Unfortunately, that theoretical development has yet to happen and
the cost and constraints of these techniques do not suit them well for use in
contemporary group processes laboratory research. Research on peripheral
nervous system activity has experienced a similar explosion. More fortunately for
our purposes, many of the processes generated in this system do lead to outcomes
that are measurable in contemporary group processes research. In fact, recent
technological advances have made many of these measures more reliable, more
affordable, and less intrusive to the point that they can now be usefully employed
even in the context of open interaction research. Accordingly, in the sections
below we will rst describe the physiological processes that appear relevant
to the basic processes we want to study and then present information about
some specic measures that might help group processes researchers to capture
those processes.
We delineate three major types of physiological emotional response behav-
ioral, autonomic, and hormonal (see Carlson, 1998; Frijda, 1986). The behavioral
component refers to activity of the skeletal muscles involved in respiration, muscle
tension, and overt movement. The autonomic nervous system primarily involves
changes in the functioning of smooth muscles and other internal organs, which
have evolved to facilitate the behaviors and provide a quick mobilization of energy
for those actions that might require vigorous motion (e.g. attacking a threatening
intruder or running from it). Hormonal responses are also produced by the au-
tonomic nervous system and reinforce the other autonomic responses by altering
the chemical processes determining blood ow and the use of stored energy by
the muscles.
These systems are obviously interrelated, in the sense that they are all three
evolved through natural selection to help us respond to the environment in ways
that will increase survival and reproduction. Escaping from threats and embracing
good fortune were types of physical activity that helped organisms prevail in
the face of changing surroundings. More problematically for measuring them as
indicators of theoretical concepts, they are related in very complex ways with ex-
perienced emotion. Since more than one action can be implied by a single emotion
Physiology of Deection and Emotion 89
(e.g. freezing or running in response to fear) and more than one emotion can be
implied by a given action (i.e. we can jump out of fear of injury or for joy), none
of these responses are likely to show a direct one-to-one correspondence with a
single, socially interpreted, labeled emotion (Gray, 1994). Neurological responses,
measured directly by brain activity, may have more potential for isolating specic
emotional response. Still, as social scientists, we may benet from the decades
of research before the current emphasis on fMRI and other direct measurement
of neurological activity shifted the emphasis in the physiological measurement of
emotion to this domain.
3
What we hope is that congurations of behavioral,
autonomic and endocrinological responses will correspond to our interests in
stress, positive or negatively valenced emotion, and the additional dimensions
of potency and activity.
The peripheral nervous systemconsists of all of the neurons throughout the body
that are not in the brain, brainstem, and spinal cord (which compose the central
nervous system). The peripheral nervous system is composed of two systems, the
somatic nervous system and the autonomic nervous system. The somatic nervous
system consists of all of our sensory and motor pathways. These include the
neurons that control our skeletal muscles. This system regulates the way that we
directly experience and act upon our environments. The autonomic nervous
system, in contrast, consists of the neurons a number of systems that operate
outside of our conscious direction including the cardiovascular system, the
respiratory system, the digestive system, and the endocrine system.
The autonomic nervous system is further divided into the sympathetic nervous
system and the parasympathetic nervous system. Neurons from the middle (i.e.
thoracic and lumbar) regions of the spinal cord lead to a variety of organs (e.g.
heart, lungs, liver, salivary glands) and comprise the sympathetic nervous system.
Neurons from the brainstem and from the lowest (i.e. sacral) regions of the spinal
cord lead to many of the same organs (e.g. heart, lungs, stomach, bladder) and
comprise the parasympathetic nervous system. Classic theories of the nervous sys-
tem understood the sympathetic and parasympathetic nervous systems to operate
in opposition to one another with the sympathetic response working essentially
to speed things up and the parasympathetic response working essentially to
slow things down. For example, activation of the sympathetic nervous system
leads to pupil dilation, decreased salivation, accelerated heartbeat, increased blood
ow to/from lungs, secretion of epinephrine and norepinephrine, production of
bile, and inhibition of bladder contractions. Parasympathetic nervous response,
on the other hand, includes pupil constriction, increased salivation, decelerated
heartbeat, decreased blood ow to/from lungs, release of bile, and bladder
contraction. These two sets of responses have traditionally been seen as two sides
of the same coin. However, the contemporary view (e.g. LeDoux, 1986) is more
90 DAWN T. ROBINSON ET AL.
complex. For example, both parasympathetic and sympathetic activity produce
related outcomes in the liver the former stimulating production of glucose and
the latter stimulating its release. And, while the sweat glands and peripheral
arterioles (small blood vessels in skin) are stimulated by the sympathetic system
(producing a cold sweat), they are not inhibited by the parasympathetic nervous
system(Carlson &Hateld, 1992). In addition, some experiences seemto produce
partial activation of both systems. For example, when looking at gruesome autopsy
photographs, subjects heart rates slowed and electrodermal activity increased
(Lacey & Lacey, 1970, as cited in Carlson & Hateld, 1992). So, the systems
do not appear correspond in any simplied way to a general sense of high and
low arousal. Rather, it appears that patterns of autonomic nervous response may
correspond in a more specic way to discrete emotional experiences.
The literature on the expression and experience of emotion is vast. It is beyond
the scope of this paper to provide a comprehensive review of the behavioral,
autonomic, and endocrinological responses to emotion. Instead, we will review
the evidence for particular patterns physiological expression that correspond to
the basic theoretical concepts invoked by the sociological theories described
above. Accordingly, we will organize our discussion below around evidence
that appears to link behavioral, autonomic, and endocrinological responses to
deection, positive affect, negative affect, potency, and activity.
The Physiological Experience of Deection
For decades, emotion researchers relied on a notion of generalized physiological
arousal that underpinned our visceral experience of emotion. Empirical evidence
accumulated during the past fty years provides no support for understanding
the physiology of emotion as the result of a singular, general, and diffuse
experience of arousal. However, the mounting empirical evidence that there
maybe physiologically distinct expressions of discrete emotional experiences
does not preclude the existence of a non-specic form of general arousal in
addition to the more emotion specic physiological responses. Affect control
theory and identity control theory both predict that individuals will respond to
deection (or discrepancy) by acting to reduce it. It might be that deection has
no translation into direct experience at all. If it is experienced as anxiety and
distress, then it should show up in measures of negative affect. On the other hand,
if it is experienced as a general sense of doubt, uncertainty, or imbalance, then it
might more closely correspond to more general indicators of stress.
Engineers use the term, stress, to refer to departures from integrity or balance.
Biologists borrow the term and use it to refer to a non-specic departure from
Physiology of Deection and Emotion 91
some preferred internal reference state, or homeostasis. Physiologists and social
scientists alike often use the term in vague ways sometimes referring to stress as
the force that brings about the state of imbalance, sometimes referring to stress as
the outcome of that imbalance. Moreover, while social scientists often recognize
that positive events can and do produce stress, stress is discussed in our research
literatures as if it is primarily negative and most research focuses on responses
to negative stressors. Recent work in the physiology of stress has led some to
doubt on the idea that stress even exists as a generalized response to departures in
homeostasis (see review in Kemeny, 2003). However, the research on stress as
a non-specic response to imbalance has been hampered by research designs
that confound stress with negativity. So, in our opinion, the idea that humans
actually do experience stress in the classic sense as a non-specic response
to departures from homeostasis remains an open question. Accordingly, we
review below evidence for some of the more promising potential measures of
generalized stress.
Behavioral Response
Little of the research on the physiological response to stress seems to provide
evidence for clear behavioral indicators of stress that would cleanly distinguish
it from the experience of emotion and valenced emotion. Consequently, at this
point we offer no suggestions about somatic measures of deection.
Autonomic Nervous Response
The research literature examining stress responses of the autonomic nervous
system is large and growing. In particular, the cardiovascular system seems to
be highly responsive to environmental stressors. Even this literature however, is
somewhat equivocal. For example, while a few notable studies report no statistical
association between blood pressure and perceived stress (Maier et al., 2003;
McCann et al., 1999), a recent meta-analysis of 15 studies looking at responses
to stressful experimental stimuli did nd that blood pressure increased with
stress, when compared to the baseline effects (Feldman et al., 1999). The same
meta-analysis reported that heart rate increased in response to stress (compared to
baseline ndings). The results of these meta-analyses encourage us to look toward
measures of heart rate and blood pressure as potential measures of deection.
Another physiological process implicated in the stress response is vagal tone.
Researchers use vagal tone to index an individuals control of the autonomic
nervous system via the vagus cranial nerve and consider vagal tone to be related
to emotion regulation. In general, heart rate increases when individuals breathe in,
and heart rate decreases when individuals breathe out. One indicator of vagal tone
is a measure of heart rate variability to respiration, or respiratory sinus arrhythmia
92 DAWN T. ROBINSON ET AL.
(RSA). In particular, researchers tend to look at how quickly RSA returns to
baseline after an emotion-provoking event (for a more in-depth description of
vagal tone see Porges et al., 1994). Ostensibly, high vagal tone reects high
regulatory control, while low vagal tone reects low regulatory control. Katz and
Gottman (1995) argue that having high vagal tone helps to buffer children from
the negative effects of marital hostility (which is a stressor), compared to those
children with low vagal tone. Larsen et al. (1986) argue that those individuals who
are able to regulate their physiological responses to stress are likely to experience
less negative emotional arousal, compared to those individuals who are less able
to regulate their physiological response to stress. Regulation of physiological
response can be captured with vagal tone. The intensity of the stressful event mod-
erates the relationship of regulatory capacity and negative emotional arousal. The
need to regulate physiological responses to stress increases as the intensity of the
stress increases. Fabes and Eisenberg (1997) argue, and nd, that in situations of
moderate to high stress, high regulatory control leads to lower negative emotional
arousal. They found that individuals who were high in regulatory control (high
vagal tone) were less likely to experience high levels of negative emotional arousal
in response to stress, but that this nding held only in situations of moderate- to
high-intensity stressors.
We think this measure may help us understand more about deection and
responses to deection. In general, the psychological literature has focused
on individual differences in vagal tone and its role in mediating the effects of
stressors on negative affect. However, if vagal tone indeed reects individuals
efforts to regulate their emotional responses to evocative events, then it might be
that this is in some way reecting the deection an individual is experiencing in
a particular situation.
Endocrine System
There is considerably more work on the relationship between stress and the
physiology of the endocrine system. Early last century, Cannon (1927) coined
the term homeostasis and argued that maintaining internal balance in the bodys
systems was a high physiological priority, arguing that we respond to assaults to
homeostasis with a non-specic ght or ight response, signaled by the secretion
of adrenaline (or epinephrine). Selye (1976) later described another non-specic
response to stressors the secretion of glucocorticoids as a general adaptation
syndrome. These non-specic stress responses bear some resemblance in indi-
viduals who are too hot, too cold, hungry, in pain, or terried. Sapolsky (2002)
argues that even though the necessary response to being too hot or too cold may
be quite different, there is enough in common about the bodys response to being
too hot, too cold, or hungry or in pain to warrant the non-specic response of the
Physiology of Deection and Emotion 93
body to stress. Specically, he argues that even when different types of stressors
throw the body out of homeostasis in different directions, some of the resources
necessary for re-establishing balance are alike. In particular, a stressed individual
needs to mobilize energy. The stress response increases levels of glucose and
oxygen in bloodstream by inhibiting energy storage, breaking down existing
stores into simpler forms, increasing breathing rate, increasing cardiovascular
tone, elevating heart rate and blood pressure, retaining water (to increase blood
volume), and shutting down important parts of the cardiovascular system. Other
non-essential processes and systems are also shutdown including digestion,
reproductive physiology and behavior, growth, tissue repair, inammation, and
pain perception.
The stress response triggers a cascade of events beginning in the brain and
ending in the adrenal cortex. The set of systems involved with this cascade are
the hypothalamus, the pituitary, and the adrenal. Physiologists refer to this system
of interrelated organs as the hypothalamus-pituitary-adrenal (HPA) axis. The
response of the HPA axis to stress begins at the base of the brain, in the hypothala-
mus. Upon perceiving a stress, the hypothalamus releases corticotropin-releasing
hormone (CRH). CRH stimulates the pituitary gland to release adrenalcorti-
cotropic hormone (ACTH) which in turn stimulates the adrenal gland to release
a class of steroid hormones called glucocorticoids. Other hormones released
by the pituitary gland during this cascade include endorphin, a natural opiate,
vasopressin (an antidiuretic) and prolactin (a reproductive hormone). In addition,
the pancreas releases glucagon, which helps to regulate carbohydrates.
The adrenal gland is the last stop in the cascade of responses triggered in the
HPA axis. The medulla, or the core of the adrenal gland, secretes epinephrine
(adrenaline). The adrenal cortex, or the outer layer of the adrenal gland, secretes
glucocorticoids in particular, cortisol. Cortisol is sometimes called the stress
hormone.
Our physiological response to stressful events has implications for more
than how we feel. In recent years, physiologists have made great strides in
understanding the relationship between endocrine responses to stress and the
immune system. We have known for some time that one of the bodys responses
to stress is to suppress the immune system. On the face of it, however, it seems
of questionable utility especially in the face of long term stressors to respond
by making the body more vulnerable to disease and illness. However, recent
research suggests that the immune response to stress is complex, leading to the
strengthening of some systems, the weakening of some systems and short term
bolstering and long term weakening of other systems.
The thymus is the primary site of T cell (immune cell) development and is one
of the central organs at the crossroads of the neuroendocrine and the immune
94 DAWN T. ROBINSON ET AL.
systems. The thymus is extremely sensitive to both acute and chronic stress.
Experiments with rats reveal that social defeat produces a long-lasting reduction
in peripheral blood T cells and as well as a persistent suppression in the process
by which T cells migrate from the bone marrow to the thymus and proliferate
(Stefanski & Engler, 1999). In these experiments, researchers expose rats to
repeated ghting in a subordinate role. Similar experiments show that social stress
directly disturbs homeostasis in the thymus (Engler & Stefanski, 2003). The HPA
axis mobilizes and the adrenal glands respond by secreting more gluticorticoids,
which interfere with the production and migration of T cells, resulting in fewer
immune cells available for attacking infections agents.
As this ongoing research provides an ever clearer specication of the mech-
anisms that link social disruption with well-being and health, it also provides
us some guidance about where to look in the short term to see human stress
responses in social situations. In particular, cortisol seems clearly linked to the
experience of stress in social and non-social situations (Brantley & Jones, 1993;
Brantley et al., 1988; Czeisler et al., 1976; Hellhammer et al., 1985; Mason
et al., 1973; Ockenfels et al., 1995; Smyth et al., 1998; van Eck et al., 1996). In
a study more reminiscent of the television show, Survivor, than a typical group
processes experiment, Jeffcoate and colleagues (Jeffcoate et al., 1986) conned
ve men in a boat for 14 days, measuring plasma levels of hormones at intervals
throughout the study. The researchers found that cortisol levels corresponded
tightly with day-to-day changes in the self-reported anxiety levels of the men.
As with other research on the physiology of stress, there is a frequent confound
between negativity and stress because most of the stressors studied are negative
in valence. Some evidence suggests that cortisol may respond more to valence of
affect than to stress (Buchanan et al., 1999; Smyth et al., 1998). However, other
work suggests that the cortisol response may be more general.
One recent study illustrates the promise cortisol response for serving as a
potential signal of deection: Gaab et al. (2003) measured salivary cortisol in
48 male university students who were exposed to a standard experimental stress
procedure the Trier Social Stress Test. Members of the treatment group received
innoculation style stress management training before being exposed to the
social stressor. Members of the control group received the same training after
the social stress test. Students who had prior innoculating experience with
the stressor in the form of the training displayed an attenuated cortisol response
compared to students in the control group. It appears that the exact same social
experience can trigger different cortisol responses in individuals with different
degrees of expectations about the event.
The endocrine system is complex and creates some unique measurement issues
to contend with in the context of group processes research. We describe some
Physiology of Deection and Emotion 95
of these in more detail in the section below. For example, the diurnal patterns
are very important in the measurement of cortisol. Steptoe et al. (2000) illustrate
this with their study of job stress among local school teachers. These researchers
found that the teachers levels of cortisol measured in saliva between 8:008:30
am positively varied with job stress, while there was no relationship between job
stress and cortisol over the rest of the day. In addition to time of day, there are
several other known sources of systematic variation in cortisol levels that might
serve as potential confounds to group processes research. We revisit these in more
detail in the section on measurement below.
The Physiological Experience of Positive Affect
Behavioral Response
Among the most studied behaviors in the physiology of emotion are those of
the facial muscles. When we emote we move our facial muscles in ways that
correspond to recognizable facial expressions. Sociologists have long attended to
the importance of emotion displays in interaction. However, since the mid-1980s,
emotions researchers have been excited about new opportunities to measure
undisplayed emotions in the face. Cacioppo and colleagues (Cacioppo et al.,
1986) proposed that visually undetectable microexpressions, corresponding to
the discrete emotions could be detected by looking at the activation of various
facial muscles using electrodes attached to the skin. For example, when we smile
we make use of a muscle in our cheeks, the zygomatic muscle. The zygomatic
major draws the corners of the lips upward (Hietanen et al., 1998). Researchers
frequently use activation of the zygomatic major to measure happiness responses
(e.g. Dimberg & Petterson, 2000). However, the empirical literature on this rela-
tionship is somewhat mixed. While Larsen et al. (1992) found that positive affect
increased zygomatic major activity, several other studies report no association
between zygomatic major activity and positive or negative affect (Cacioppo
et al., 1992; Wexler et al., 1992). Bradley and Lang (2000) found no association
with the zygomatic major when participants listened to pleasant vs. unpleasant
sounds (Bradley & Lang, 2000). As suggested by Hietanen et al. (1998), perhaps
this lack of association is due to the fact that the muscles in the lower face are
under more voluntary control of the individual, the zygomatic major being one of
those muscles.
The orbicularis oclui muscle is also involved in the expression of positive
emotion. The orbicularis oculi is the facial muscle that causes the corners of
the eye to wrinkle and bag under the eye (Hietanen et al., 1998). This muscle
is thought be under less voluntary control of the individual than the zygomatic
96 DAWN T. ROBINSON ET AL.
major. Hearing content voices increases responses in the orbicularis oculi muscle
compared to angry voices (Hietanen et al., 1998).
Another facial movement that has received considerably recent attention as a
discriminator of positive and negative affect is the eye blink response to startle.
The eye blink startle response is reexive and not thought to be under voluntary
control. However, it does seem to vary with affect. The empirical literature on
the relationship between positive affect and the startle blink response is not
conclusive, but positive emotion seems to dampen the relationship between
startle/arousal and blinking. Schupp and colleagues found a suppressed startle
blink response while participants were viewing pleasant pictures, compared to
unpleasant pictures (Schupp et al., 1997). The suppression of the startle blink
response also holds when participants are exposed to other types of positively
valenced or appealing stimuli (Codispoti et al., 2001; Ehrlichman et al., 1997;
Lang et al., 1990; Sutton et al., 1997).
However, not all researchers have been able to detect an association between
the suppression of the blink response and positive affect either. Other researchers
looking report no signicant associations between eye blink suppression and
exposure to positive stimuli (pictures, lms, sounds, odors or images), compared
to neutral stimuli (Bradley & Lang, 2000; Bradley et al., 1990, 1996; Cook
et al., 1991; Ehrlichman et al., 1995; Jansen & Frijda, 1994; Larson et al., 2000;
Miltner et al., 1994; Vrana, 1995; Vrana et al., 1988). Skolnick and Davidson
(2002) suggest that this lack of association could be due to the nature of the
tasks. In many of these designs, there may be a confound between positive affect
and arousal. Skolnick and Davidson (2002) argue that the level of arousal of the
stimuli has a greater impact on positive affects suppression of the startle eye
blink response than the impact of negative affects enhancement of the startle eye
blink response, which is discussed in the next section. Signicant suppression of
the startle eye blink response is usually observed when the stimuli are arousing
in nature (see Vrana et al., 1988 for an example).
Autonomic Nervous Response
As an indicator of general sympathetic arousal, some researchers have looked to
cardiovascular activity as an indicator of positive emotion. Compared to neutral
affect, positive affect does seemto be associated with an increase in blood pressure
(Maier et al., 2003). The relationship between positive affect and heart rate is a
bit mixed. While Maier et al. (2003) found a marginal association with positive
affect, Bradley and Lang (2000) found no association with heart rate and pleasant
sounds, compared to neutral sounds. Codispoti et al. (2001) found no association
with heart rate and the valence of the pictures presented.
Another autonomic response researchers have tried to link to positive affect is
skin conductance. Skin conductance refers to the electrical conductivity (or lack
Physiology of Deection and Emotion 97
of electrical resistance) of the skin. Changes in conductance are a function of
sweat gland activity and the skins pore size. An increase in conductivity arises
through increased skin moisture, pre-secretory activity of the sweat gland cell
membranes or both. The relationship between positive affect and skin conductance
activity is not quite clear. While Bradley and Lang (2000) found an increase
in skin conductance activity with pleasant sounds, compared to neutral sounds,
they found no association with pleasant sounds, compared to unpleasant sounds.
Codispoti et al. (2001) observed a similar pattern. While they found an increase
in skin conductance activity with pleasant pictures, compared to neutral pictures,
they did not nd an association with pleasant pictures compared to unpleasant
pictures. It appears as if the increase in skin conductance activity could just be
due to the existence of an emotion, but it does not differentiated between positive
and negative affect. Given the overall pattern of empirical ndings it seems
most likely to us that electrodermal activity is more related to activation than to
affect valence. We will discuss this further in the section on the physiology of
activity below.
Endocrine Response
We could nd no research that attempts to link endocrinological response directly
to positive emotion. As we suggested in our discussion of cortisol responses to
stress, we think that direct examination of the relationship between positive events
and cortisol response will go a long way toward helping us untangle the relationship
between stress, affect, and cortisol.
The Physiological Response to Negative Affect
Behavioral Response
The major facial muscle involved with the expression of negative affect is the
corrugator supercili activity (Larsen et al., 2003). This is the muscle we use to pull
together our brow. When we do so in an upward direction we appear sad. When
we do so in a downward direction we appear angry. Most of the research on the
corrugator muscle seems to focus on anger. When Dimberg and Petterson (2000)
presented angry stimuli to participants, they observed an increase in corrugator
supercili activity. Hietanen et al. (1998) observed increased corrugator supercili
activity increased after participants heard angry voices compared to content
voices. When Jackson et al. (2000) asked participants to enhance their negative
affect, they observed increased corrugator supercili. In contrast, when Jackson
et al. (2000) asked participants to suppress negative affect, they observed
decreased corrugator supercili activity. Other researchers report increased cor-
rugator supercili activity when participants are presented with unpleasant stimuli
98 DAWN T. ROBINSON ET AL.
(scenes, sounds, pictures), compared to pleasant stimuli (Bradley & Lang, 2000;
Cacioppo et al., 1986; Codispoti et al., 2001; Lang et al., 1990). It does appear
that activation of the corrugator supercili may be a useful behavioral indicator of
negative emotion.
The same startle eye blink response that is inhibited by positive affect is
intensied by negative affect. Numerous studies report signicant associations
between the startle eye blink and negative stimuli (pictures, lms, sounds, orders,
or images), compared to neutral stimuli or positive stimuli (Bradley &Lang, 2000;
Bradley et al., 1990, 1996; Cook et al., 1991; Ehrlichman et al., 1995; Jansen &
Frijda, 1994; Lang et al., 1990; Larson et al., 2000; Miltner et al., 1994; Schupp
et al., 1997; Vrana, 1995; Vrana et al., 1988). In a different demonstration of the
relationship between negative affect and the startle blink, Jackson et al. (2000)
asked participants to either suppress or enhance negative affect. They found that
smaller startle eye blinks occurred when participants were asked to suppress
negative affect, while larger startle eye blinks occurred when participants were
asked to enhance negative affect.
Autonomic Nervous Response
The relationship between cardiovascular response and negative affect is unclear.
Both Brondolo et al. (1999) and Maier et al. (2003) found no association between
negative affect and blood pressure. Kamarack et al. (1998), on the other hand, did
nd that negative affect increased blood pressure. And, in a meta-analysis of 15
studies, Feldman et al. (1999) found that increases in undifferentiated negative
emotion were associated with increases in blood pressure. The picture looks more
hopeful and even more complicated when we consider discrete emotions.
For example, some studies show that diastolic blood pressure rises with anger,
compared to fear (Schwartz et al., 1981; Roberts & Weerts, 1982) or sadness
(Schwartz et al., 1981). Several studies report higher nger temperatures for anger
than for fear (see review in Cacioppo et al., 1993).
The relationship between heart rate and negative affect is also unclear in the
literature. Bradley and Lang (2000) found that unpleasant sounds, compared
to pleasant, were associated with heart deceleration. Gross et al. (1994) found
crying to be associated with heart deceleration. However, Codispoti et al. (2001)
found no association between heart rate and valence of picture presented. In
the meta-analysis conducted by Feldman et al., they found that increases in
undifferentiated negative emotion were associated with increases in heart rate.
However, the associations Feldman et al. nd are small. They argue that this
could be due to the tasks involved in the experiments. They also discuss the
methodological problems of measuring emotions via subjective response as a
reason for the small associations. Taken together, the empirical evidence suggests
Physiology of Deection and Emotion 99
to us that cardiovascular responses may vary more closely with potency and
activity than with simple valence. So, we will revisit this literature in the next
two sections.
Researchers have also investigated the relationship between vagal tone and
negative affect. Fabes and Eisenberg (1997) found vagal tone be correlated with
frustration (Fabes & Eisenberg, 1997). Gross et al. (1994) argue that crying is
associated with slower respiration rates and that this independent association may
lead to an increase in RSA. Rottenberg et al. (2003) found that RSA was elevated
both at the onset of crying and at the resolution of crying. Rottenberg et al. (2003)
compared clinically depressed individuals to non-depressed individuals. They
found that depressed individuals did not have an elevated RSA at the onset of cry-
ing. This does not encourage us to look toward vagal tone as a general indicator of
negative affect.
The relationship between negative affect and skin conductance is a bit muddied
as well. While Bradley and Lang (2000) found skin conductance activity to
increase with unpleasant sounds, compared to neutral sounds, they found no
association with unpleasant sounds, compared to pleasant sounds. Codispoti
et al. (2001) found the exact same pattern. While they found skin conductance
activity to increase with unpleasant pictures, compared to neutral pictures, they
found no association with unpleasant pictures, compared to pleasant pictures. As
mentioned in the section on positive affect, it appears as if the association between
negative affect and the increase in skin conductance is due to the presence of an
emotion, the actual valence of the emotion does not seem to matter.
Endocrine Response
As reviewed in the section on deection/stress, there is considerable indication
that cortisol may be a good indicator of general negative affect. Because of the
theoretical advantages of nding a measure of deection that is independence
of emotion valence, we are holding out hope that this is not true. However, it
remains an open question and one that we hope group processes researchers might
consider a useful one to pursue.
The Physiological Experience of Potency
Affect control theory distinguishes between the potency of an emotional experi-
ence and the potency of an actors transient identity impressions. In affect control
theorys emotion model, the powerfulness of ones transient identity impressions
correlates with the powerfulness of ones experienced emotion in a situation.
However, this correspondence is not as strong as it is on the evaluation dimension
100 DAWN T. ROBINSON ET AL.
(as demonstrated in the reduced formula presented above). Hence, it is possible
to feel a weak emotion while operating in a powerful identity, and vice versa. We
can imagine, then, reasons for wanting to be able to capture both of these aspects
of potency. So, in this section we review evidence for physiological measures
that might be related either to the potency of discrete emotions, or the potency of
transient identity impressions.
Behavioral Responses
Recall, that corrugator supercili is the muscle above the eye thought to be related
to anger (Dimberg & Petterson, 2000; Dimberg & Thunberg, 1998). More specif-
ically, the movements of the zygomatic major distinguish between the negative,
powerful emotion of anger and the negative, weak emotion of sadness. This is the
muscle that we use to scrunch the brow together when we are angry or sad. When
we are angry, we bring the center part of our brow together and down (Ekman &
Friesen, 1975). When we are sad, the center part of our brow comes together in
an upward direction (Ekman & Friesen, 1975). Thus, when controlling for nega-
tive affect, activation of the corrugator supercili might be useful in distinguishing
between more and less potent emotions.
Creative work done by Stanford Gregory (Gregory & Gallagher, 2002; Gregory
& Webster, 1996; Gregory et al., 1997) reveals another interesting physiological
response that we might use to signal potency, vocal frequencies. Gregory and
colleagues point out that as air moves through the speech organs (e.g. pharynx,
sinuses, nasal cavities, mouth) those organs act as lters or resonators for the
frequencies passing through. Using spectral analysis to look at non-verbal vocal
frequencies, Gallagher and Webster (2002) measured the fundamental frequency
of presidential candidates voices during presidential debates. They found that
the fundamental frequency of the candidates voices predicted relative social
dominance as well as popular vote. While this vocal stress shows promise, then,
as a measure of relative potency, earlier work by Gregory and Webster (1997)
leads us to pose a caveat. In this study, the researchers found that, as with other
paralinguistic behaviors, lower status speakers tend to accommodate their vocal
frequencies to those of higher status partners. Research attempting to use vocal
frequency as a signal of transient potency would also have to bear in mind the
effects of local relative status on accommodation patterns.
Autonomic Nervous Response
There is some evidence to think that cardiovascular response may be a good place
to look for signals of emotion potency. As reviewed in the section on negative
affect, diastolic blood pressure rises with anger, compared to fear (Roberts &
Weerts, 1982; Schwartz et al., 1981) or sadness (Schwartz et al., 1981). Also, nger
Physiology of Deection and Emotion 101
temperatures tend to rise more for anger than for fear (see review in Cacioppo
et al., 1993).
It is also possible that skin conductance maybe be a useful indicator of potency.
For example, Markovsky (1988) found an increase in skin conductance reactivity
in response to a perceived injustice. However, it is far from clear whether this
response was a result of transient potency, or deection. Most studies that
compare skin conductance levels across between emotion conditions and control
conditions nd signicant effects (see review by Cacioppo et al., 1993). To date,
however, skin conductance has not proved to be a replicable way to distinguish
more critically between disparate emotions.
Endocrine Response
Endocrine response may well provide some insight into potency dynamics
among humans. Testosterone, produced by the gonads, is another one of the
steroid hormones. Losing a competition decreases testosterone (Elias, 1981;
Mazur, Booth & Dabbs, 1992). Winning a game of chance increases testosterone
(McCaul et al., 1992). Jeffcoate et al. conned ve men to a boat for fourteen
days, measured cortisol prolactin and testosterone, found testosterone varied with
position in emergent dominance hierarchy. There is also some indication that
cortisol levels vary by gender and socioeconomic status (Kunz-Ebrecht et al., in
press; Steptoe & Marmot, 2002; Steptoe et al., 2002).
Finally, work on non-human primates nds evidence for a relationship between
rank and stress. In species with stable hierarchies (e.g. rhesus monkeys, baboons,
rats), low ranking individuals experience greater stress. In species with less clear
hierarchies (e.g. marmosets and tamarins), there is little relationship between rank
and stress. In contrast to both of those, among mongooses, whose hierarchies
are more turbulent, high ranking members display the greatest stress responses
(Sapolsky, 2002). Rather thanspeakingtostress hormone (i.e. cortisol) as anindica-
tor of potency, this suggests an interaction between potency, deection, and social
structure that we might be able to assess with measures of stress (such as cortisol).
The Physiological Experience of Activity
Behavioral Responses
We might expect individuals who experience more active emotions or more
expressive transient impressions to move more and with larger motions. However,
beyond classical approaches to coding body motions from videotape, there maybe
more systematic physiological responses that could help us distinguish activity
from other dimensions of affect. Although, the relationship is not straightforward,
102 DAWN T. ROBINSON ET AL.
the startle eye blink is one possibility. As we mentioned in our reviewon the startle
eye blink above, the startle blink response seems to be mediated by the level of
overall arousal. The higher the arousal (either positive or negative), the greater
the startle response (Cuthbert et al., 1996).
Autonomic Nervous Response
The ndings on activity and the cardiovascular system are much less muddy.
General activity and arousal are clearly associated with increases in blood pressure
(Gellman et al., 1990; Jacob et al., 1999; Pollard &Schwartz, 2003; Schwartz et al.,
1994). Inaddition, ingeneral, heart rate seems tobe higher for more active emotions
(e.g. anger) than for quieter emotions (e.g. happiness, sadness) (see review in
Cacioppo et al., 1993).
Finally, there seems to be a relationship between arousal and electrodermal
response. The assumption of this relationship underlies the use of polygraph
techniques which use electrodermal responses along with respiration, pulse,
and relative blood pressure to measure arousal in response to (supposed) lying.
Markovskys (1988) ndings on injustice and galvanic skin response referred to
above are consistent with the idea that arousal increases electrodermal activity.
In addition, Bradley and Lang (2002) found that arousing (pleasant) sounds,
compared to neutral sounds, generated increased electrodermal responses. Taken
together, we think that electrodermal activity may be a promising physiological
marker of activation.
Endocrine Response
In our review of the literature, we found no compelling evidence for looking to the
endocrine system to help us distinguish activity from other dimensions of affect.
PULLING IT TOGETHER: PROVISIONAL
PHYSIOLOGICAL PROFILES OF
SOCIOLOGICAL CONCEPTS
As should be clear from our review of the information above, the physiological
literature on the measurement of stress and emotion is undergoing rapid develop-
ment and revision. However, from the existing literature, we have suggested some
patterns that may advance our ability to test theoretical ideas in the sociology of
identity and emotion. Table 1 summarizes our conclusions drawn from the previ-
ous sections. More specically, Table 1 shows our tentative suggestions for linking
proles of measurable physiological response to the ve theoretical concepts
P
h
y
s
i
o
l
o
g
y
o
f
D
e

e
c
t
i
o
n
a
n
d
E
m
o
t
i
o
n
1
0
3
Table 1. Provisional Physiological Proles of Concepts in Control Theories of the Sociology of Emotion and Identity.
Theoretical Concept
Deection Positive Affect Negative Affect Potency Activity
Predicted
physiological index
Higher levels
of cortisol
Activation of
zygomatic muscle
Activation of
corrugator muscle
Up/down movement
of corrugator muscle
Pupil dilation
Increased
electrodermal activity
Increase in diastolic
blood pressure
Increase in nger pulse
Increased
heart rate
Activation of
periocular region
Higher levels of
cortisol?
Higher levels of
testosterone
Increased
blood
pressure
Decreased eye
blink response to
startle
Increased eye blink
response to startle
Lower fundamental
frequencies in vocal
spectra
Changes in
vagal tone
Increase in nger
temperature
Increase in face
temperature
104 DAWN T. ROBINSON ET AL.
critical for testing ideas derived from affect control theory and identity control
theory. In the following section we revisit these physiological responses with
suggestions about how to measure them in the context of typical group processes
research settings.
MEASURING PHYSIOLOGICAL RESPONSE
Behavioral Responses
Facial Muscular Activity
There are two main approaches to studying facial muscular activity. The more
traditional approach uses videotapes of actors and hand coding of emotion
displays. The most developed approach in this tradition is the Facial Action
Coding System (FACS) developed by Ekman and Friesen (1978). This approach
uses frame-by-frame visual coding of muscle movements in order to look for both
intended and unintended (leaked) emotion displays. The second approach, Facial
electromyography (EMG) measures the electrical activation of specic muscles in
the face. These methods require attaching electrodes to the skin in order to assess
electrical activation of the specic muscles at a specic location. The advantage
of EMG over FACS is that it is able to measure activation in muscles whose
movements are not visible to the eye (e.g. a suppressed smile). One disadvantage of
this method requires that a researcher must choose specic muscles to investigate,
prior to collecting data. Consequently, the method is not well-suited for collecting
information about multiple emotions over the course of an interaction. In fact, this
method is not well suited for collecting data during interaction. This method is
fairly socially intrusive and somewhat sensitive to motion. Electrodes hang from
the face and connect an individual to a data collection unit. These data must be
collected in laboratory setting, rather than in the eld. Another drawback of EMG
is that it is unable to assess how the specic facial movement moved, it can only
determine if there was muscle activity. More traditional methods (such as FACS)
also enable a researcher to determine the character of a particular facial muscle
movement. For example, while EMG could only reveal that there was brow activ-
ity; with FACS a researcher could determine whether the browwent up or down. A
newpotential alternative to EMGor the traditional hand coding of videotapes is the
use of high-denition thermal imaging. New thermal imaging cameras can detect
increased blood ow to parts of the face in the form of spectral images reecting
warming patterns. This option may be an answer to the lack of specicity and
intrusiveness of the EMG and the cost and measurement errors associated with the
time-consuming FACS procedures.
Physiology of Deection and Emotion 105
Autonomic Nervous Response
Cardiovascular Activity
The cardiovascular system is part of the autonomic nervous system. There are
many different measures are used by researchers to capture cardiovascular activity,
including heart rate, heart rate variability, respiratory sinus arrhythmia (which
captures vagal tone), blood pressure, and peripheral vasoconstriction.
Heart rate and heart rate variability can both be measured either by nger
plethysmograph (FP) or electrocardiogram (ECG). The use of FP is fairly
non-intrusive and fairly accurate. It allows researchers to record continuous real
time data and link it to other aspects of interaction. However, ECG measures tend
to be less prone to measurement error especially those produced by motion.
Giardino et al. (2002) conducted two experiments comparing FP and ECG
and found that the FP is more sensitive to non-resting states, thus less reliable.
Giardino et al. (2002) argue that while FP may yield reliable results under resting
conditions, they express caution in its use under non-resting states.
One noninvasive method for indexing vagal tone is respiratory sinus arrhythmia
(RSA). RSA is used as a measure of parasympathetic nervous system activity.
RSA is a measure of heart rate variability to respiration patterns. In order to
measure RSAone must be able to detect both the heart beat and the timing between
heart beats, which are referred to as heart periods (for a detailed description of
the calculation of RSA see Fabes & Eisenberg, 1997 or Porges et al., 1994).
Blood pressure is obtained as a multiplicative function of the cardiac output
and total peripheral resistance (Maier et al., 2003). Cardiac output is the amount
of blood pumped by the heart per minute. Total peripheral resistance is the
amount of resistance that is exerted on the blood ow throughout the body. Blood
pressure is often measured using a sleeve, making it a somewhat more intrusive
measure. It can easily be collected in a laboratory setting or in the eld. However,
sleeve-based measurement is not well suited for collection during interaction.
All of the nger transducer collected heart measures are fairly non-intrusive
and can comfortably be used during ongoing interactions. These measurements
can be collected either in a laboratory setting or in the eld. However, all of the
cardiovascular measures are somewhat sensitive to movement and are somewhat
impaired if collected during unrestricted interaction. To maximize the reliability
of these measures, one would want to either minimize movement of research
participants, or nd ways to statistically control for movement.
Electrodermal Activity
Galvanic skin response is a measure of the skins conductance between two
electrodes. Skin conductance is typically measured by applying a small current
106 DAWN T. ROBINSON ET AL.
through two electrodes, placed on the ngers or toes, and the response is seen as a
change in conductance (decrease in resistance) of the skin with time. Electrodermal
activity can be measured using small, relatively non-intrusive nger transducers.
Using portable data collection units enables research in the eld as well as in the
laboratory setting.
Endocrine System Response
The endocrine system is activated by the hypothalamus-pituitary-adrenal (HPA)
axis, which is elaborated on in the section on stress. Cortisol and testosterone
are two steroid hormones that are both measurable in saliva in ways that reliably
relate to blood borne amounts. In the bloodstream, there are both free and bound
hormones. Free (or unbound) hormones are those hormones that have yet to be
bound to any other compound. Once hormones are bound, they cannot traverse
into the brain. In other words, once hormones are bound, they are unable to impact
the brain. The proportion of free hormones in the bloodstream is what is found in
saliva. Only the unbound hormones are able to get into the saliva.
Prior to collecting saliva samples, individuals are asked to chew sugar-free
gum, in order to stimulate the ow of saliva. There are two dominant methods to
collect saliva. The rst method is to have individuals put swabs of cotton, called
a salivette, into their mouths, having them swish it around their mouth, saturating
the swabs with their saliva (for example see Dabbs, 1990). However, cotton
swabs may interfere with the measurement of certain hormone levels (see Dabbs,
1991; Shirtcliff et al., 2001). Specically, it articially increases testosterone
concentrations but not cortisol concentrations. The second method is to have
individuals spit into tubes (for an example see Schultheiss et al., 1999). This has
the added advantage of allowing for the collection of larger amounts of saliva in
order to test for multiple hormones, conduct repeated tests for reliability analyses,
and to save samples for later re-analysis.
The two means of processing saliva are with radiation (radioimmunoassay,
RIA) or chemicals (enzymeimmunoassay, EIA). The advantages to using EIA
are: (a) that it does not require the use of radioactive materials and so there is no
need for special licenses to conduct analyses; (b) EIA requires a lower sample
volume; and (c) the measurement equipment necessary to conduct analyses
is less expensive than the equipment required for RIA. There are also critical
disadvantages to this method. First, the actual kits required to perform an EIA
analysis are much more expensive than the kits required to perform an RIA,
increasing the per sample cost immensely. For the size of studies that group
processes researchers conduct (compared to many medical studies) this cost
Physiology of Deection and Emotion 107
difference could get prohibitive. Second, the EIA method requires more steps in
the processing of the saliva, which introduces greater potential for measurement
error. Finally, recent research suggests that EIA may inate measures of salivary
cortisol, compared to RIA. Perhaps the best news to a sociologist who is interested
in looking at steroid hormones in saliva, but is not interested in processing the
saliva, is that for either the EIA or RIA method, researchers can collect and freeze
the saliva samples and send them out for processing.
Saliva samples can be collected in a laboratory setting or in the eld. Some
studies have used self-collection, although some of this research shows a fairly
high report of cheating (e.g. not collecting samples on the intended schedule,
misreporting details of sample collections; Broderick et al., in press).
There are several things one must keep in mind when designing a study that
involves collecting saliva samples. Steroid hormones are impacted by cyclical
processes (to be discussed below), making it important to use a repeated-measures
design, to control for within-person variability. In experiments, one should collect
several (at least two) samples prior to the experimental manipulation in order to
establish an individuals baseline. Furthermore, it is important to take a baseline
as close to the treatment as possible in order to help rule out the possibility that
the difference between the baseline and the treatment is due to the individuals
regular cycle. Salivary glands take between 5 and 10 minutes to rell, so it is
important to keep the time between samples at a minimum of 10 minutes.
Another thing to keep in mind in designing a study is that after HPA activation,
it takes about 1520 minutes for the hormones to show up in the saliva. Depending
on the length of the experimental manipulation, changes in the steroid hormones
may already be observable at the end of the treatment. So, if the manipulation is
long (more than 10 minutes), it is worth trying to include an additional sample
during the treatment. If a researcher can only afford to do one sample after the
treatment, a sample should be taken about 15 minutes after the treatment. It
is better, however, to have multiple measurements every 15 minutes after the
experimental treatment.
Steroid hormone levels are sensitive to diet, medications, and time of day
(circadian changes). For example, both testosterone and cortisol levels drop over
the course of the day, with the largest decline in the morning (Dabbs, 1990).
When collecting saliva samples it is necessary to collect samples at the same time
of the day (or control for the time of day) and collect information about medical
conditions, medications, and certain habits (e.g. smoking, alcohol, and caffeine).
It is best to advise potential participants to not consume any alcohol the night prior
to saliva collection. It is best to only use individuals who are not on any kind of
medication. Since age matters too, it is best to use participants of approximately
the same age (or control for age) (Riad-Fahmy et al., 1987). Participants should
108 DAWN T. ROBINSON ET AL.
also be advised not to brush teeth their teeth for an hour before saliva collection.
Brushing ones teeth cause micro lacerations, which can cause blood to be in the
saliva, which inates the hormone concentrations in the saliva.
There are also sex differences in the levels of salivary hormones. If one is using
female participants it may be important to collect information on their menstrual
cycle. Some hormones are impacted by the menstrual cycle, specically estradiol
and progesterone.
4
Dabbs (1990) found support for the notion that the menstrual
cycle does not impact the level of testosterone in saliva. Also, it is also important
to nd out who is taking oral contraceptives. These impact steroid levels as
well including both testosterone and cortisol (Schultheiss et al., 2003). Finally,
there is some evidence that salivary testosterone measures may underestimate
testosterone levels for females, but not males (Shirtcliff et al., 2002).
CONCLUSIONS AND CAVEATS
In this paper we have outlined a set of concepts that have similar uses in a set of
cybernetic theories that link action and emotion, along with a set of physiological
measures that might be useful for assessing those concepts within the context of
social interaction. The basis of the physiological measures made us condent this
would not be an easy task. Since the autonomic, endocrinological and behavioral
systems that are activated in emotional response have evolved to enable a variety
of overlapping lines of action, we could be sure that the research linking them
to social stimuli would be mixed. Indeed, we would not expect to nd single
physiological measures that corresponded to single emotional states; instead, we
have attempted to develop proles of physiological responses that can be linked
to theoretical concepts.
We wish to emphasize that even these proles are highly speculative at this
point. Since the research using these physiological responses has not been
conducted within the theoretical paradigms discussed here, we are in the very
preliminary stages of linking the two. We do, however, have clear ideas about how
the investigation should proceed. The theoretical research programs reviewed
above can generate clear propositions about how their constitutive processes and
concepts are related. If these propositions are then translated into hypotheses about
patterns of both conventional and physiological measures that will be observed
under different conditions, we should be able to establish the validity of the
measures in the context of the larger research enterprise. Basically, the theoretical
ideas and the physiological measurement of concepts will co-evolve as we learn
more about both. Then, the physiological measures may be able to help us distin-
guish among some of the more subtle differences in the theoretical predictions.
Physiology of Deection and Emotion 109
At the very least, such a program would help us understand how the physiological
aspects of emotion are related to its social construction within the context of
interaction. By having a clearer picture of both, we are bound to gain insights about
their interrelationship.
NOTES
1. Those lucky enough to have been active scholars during this feisty period may
remember roundtable sessions at the American Sociological Association meetings attended
by 3040 people per table, culminating with shouted discussions between Dave Kemper,
Norman Denzin, and Tom Scheff.
2. Carver and Scheier (2000) note that such loops must be embedded with negative
feedback loops to be stable.
3. Gray (1994) uses the old joke about the man looking for his watch under a lamppost
(. . . the light is better here . . .) to illustrate how physiologically oriented psychologists
used the techniques that they had available to look, perhaps in the wrong places, for phys-
iological analogues to specic emotions before precise neurological measurements were
available. While the neurological measures may be more complex than a rst enthusiasm
would indicate, we hope to make use of the wealth of data that were generated while
physiological psychologists were looking under the lamppost at the ANS and ES rather than
neuroimaging.
4. If one is interested in looking at estradiol or progesterone, it is essential to collect
information on the menstrual cycle. Specically, you need to determine the date of last
two menstrual periods. The menstrual cycle is a monthly hormone change so it is best to
test all female participants at the same stage of the menstrual cycle.
ACKNOWLEDGMENTS
We would like to acknowledge the support of National Science Foundation grant
SES 0111291 to Robinson and SES 0110599 to Smith-Lovin.
REFERENCES
Averett, C., & Heise, D. R. (1987). Modied social identities: Amalgamations, attributions, and
emotions. Journal of Mathematical Sociology, 13, 103132.
Bradley, M. M., Cuthbert, B. N., & Lang, P. J. (1990). Startle reex modulation: Emotion or attention?
Psychophysiology, 27, 513522.
Bradley, M. M., Cuthbert, B. N., & Lang, P. J. (1996). Lateralized startle probes in the study of
emotion. Psychophysiology, 33, 156161.
Bradley, M. M., & Lang, P. J. (2000). Affective reactions to acoustic stimuli. Psychophysiology, 37,
204215.
110 DAWN T. ROBINSON ET AL.
Brantley, P. J., Dietz, L. S., McKnight, G. T., Jones, G. N., & Tulley, R. (1988). Convergence between
the daily stress inventory and endocrine measures of stress. Journal of Consulting and Clinical
Psychology, 56, 549551.
Brantley, P. J., & Jones, G. N. (1993). Daily stress and stress-related disorders. Annals of Behavioral
Medicine, 15, 1725.
Broderick, J. E., Arnold, D., Kudielka, B. M., & Kirschbaum, C. (in press). Salivary cortisol sampling
compliance: Comparison of patients and healthy volunteers. Psychoneuroendocrinology.
Brondolo, E., Rosen, R. C., Kostis, J., & Schwartz, J. E. (1999). Relationship of physical symptoms
and mood to perceived and actual blood pressure in hypertensive men: A repeated-measure
design. Psychosomatic Medicine, 61, 311318.
Buchanan, T. W., alAbsi, M., & Lovallo, W. R. (1999). Cortisol uctuates with increases and
decreases in negative affect. Psychoneuroendocrinology, 24, 227241.
Burke, P. J. (1991). Identity processes and social stress. American Sociological Review, 56, 836849.
Burke, P. J. (1996). Social identities and psychsocial stress. In: H. B. Kaplan (Ed.), Psychosocial
Stress: Perspectives on Structure, Theory, Life Course and Methods (pp. 141174). Orlando,
FL: Academic Press.
Burke, P. J., & Reitzes, D. C. (1991). An identity theory approach to commitment. Social Psychology
Quarterly, 54, 239251.
Cacioppo, J. T., Bush, L. K., & Tassinary, L. G. (1992). Microexpressive facial actions as a function
of affective stimuli: Replication and extension. Personality and Social Psychology Bulletin,
18, 515526.
Cacioppo, J. T., Petty, R. E., Losch, M. E., & Kim, H. S. (1986). Electromyographic activity over
facial muscle regions can differentiate the valence and the intensity of affective reactions.
Journal of Personality and Social Psychology, 50, 260268.
Cannon, W. B. (1927). The James-Lange theory of emotions: Acritical examination and an alternative.
American Journal of Psychology, 39, 106124.
Carlson, N. R. (1998). Physiology of behavior. Boston, MA: Allyn & Bacon.
Carlson, J. G., & Hateld, E. (1992). Psychology of emotion. Fort Worth, TX: Harcourt Brace
Jovanovich College Publishers.
Carver, C. S., & Scheier, M. F. (1981). Attention and self-regulation: A control theory approach to
human behavior. New York: Springer-Verlag.
Carver, C. S., &Scheier, M. F. (1990). On the self-regulation of behavior. NewYork: Cambridge Press.
Carver, C. S., & Scheier, M. F. (2000). In: M. Boekaerts, P. R. Pintrich & M. Zeidner (Eds), Handbook
of Self-Regulation (pp. 4184). New York: Academic Press.
Codispoti, M., Bradley, M. M., & Lang, P. J. (2001). Affective reactions to briey presented pictures.
Psychophysiology, 38, 474478.
Cook, E. W., III, Hawk, L. W., Jr., Davis, T. L., & Stevenson, V. E. (1991). Affective individual
differences and startle reex modulation. Journal of Abnormal Psychology, 100, 513.
Cuthbert, B. N., Bradley, M. M., & Lang, P. J. (1996). Probing picture perception: Activation and
emotion. Psychophysiology, 33, 103111.
Czeisler, C. A., Ede, M. C., Regestein, Q. R., Kisch, E. S., Fang, V. S., & Ehrlich, E. N. (1976).
Episodic 24-hour cortisol secretory patterns in patients awaiting elective cardiac surgery.
Journal of Clinical Endocrinology and Metabolism, 42, 273283.
Dabbs, J. M., Jr. (1990). Salivary testosterone measurements: Reliability across hours, days, and
weeks. Physiology and Behavior, 48, 8386.
Dabbs, J. M., Jr. (1991). Salivary testosterone measurements: Collecting, storing, and mailing saliva
samples. Physiology and Behavior, 49, 815817.
Physiology of Deection and Emotion 111
Darwin, C. (1872). The expression of the emotions in man and animals. Chicago: University of
Chicago Press.
Dimberg, U., & Petterson, M. (2000). Facial reactions to happy and angry facial expressions: Evidence
for right hemisphere dominance. Psychophysiology, 37, 693696.
Dimberg, U., & Thunberg, M. (1998). Rapid facial reactions to different emotionally relevant stimuli.
Scandinavian Journal of Psychology, 39, 3945.
Ehrlichman, H., Brown, S., Zhu, J., & Warrenburg, S. (1995). Startle reex modulation during
exposure to pleasant and unpleasant odors. Psychophysiology, 32, 150154.
Ehrlichman, H., Kuhl, S. B., Zhu, J., & Warrenburg, S. (1997). Startle reex modulation by pleasant
and unpleasant odors in a between-subjects design. Psychophysiology, 34, 726729.
Ekman, P., & Friesen, W. V. (1975). Unmasking the face. Englewood Cliffs, NJ: Prentice-Hall.
Ekman, P., & Friesen, W. V. (1978). The facial action coding system: A technique for measurement of
facial movement. Palo Alto, CA: Consulting Psychologists Press.
Elias, M. (1981). Serum cortisol, testosterone, and testosterone-binding globulin responses to
competitive ghting in human males. Aggressive Behavior, 7(3), 215224.
Engler, H., & Stefanski, V. (2003). Social stress and T cell maturation in male rats: transient and
persistent alterations in thymic function. Psychoneuroendocrinology, 28(8), 951969.
Fabes, R. A., & Eisenberg, N. (1997). Regulatory control and adults stress-related responses to daily
life events. Journal of Personality and Social Psychology, 73, 11071117.
Feldman, P. J., Cohen, S., Lepore, S. J., Matthews, K. A., Kamarack, T. W., & Marsland, A. L. (1999).
Negative emotions and acute physiological response to stress. Annals of Behavioral Medicine,
21, 216222.
Frijda, N. H. (1986). The emotions. New York: Cambridge.
Gaab, J., Blattler, N., Menzi, T., Pabst, B., Stoyer, S., & Ehlert, U. (2003). Randomized controlled
evaluation of the effects of cognitive-behavioral stress management on cortisol responses to
acute stress in healthy subjects. Psychoneuroendocrinology, 28, 767779.
Gellman, M., Spitzer, S., Ironson, G., Llabre, M., Saab, P., DeCarlo Pasin, R. et al. (1990). Posture,
place, and mood effects on ambulatory blood pressure. Psychophysiology, 27, 544551.
Giardino, N. D., Lehrer, P. M., & Edelberg, R. (2002). Comparison of nger plethysmograph of ECG
in the measurement of heart rate variability. Psychophysiology, 39, 246253.
Gordon, S. L. (1990). Social structural effects on emotions. In: T. D. Kemper (Ed.), Research Agendas
in the Sociology of Emotions (pp. 134179). Albany, NY: SUNY Press.
Gray, J. A. (1994). Three fundamental emotion systems. In: P. Ekman & R. Davidson (Eds), The
Nature of Emotion: Fundamental Questions (pp. 243247). New York: Oxford University
Press.
Gross, J. J., Fredrickson, B. L., & Levenson, R. W. (1994). The psychobiology of crying. Psychophys-
iology, 31, 460468.
Heise, D. R. (1979). Understanding events: Affect and the construction of social action. Cambridge:
Cambridge University Press.
Heise, D. R., & Calhan, C. (1995). Emotion norms in interpersonal events. Social Psychology
Quarterly, 58, 223240.
Heise, D. R., & MacKinnon, N. J. (1987). Affective bases of likelihood judgments. Journal of
Mathematical Sociology, 13, 133151.
Hellhammer, D. H., Hubert, W., & Schurmeyer, T. (1985). Changes in saliva testosterone after
psychological stimulation in men. Psychoneuroendocrinology, 10, 7781.
Hietanen, J. K., Surakka, V., & Linnankoski, I. (1998). Facial electromyographic responses to vocal
affect expressions. Psychophysiology, 35, 530536.
112 DAWN T. ROBINSON ET AL.
Higgins, E. T. (1987). Self-discrepancy: A theory relating self and affect. Psychological Review, 94,
319340.
Higgins, E. T. (1989). Self-discrepancy theory: What patterns of self-beliefs cause people to suffer?
In: L. Berkowitz (Ed.), Advances in Experimental Social Psychology (pp. 93136). New York:
Academic Press.
Hochschild, A. (1983). The managed heart: Commercialization of human feeling. Berkeley, CA:
University of California Press.
Hochschild, A. (1990). Ideology and emotion management: A perspective and path for future work.
In: T. D. Kemper (Ed.), Research Agendas in the Sociology of Emotions (pp. 117144). Albany,
NY: SUNY Press.
Jackson, D. C., Malmstadt, J. R., Larson, C. L., & Davidson, R. J. (2000). Suppression and
enhancement of emotional responses to unpleasant pictures. Psychophysiology, 37, 515522.
Jacob, R., Thayer, J., Manuck, S., Muldoon, M., Tamres, L., Williams, D. et al. (1999). Ambulatory
blood pressure responses and the circumplex model of mood: A 4-day study. Psychosomatic
Medicine, 61, 319333.
Jansen, D. M., & Frijda, N. H. (1994). Modulation of the acoustic startle response by lm-induced
fear and sexual arousal. Psychophysiology, 31, 565571.
Jeffcoate, W. J., Lincoln, N. B., Selby, C., & Herbert, M. (1986). Correlation between anxiety and
serum prolactin in humans. Journal of Psychosomatic Research, 30, 217222.
Kamarack, T., Shiffman, S., Smithline, L., Goodie, J., Paty, J., Gnys, M., & Jong, J. (1998). Effects
of task strain, social conict, and emotional activation on ambulatory cardiovascular activity:
Daily life consequences of recurring stress in a multiethnic adult sample. Health Psychology,
17, 1729.
Katz, L. F., & Gottman, J. M. (1995). Vagal tone protects children from marital conict. Development
and Psychopathology, 7, 8392.
Kemeny, M. E. (2003). The psychobiology of stress. Current Directions in Psychological Science, 12,
124129.
Kemper, T. D. (1978). A social interactional theory of emotions. New York: Wiley.
Kemper, T. D. (1987). How many emotions are there? Wedding the social and the autonomic
components. American Journal of Sociology, 93, 263289.
Kunz-Ebrecht, S. R., Kirschbaum, C., Marmot, M., & Steptoe, A. (in press). Differences in cortisol
awakening response on work days and weekends in women and men from the Whitehall II
cohort. Psychoneuroendocrinology.
Lacey, J. I., & Lacey, B. C. (1970). Some autonomic-central nervous system interrelationships. In:
P. Black (Ed.), Physiological Correlates of Emotion (pp. 205227). New York: Academic
Press.
Lane, R. D., & Nadel, L. (2000). Cognitive neuroscience of emotion. New York: Oxford University
Press.
Lang, P. J., Bradley, M. M., & Cuthbert, B. N. (1990). Emotion, attention, and the startle reex.
Psychological Review, 97, 377395.
Larsen, R. J., Diener, E., & Emmons, R. A. (1986). Affect intensity and reactions to daily life events.
Journal of Personality and Social Psychology, 51, 803804.
Larsen, J. T., Norris, C. J., & Cacioppo, J. T. (2003). Effects of positive and negative affect on elec-
tromyographic activity over zygomaticus major and corrugator supercilii. Psychophysiology,
40, 776785.
Larson, C. L., Ruffalo, D., Nietert, J. Y., & Davidson, R. J. (2000). Temporal stability of the
emotion-modulated startle response. Psychophysiology, 37, 92101.
Physiology of Deection and Emotion 113
LeDoux, J. E. (1986). The neurobiology of emotion. In: J. E. LeDoux & W. Hirst (Eds), Mind and
Brain: Dialogues in Cognitive Neuroscience (pp. 301354). Cambridge, England: Cambridge
University Press.
LeDoux, J. (1996). The emotional brain: The mysterious underpinnings of emotional life. New York:
Simon & Schuster.
MacKinnon, N. J. (1994). Symbolic interactionism as affect control. Albany, NY: State University of
New York Press.
MacKinnon, N. J., & Keating, L. J. (1989). The structure of emotions: CanadaUnited States
comparison. Social Psychology Quarterly, 52, 7083.
Maier, K. J., Waldstein, S. R., &Synowski, S. J. (2003). Relation of cognitive appraisal to cardiovascular
reactivity, affect, and task engagement. Annals of Behavioral Medicine, 26, 3241.
Mandler, G. (1982). Stress and thought processes. In: L. Goldberger & S. Breznitz (Eds), Handbook
of Stress (pp. 88104). New York: Free Press/Macmillan.
Markovsky, B. (1988). Anchoring justice. Social Psychology Quarterly, 51, 213224.
Mason, J. W., Harltey, H., Kotchen, T. A., Major, M. C., Mougey, E. H., Ricketts, P. T., Jones, L. G.,
& Colonel, M. C. (1973). Plasma cortisol and norepinephrine responses in anticipation of
muscular exercise. Psychosomatic Medicine, 35, 406414.
Mazur, A., Booth, A., & Dabbs, H. M., Jr. (1992). Testosterone and chess competition. Social
Psychology Quarterly, 55, 7077.
McCann, B. S., Benjamin, G. A. H., Wilkinson, C. W., Retzlaff, B. M., Russo, J., & Knopp, R. H.
(1999). Plasma lipid concentrations during episodic occupational stress. Annals of Behavioral
Medicine, 21, 103110.
McCaul, K. D., Gladue, B. A., & Joppa, M. (1992). Winning, losing, mood, and testosterone.
Hormones And Behavior, 26(4), 486504.
Miller, G. A., Galanter, E., & Pribam, K. H. (1960). Plans and the structures of behavior. New York:
Holt.
Miltner, W., Matjak, M., Braun, C., Diekmann, H., & Brody, S. (1994). Emotional qualities of odors
and their inuence on the startle reex in humans. Psychophysiology, 31, 107110.
Morgan, R. L., & Heise, D. R. (1988). Structure of emotions. Social Psychology Quarterly, 51, 1931.
Ockenfels, M. C., Porter, L. S., Smyth, J. M., Kirschbaum, C., Hellhammer, D. H., & Stone, A. A.
(1995). Effect of chronic stress associated with unemployment on cortisol: Overall cortisol
levels, diurnal rhythm, and acute stress reactivity. Psychosomatic Medicine, 57, 460467.
Osgood, C. E., May, W. H., & Miron, M. S. (1975). Cross-cultural universals of affective meaning.
Urbana, IL: University of Illinois Press.
Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957). The measurement of meaning. Urbana, IL:
University of Illinois Press.
Pollard, T. M., & Schwartz, J. E. (2003). Are changes in blood pressure and total cholesterol
related to changes in mood? An 18-month study of men and women. Health Psychology, 22,
4753.
Porges, S. W., Doussard-Roosevelt, J. A., & Maiti, A. K. (1994). Vagal tone and the physiological
regulation of emotion. The Development of Emotion Regulation: Biological and Behavioral
Contributions. Monographs of the Society for Research in Child Development, 59(23,
Serial No. 24).
Powers, W. T. (1973). Behavior: The control of perception. Chicago, IL: Aldine Publishing Co.
Riad-Fahmy, D., Read, G. F., Walker, R. F., Walker, S. M., & Grifns, K. (1987). Determination of
ovarian steroid hormone levels in saliva: An overview. Journal of Reproductive Medicine, 32,
254272.
114 DAWN T. ROBINSON ET AL.
Robinson, D. T., &Smith-Lovin, L. (1992). Selective interaction as a strategy for identity maintenance:
An affect control model. Social Psychology Quarterly, 55, 1228.
Robinson, D. T., & Smith-Lovin, L. (1999). Emotion display as a strategy for identity negotiation.
Motivation and Emotion, 23, 73104.
Rottenberg, J., Wilhelm, F. H., Gross, J. J., & Gotlib, J. H. (2003). Vagal rebound during resolution
of tearful crying among depressed and nondepressed individuals. Psychophysiology, 40, 16.
Sapolsky, R. M. (2002). Endocrinology of the stress-response. In: J. B. Becker, S. M. Breedlove,
D. Crews & M. M. McCarthy (Eds), Behavioral Endocrinology (pp. 409450). Cambridge,
MA: MIT Press.
Schultheiss, O. C., Campbell, K. L., & McClelland, D. C. (1999). Implicit power motivation
moderates mens testosterone responses to imagined and real dominance success. Hormones
and Behavior, 36, 234241.
Schultheiss, O. C., Dargel, A., & Rhodes, W. (2003). Implicit motives and gonadal steroid hormones:
Effects of menstrual cycle phases, oral contraceptive use, and relationship status. Hormones
and Behavior, 43, 293301.
Schupp, H. T., Cuthbert, B. N., Bradley, M. M., Birbaumer, N., & Lang, P. J. (1997). Probe P3 and
blinks: Two measures of affective startle modulation. Psychophysiology, 34, 16.
Schwartz, J. E., Warren, K., & Pickering, T. (1994). Mood, location and physical position as predictors
of ambulatory blood pressure and heart rate: Application of a multi-level random effects
model. Annals of Behavioral Medicine, 16, 210220.
Selye, H. (1976). The stress of life. New York: McGraw-Hill.
Shirtcliff, E. A., Granger, D. A., & Likos, A. (2002). Gender differences in the validity of testosterone
measured in saliva by immunoassay. Hormones And Behavior, 42, 6269.
Shirtcliff, E. A., Granger, D. A., Schwartz, E., & Curran, M. J. (2001). Use of salivary biomarkers
in biobehavioral research: Cotton-based sample collection methods can interfere with salivary
immunoassay results. Psychoneuroendocrinology, 26, 165173.
Skolnick, A. J., & Davidson, R. J. (2002). Affective modulation of eyeblink startle with reward and
threat. Psychophysiology, 39, 835850.
Smith-Lovin, L. (1990). Emotion as the conrmation and disconrmation of identity. In: T. D. Kemper
(Ed.), Research Agendas in the Sociology of Emotions (pp. 238270). Albany, NY: SUNYPress.
Smith-Lovin, L. & Heise, D. R. (1988). Analyzing social interaction: Advances in affect control
theory. New York: Gordon & Breach.
Smyth, J., Ockenfels, M. C., Porter, L., Kirschbaum, C., Hellhammer, D. H., & Stone, A. A. (1998).
Stressors and mood measured on a momentary basis are associated with salivary cortisol
secretion. Psychoneuroendocrinology, 23, 353370.
Stefanski, V., & Engler, H. (1999). Social stress, dominance and blood cellular immunity. Journal of
Neuroimmunology, 94, 144152.
Steptoe, A., Cropley, G., & Kirschbaum, C. (2000). Job strain and anger expression predict early
morning elevations in salivary cortisol. Psychosomatic Medicine, 62, 286292.
Steptoe, A., Feldman, P. M., Kunz, S., Owen, N., Willemsen, G., & Marmot, M. (2002). Stress
responsivity and socioeconomic status: A mechanism for increased cardiovascular disease
risk? European Heart Journal, 23, 17571763.
Steptoe, A., & Marmot, M. (2002). The role of psychobiological pathways in socio-economic
inequalities in cardiovascular disease risk. European Heart Journal, 23, 1325.
Stets, J. A. (2003). Justice, emotion and identity theory. In: P. J. Burke, T. J. Owens, R. T. Serpe &
P. A. Thoits (Eds), Advances in Identity Theory (pp. 105122). New York: Kluwer/Plenum.
Stryker, S. (1980). Symbolic interactionism: A social structural version.
Physiology of Deection and Emotion 115
Stryker, S. (1987). The vitalization of symbolic interactionism. Social Psychology Quarterly, 50, 8394.
Sutton, S. K., Davidson, R. J., Donzella, B., Irwin, W., & Dotti, D. A. (1997). Manipulating affective
state using extended picture presentations. Psychophysiology, 34, 217226.
Thoits, P. A. (1989). The sociology of emotions. Annual Review of Sociology, 15, 317342.
Turner, J. H. (1996). The evolution of emotions in humans: A Darwinian-Durkheimian analysis.
Journal for the Theory of Social Behavior, 26, 133.
Turner, J. H. (1999a). The neurology of emotion: Implications for sociological theories of interper-
sonal behavior. In: D. D. Franks & T. S. Smith (Eds), Mind, Brain and Society: Toward a
Neurosociology of Emotion (pp. 81108). Stamford, CT: JAI Press.
Turner, J. H. (1999b). Toward a general sociological theory of emotions. Journal for the Theory of
Social Behavior, 29, 132162.
Turner, J. H. (2000). On the origins of human emotions: A sociological inquiry into the evolution of
human affect. Stanford, CA: Stanford University Press.
van Eck, M., Berkhof, H., Nicolson, N., &Sulon, J. (1996). The effects of perceived stress, traits, mood
states, and stressful daily events on salivary cortisol. Psychosomatic Medicine, 58, 447458.
Vrana, S. R. (1995). Emotional modulation of skin conductance and eyeblink responses to a startle
probe. Psychophysiology, 32, 351357.
Vrana, S. R., Spence, E. L., & Lang, P. J. (1988). The startle probe response: A new measure of
emotion? Journal of Abnormal Psychology, 97, 487491.
Wexler, B. E., Warrenburg, S., Schwartz, G. E., & Janer, L. D. (1992). EEG and EMG responses
to emotion-evoking stimuli processed without conscious awareness. Neuropsychologia, 30,
10651079.
VIOLENT MALES: A THEORY
OF THEIR EMOTIONAL/
RELATIONAL WORLD

Thomas J. Scheff
ABSTRACT
After a brief review of the origins of this work, a theory of the emo-
tional/relational origins of male violence is outlined, and illustrated by
episodes in Hitlers life. Drawing on earlier work on aggression and
violence, I propose that three conditions lead to rage and violence: (1) No
affectional attachments. (2) A single overarching obsession. (3) Complete
repression of shame. Key features of the theory are illustrated by details in
Hitler biographies. This case suggests a way in which emotions unite leaders
and led, leading to collective violence. Finally, a method that would provide
a preliminary test of the theory is suggested.
INTRODUCTION
At the editors request I rst describe the origins of my interest in the emo-
tional/relational world. Although they involve my personal life in the late 1960s,
for brevity I begin in early 1970s. My interest in emotions at this time had led me
to acquire a marriage counselors license in California. Fortunately for me and
for my clients, I didnt venture into a solitary private practice. Instead I joined a

This chapter is based in part on Scheff (2003).


Theory and Research on Human Emotions
Advances in Group Processes, Volume 21, 117139
Copyright 2004 by Elsevier Ltd.
All rights of reproduction in any form reserved
ISSN: 0882-6145/doi:10.1016/S0882-6145(04)21005-0
117
118 THOMAS J. SCHEFF
group that contained a brilliant and experienced social worker, Marilyn Nadler
(Now living in Santa Cruz). In an unostentatious way, she supervised my work
and that of the other members of our group.
During the two years of Marilyns barely visible guidance, our group thrived.
Our success with cases, many of them quite difcult, led us to believe we were all
good therapists. But when she moved to another town, I realized, as did the other
group members, that our effectiveness depended on her. With minimum training
and experience, we simply werent up to the job. Rather than struggling on, I
decided I would give up my practice, in order to write about what I had learned.
One of the things that I learned during my two years of practice became the
rst step in the development of the present theory. I had noticed that even with
my most successful sessions, that there was one aspect of my training that usually
was useless. I had been taught to encourage angry clients, and those who I thought
might harbor hidden anger, to vent their anger. That is, I suggested that they yell
and scream and hit pillows or empty boxes. The upshot is that this procedure was
seldomeffective. Typically the client that I thought had hidden anger would just go
through the motions of being angry, or even refuse entirely. And the openly angry
client would usually report that he or she felt even worse after acting out the anger
than before they did it. There were exceptions, clients who seemed emboldened,
but these were few and far between.
So I left practice with a puzzle: how does one manage anger? In the late 1970s,
a single sentence caught my eye in a book that seemed otherwise impenetrable:
Lewiss empirical study of emotions in psychotherapy sessions, Shame and Guilt
in Neurosis (1971). The sentence was: Shame and anger have a deep afnity.
Out of the whole book, one sentence resonated. Now I would have to understand
the rest of the book, so I could nd out what that sentence meant.
After struggling for many months, I asked my partner, Suzanne Retzinger, to
explain. She said, in effect, Its not the anger, stupid, its the shame. She and
I learned very quickly how much it helps to look for rejection or insult lying
behindanger. Whenhiddenshame or embarrassment is identied, most of the anger
vanishes. With this recognition, I started on the path of trying to understand shame
and embarrassment, and later, just as important, the nature of close relationships.
MALE VIOLENCE
There is an overwhelming amount of evidence that most violence is committed
by men. Indeed, this evidence is daily reafrmed in the mass media. There is
also a smaller but substantial literature that suggests that men manage emotions
and relationships differently then women. This article proposes that the emo-
tional/relational development of males is one cause of male violence. Although
Violent Males: A Theory of Their Emotional/Relational World 119
there may be other causes as well, this paper focuses on a theory that links the
emotional/relational life of men to a propensity to violence.
There is a substantial literature showing that violent behavior is predominantly
a male domain (Archer, 1994). But little headway has been made in demonstrating
why this is the case. There was a urry of hope for a genetic explanation when the
XYY chromosome in males was discovered, but subsequent research has shown
that it is only marginally related to violence, at best (Karli, 1997). There is also
the possibility that high levels of testosterone may be implicated, but this link
has not been demonstrated (Sapolsky, 1997, pp. 147160). More promising is the
work on depression, gender, and aggression (Feshback et al., 1997). Although
there seems to be a relationship, these studies have been unable to provide a
causal explanation. The theory offered in the present article may be relevant to
this problem, as will be explained below.
The causal explanations that are best documented propose that many different
kinds of factors contribute to the causation of violence. One such statement is by
Gilbert (1994), who documents genetic, economic, cultural, and emotional causes
of male violence. One of his theses, the wayinwhichshame/anger sequences leadto
aggression, is similar to the main argument of the present article. Without denying
any of the other variables that contribute to violence, the purpose of this article will
tofocus onanexplanationof role of emotions andrelationships incausingviolence,
taking this argument somewhat further than Gilbert and its other proponents do.
One last reference to the literature on violence will be necessary before
beginning my own argument. Hinde (1997) suggests that we must decouple
individual aggression fromthe causes of war. He argues that aggressive tendencies
in soldiers are somewhat irrelevant, because they are, for the most part, only
fullling social obligations. War, he proposes, is a social institution, so that in
order to understand its causes, we need in the rst instance to examine social
structures, rather than focus on aggressiveness.
I think that Hindes proposal is valid, in that it adds one more line of study
to the many offered by Gilbert (1994), as explained before. But I think Hinde
overlooks one important way in which individual aggressiveness contributes, on
a mass scale, to violence: the way in which the aggressiveness of the leaders
of nations is reciprocally linked to the aggressiveness of the publics the leaders
serve. I will return to this issue below in my analysis of Hitlers biographies.
EMOTIONAL/RELATIONAL
CONTRIBUTIONS TO VIOLENCE
In the 1980s in the U.S. and Europe, psychotherapists became aware that many
of their patients had as their main complaint numbness, blankness, and lack
120 THOMAS J. SCHEFF
of feeling (Krystal, 1988; Taylor et al., 1997). This condition has been named
alexithymia. It is not clear whether these were new symptoms or that they were
noticed more. Although neither Krystal nor Taylor discuss gender, at least three
out of four of the cases they mention are men. Especially in Western culture, men
are socialized to be alexithymic in varying degrees.
Men all over the world have been socialized to be strong, brave and competent,
which usually has meant suppressing vulnerable emotions, especially fear, grief,
and shame. In most male adults, these emotions are hidden, disguised or suppressed
so consistently that early in childhood being aware of them becomes problematic.
There is also a substantial literature that men are socialized to pay less
attention to attachments to others than women do. To achieve success or at least
survival, men are trained to be less interested in affectional bonds than women
(Lewis, 1976). Suppression of emotion and detachment from others are closely
interrelated. The more one suppresses ones emotions, the more difculty others
have relating to you. And the more isolated from others, the easier it is to suppress
emotions. On ones own, without having to attend to others, one can try to
organize ones life in a way that avoids emotion.
By suppressing emotions and relationships, most men become mobile to seek
accomplishments and jobs. But such mobility comes at a high price: isolation
from self and others, and, as I will argue, a propensity for aggression and violence.
Men are socialized to deal with the outer world, but also to mostly ignore the
emotional/relational (e/r) world. How does isolation from the e/r world translate
into violence?
There are several studies that suggest a somewhat counter-intuitive answer
to this question. These studies propose that the management of one particular
emotion, shame, is crucial. They propose that men are particularly socialized to
suppress this emotion: the sense of being weak, inadequate, powerless, helpless,
impotent, or incompetent. Rather than experience these painful feelings or let
others see them undergoing them, men usually become blank or angry. Shame
itself is harmless, indeed, necessary. Shame is a prime component of conscience,
modesty, and morality. It becomes a problem only if covered over. That is, one
ingredient of violence, its incredible energy, is produced by masking shame with
blankness or anger. The rest of this article will explain how this might happen,
especially when it occurs in conjunction with isolation from others.
SHAME AND ANGER STUDIES
In a clinical study of rage, Gaylin (1984) has suggested that rage usually has its
source in two other emotions, fear and shame, especially shame in the form of
Violent Males: A Theory of Their Emotional/Relational World 121
humiliation. However, it seems obvious that fear and shame could not alone cause
violent reactions because these emotions are so widespread. Like Gaylin, Gilbert
(1994) doesnt deal with his issue in his proposal that shame is an underlying cause
of violence. What other conditions are necessary, in addition?
Gilligan (1996) has considered this question. On the basis of his contacts as
a prison psychiatrist with men convicted of crimes of violence, he suggests three
conditions under which shame results in violence: (1) The shame must be a secret
(often hidden behind a defensive mask of bravado, arrogance. . . . or indifference
p. 111). (2) The offender perceives no other alternative than violence. And (3)
The offender lacks the inhibiting feelings of love, guilt or fear.
This formulation by Gilligan is a good beginning, because it brings the issue of
conditions out in the open. This article will attempt to state the exact conditions
necessary for rage to end in destructive violence. There are several ambiguities
in Gilligans formulation, the rst being what is meant by shame being a secret.
I will deal with each of his three conditions in turn.
Although not stated explicitly, Gilligan seems to imply that although violent
men keep their shame a secret from others, they themselves are aware of it.
There are other studies, however, that suggest that the kind of shame that leads
to violence is a secret not only to others, but also to self. The perpetrators lack
of awareness of his own shame is shown in studies of family violence by Lansky
(1984, 1987, 1989). His cases suggest that in order to lead to blind rage, the
shame component in the emotions that are aroused must be outside of awareness.
UNACKNOWLEDGED SHAME
Hidden shame as the main source of anger is a key theme in the work of Helen
Lewis. It is one of the central themes of her study of differences in emotional
development of men and women (1976). However, she rst discovered the link
between hidden shame and anger in her earlier study (1971) of psychotherapy
sessions. She used a systematic technique to locate shame episodes in several
hundred transcripts, then analyzed each episode, second by second, in the context
in which it occurred.
Unlike most other shame researchers, Lewis made relationship issues equal
to emotional ones. Her analysis of emotions is also considerably more detailed
and documented than other studies. For these reasons, it provides one of the main
sources for the present paper.
Lewis was both a research psychologist and a practicing psychoanalyst at the
time of her rst study of shame (1971). One of her major contributions is the
idea that shame is inherently a social emotion. Her formulation was bio-social:
122 THOMAS J. SCHEFF
human beings are social by biological inheritance. That is, she saw shame as an
instinct that has the function of signaling threats to the social bond. Just as the
instinctual emotion of fear signals danger to life and limb, shame also signals a
potential threat to survival, especially for an infant, threat to a social bond. In this
same vein, Kaufman (1989) proposed that shame dynamics form the interpersonal
bridge that connects individuals who would otherwise be isolated.
In Lewiss empirical study of shame (1971) she encountered shame because
she used a systematic method for identifying emotions in verbal transcripts, the
Gottschalk-Gleser method (1969, 1995). This method involves long lists of key
words that are correlated with specic emotions, such as anger, grief, fear, anxiety,
and shame.
Counting keywords, Lewis found that anger, fear, grief, and anxiety cues
showed up from time to time in the transcripts. But she was unprepared for was
the massive frequency of shame cues. Her methodology was complex, in that
once a shame episode was located by Gottschalks method, Lewis also applied a
qualitative method, analyzing each episode word by word.
Lewiss ndings:
(1) Prevalence: Lewis found a high frequency of shame markers in all sessions,
far outranking markers of the other emotions combined. This nding suggests
that shame was a dominant force in the sessions she analyzed.
(2) Lack of reference: Lewis noted that although shame markers were very fre-
quent, patient or therapist almost never used the word shame or its near
cognates. Even the relatively mild word embarrassment was seldom used.
In analyzing the context in which shame markers occurred, Lewis identied
situations in which the patient recounted a shameful memory, or seemed to
feel distant from, criticized, or exposed by the therapist. These two contexts
generated a cloud of shame markers. Both contexts t the proposition that
shame arises from seeing ones self negatively from the point of view of the
other (Cooley, 1922; Darwin, 1872).
However, patients had two different, seemingly opposite responses in the shame
context. In one, the patient seemed to be suffering psychological pain, but failed
to identify it as shame. Lewis called this form overt, undifferentiated shame. A
patient would usually refer to an emotion or feeling, but the reference misidentied
the shame feeling (This is an awkward moment for me.)
In a second type of response, the patient seemed not to be in pain, revealing
an emotional response only by rapid, obsessional speech on topics that seemed
slightly removed from the dialogue. Lewis called this second response bypassed
shame. Identifying or calling shame by its right name seems to be an important
aspect of understanding and managing it.
Violent Males: A Theory of Their Emotional/Relational World 123
(3) Finally, Lewis noted that there was an afnity between shame and anger. She
found that anger markers in the patients speech were always preceded by
shame markers
1
. Apparently one way of hiding shame is to become angry.
This link is the key to the explanation of rage and violence offered in this
article.
In her study of differences in the way men and women manage emotions
(1976), Lewis cites studies suggesting that the overt, undifferentiated form of
unacknowledged shame is more characteristic of women than of men, and the
bypassed form more characteristic of men than of women. She uses this difference
in the management of shame to explain the higher rates of depression in women
than in men, and the higher level of aggression in men. This difference, I think,
can help explain the causal links between gender, depression, and aggression in
boys and girls that have been reported by Feshback et al. (1997).
However, before continuing on the main theme, it will be necessary to elaborate
on Lewiss idea that most shame is hidden. Since her 1971 study was based
entirely on transcripts of sessions that she had not herself witnessed, she was a
bit coy in naming hidden shame. She called it unacknowledged shame. That is,
she noted that the word shame or its near relatives (embarrassment, humiliation)
was almost never used by either the patient or the therapist in reference to shame
episodes. By calling such shame unacknowledged, she reserved judgment on the
issue of whether the patient was conscious of shame, but not mentioning it.
After completing the study, in her clinical work she realized that patients
seldom acknowledged shame because they were usually unaware of it. They were
in a state of shame, as Lewis put it, but did not feel ashamed. Here is one of the
many examples from clinical practice she provided:
P: Well sometimes it may sound silly, but sometimes on the train when Im riding
a train or something I just . . . if Im doing . . . if Im sitting down or something
(laugh) you know you may think that some people may be staring at you [5]
or you just sort of wonder what type of well . . . why theyre staring at you [5]
(slight laugh). I dont know, I just you know, If, if youre sitting down and
somebody keeps staring at you looking [5] . . .
T: What do you think when that happens? What passes through your mind?
P: Well just that maybe I dont knowif I maybe lookingawkward[1] or something.
I dont know. I cant think of what [6 (blankness)].
T: Anything else cross your mind?
P: You mean concerning what people think about me? Well, most cases, I mean
people in most cases, I think people, you know people I dealt with who might
[have] any bad things, have any bad feelings-or ill will toward me, like [1], you
know other people whom you dont really know too well, I mean you might
124 THOMAS J. SCHEFF
have various little acquaintances dealing with them, and they dont understand
you too well, you know. And I feel very repelled by them [3] (slight laugh)
(1971, pp. 248249).
Lewis says In this excerpt . . . a by-passed shame reaction is experienced entirely
without shame affect. . . . the patient experiences hostile feeling; he is repelled
by his imagined hostile watchers [5]. Thus he experiences his own retaliatory
hostility, devaluing the scornful viewers [1]. There is no shame affect, only shame
imagery and ideation. . . . this patient connected the occurrence of chest pains [3]
with the ideation of being watched or looked at by others [5], sometimes in the
wake of a failure to live up to his ego-ideal.
Unlike any other emotion, shame arousal depends entirely on the specics of
the social relationship. Grief is also a social emotion, but a much simpler one; it
registers loss of a love object. Although fear may be generated by another person,
it is not always social. It is an instinctive reaction to danger to life and limb,
regardless of source. Similarly, anger is usually, but not always socially generated.
Its origin is frustration, of whatever kind. Unlike any other emotion, shame is
relationship specic. Because social relationships are so complex, shame, the
affect generated by them, is much more complex than grief, fear and anger.
THE THEORY IS INCOMPLETE
Although it may be true that unacknowledged shame is one of the causes of
violence, perhaps even the main cause, it would also seem that there must be other
ingredients as well. It is clear that unacknowledged shame, particularly in men,
is not unusual. Indeed, it may be that for most men, shame goes unacknowledged
for most of their lives.
An earlier essay (Scheff, 2002) makes this point by a secondary analysis of
two classic texts: Williss (1977) study of working class boys in an English high
school, and Sennett and Cobbs (1972) study of working class men. Sennett and
Cobb, since they stay with the words of their informants, dont say that the hidden
injury suffered by these men is chronic shame, but the informants words imply
it. Similarly, Willis does not use the language of shame and anger to explain the
pranks and laffs that the lads have at their teachers expense, but their words
and behavior imply it.
There is a parallel in studies of the U.S. prison system in the 1950s and 1960s.
They investigated the different kinds of pain that imprisonment causes. One of
the most intense pains, they argued, was the loss of status (degradation). They
go on to suggest that prisoners react to their loss of status by forging a prisoner
Violent Males: A Theory of Their Emotional/Relational World 125
culture that attempts to restore at least some of their damaged status. These
studies propose that since prisoners feel rejected by their keepers and by their
society, they unite in a culture which rejects the rejecters (McCorkle & Korn,
1954). Hostility from the guards and from society is met by counter-hostility from
the prisoners.
Although shame is not explicitly mentioned in these studies, they imply a
shame/anger dynamic. Rather than acknowledge the shame of rejection, the
prisoners mask it with anger and hostility, much like the lads in Williss study
react to their teachers. If high school boys, working class men, and offenders
in prisons repress shame, then there must be something more to the creation of
violence, as Gilligan suggests. There is clear evidence of the repression of shame
in these men, but only a few of them perpetrate violent acts.
ELABORATING THE
CONDITIONS FOR VIOLENCE
It seems to me that Gilligans second and third condition need to be considered
further, as I have already done with his rst condition, secrecy. His second
condition, that one has run out of alternatives to violence, is called entrapment
in the literature on the conditions for conict (e.g. Brockner & Rubin, 1985). It
may be based, at least in part, on reality, as is the case with those who occupy
low status in ghettoes. Or it may be mostly illusory, as it sometimes is with upper
class perpetrators.
One issue that needs further exploration is the possibility that both Gilligans
second and third conditions are so closely related to unacknowledged shame
that they are facets rather than independent dimensions. Fluster and disruption
of perception and thought is a key feature of shame and embarrassment states,
whether acknowledged or not. The feeling of entrapment, in some cases, might
be entirely due to shame states.
Similarly, the absence of feelings of love, guilt, or fear might also be traced to
the shame process. As already indicated, most men are socialized to be ashamed
of feeling, especially fear, but also feelings of tenderness, love and affection.
The third emotion mentioned by Gilligan, guilt, is itself a shame derivative,
compounded of shame and anger, with the anger directed inward. As indicated
above, many men experience grief, fear and shame as blankness or obsession
rather than emotions. Like feelings of entrapment, the absence of love, guilt and
fear might be facets of the complete repression of shame.
If that is the case, then the question of further conditions for the development of
violence out of shame states might be an issue of the degree to which shame states
126 THOMAS J. SCHEFF
are completely inaccessible. Perhaps violence is generated out of unacknowledged
shame only in the most extreme cases of repression.
One indicator of such extreme cases would be a total lack of humor and
laughter, since laughter directed at self, especially, provides relief or even release
from shame. Although there are recorded instances of laughter by Stalin, they
were not directed at self, but at his victims. Knowledge of the lives of terrorists
like Timothy McVeigh and Osana Ben Laden are much less complete, but so far
I have not come across any instances of humor or laughter in their histories.
THE MECHANISM OF HUMILIATED FURY
Although the studies of shame and anger by Lewis, Lansky, and Gilligan are
basic resources for an understanding of male emotions and violence, they are
lacking in one respect. These earlier studies do not provide a basis for explaining
the incredible energy of states of humiliated fury. These states seem to have an
intensity that is usually absent in normal behavior, and, even more unusual, they
can last a lifetime. To explain this phenomenon, an explicit theory of process,
second by second, would be necessary: how an isolated male progresses from
stimuli to shame, to a shame response, to rage, and nally to aggression and/or
violence.
Lewiss moment by moment analysis of shame/anger episodes in discourse
(1971) suggests a shame mechanism that I have called spiraling (Scheff, 1987,
1994). It seems to me that shame/anger spirals explain the emotional basis for
the high energy level of humiliated fury. Lewis herself has provided a cognitive
explanation that complements the emotional one. According to Lewis (1971),
the predominant cognitive feature of bypassed shame is what she calls obsessive
preoccupation, the narrowing of focus onto a single issue. When an individual
has a propensity to isolation from others, these two processes serve to further
isolate him or her. They also can help explain a lifetime of anger, aggression
and violence.
In this framework, there are three conditions for long term aggression and
violence. The rst is social: isolation, the absence of affectional attachments. The
second is cognitive: obsessive preoccupation. The third is emotional: complete
repression of shame in the form of shame/anger spirals. The rst and second
conditions are straightforward, isolation and obsession. Since the third condition,
repression of shame as shame/shame and shame/rage anger sequences, is not
widely known, I will describe it further.
Most emotional responses are very brief. These responses, since they usually
serve as signals to pay attention, may last only a second or two. Then how can
Violent Males: A Theory of Their Emotional/Relational World 127
emotions such as fear or rage last for hours? Silvan Tomkins (1963) suggested
that the basis for long lasting emotions was what he called emotion binds:
one emotion being bound by another, particularly by shame. Helen Lewis (1971)
made a similar suggestion. She used the term feeling traps. Again, her specic
example involved shame: she thought that shame could be masked by anger, but
then heightened by shame about being angry, and so on.
Although not stated explicitly by Tomkins or Lewis, both seem to imply
that emotions can form closed loops, a self-perpetuating emotional episode that
refuses to subside. A familiar example are people who are blushers. They are so
self-conscious about their blushing that they are ashamed of it. But their shame
about blushing increases the blush, and so on. This particular example suggests a
loop that is not mentioned by either Tomkins or Lewis: shame/shame. But it is this
loop, I believe, that gives rise to the most prevalent form of shame spirals, those
that lead to blankness and withdrawal, as in the case of Sennett and Cobbs working
class men.
The two kinds of shame spirals give rise to two different paths: withdrawal
and silence (shame/shame) and anger, aggression and violence (shame/anger).
The emotional/relational theory of violence outlined here would seem to be
particularly applicable to instances involving long term violence on a massive
scale. Hitlers life history will be used to illustrate the major aspects of this theory.
His history is particularly useful because there are many detailed glimpses of him
both as a child and as a man.
We know that terrorists are usually isolated men like Timothy McVeigh, Ted
Kaycinski, and Osana Ben Laden. Often, as is the case with Ben Laden, their
rhetoric makes many references to shame and humiliation. But little is known
about their life histories. In comparison with Hitler, their biographies are largely
unknown. To what extent did Hitlers biographies suggest the three conditions for
violence suggested by the theory outlined here: isolation from others, a single,
overarching obsession, and complete repression of shame?
HITLER STUDIES
Before embarking on yet another attempt to understand the sources of Hitlers
behavior, it is necessary to discuss the present status of such studies. Rosenbaum
(1998) has provided a comprehensive review of a large number of the most widely
known studies, and in addition, taken the unusual step of interviewing most of the
authors of the studies reviewed. His conclusion is that since there is no consensus
at all among the many authors, that the causes of Hitler actions are still unknown,
and indeed, may be unknowable.
128 THOMAS J. SCHEFF
Rosenbaum notes (1998, p. 85) that many of the authors of Hitler studies are
what has been called exceptionalists. That is, they believe that Hitler and his
actions were so extreme that they cannot be explained, and shed no light whatsoever
on the human condition. Rosenbaum himself falls clearly into this category.
My impression is that the position the various authors take on this issue is
closely related to their training. Those with a literary background are usually
exceptionalists, and the social scientists and psychiatrists are not. Rosenbaum, a
journalist, has a Ph.D in English. He is candid about his bias in the Introduction
(p. xlii). After stating that one of the authors he interviewed proposed that the best
one could do as a Hitler scholar was to be an educated consumer, Rosenbaum
says about himself: . . . if the particular nature of the way this . . . consumer
(Rosenbaum) was educated has any bearing on the book that resulted it may lie
in a predispostion to Empsonian ambiguity and uncertain rather than the certainty
of theory.
Since he was predisposed to nd ambiguity and uncertainty, that is what he
found. Also indicated in the quote above is Rosenbaums nave idea of the role
of theory in social science, that it involves certainty. The role of theory is
that of a hypothesis to be tested, or in the present article, illustrated, rather than
representing certainty of any kind.
Like Rosenbaum and the other authors of Hitler studies, my own predisposition
can be linked to my background as a scientist. Although now a social scientist, my
rst graduate studies were in physics, at UC Berkeley. Early on in physics studies,
one learns an elementary technique that seems applicable to the present instance.
To get some preliminary sense of the nature of a complex equation, the rst
step is usually to ascertain how the equation behaves under extreme conditions,
namely zero and innity. Hitlers case certainly represent a near innity of human
malignity. What case would represent zero?
The instances that comes rst to mind are the saints of human history, such
as Havel, Gandhi, Teresa, or Jesus. Indeed, I know of a scholarly sermon on the
emotional history of Jesus which was a zero-seeking response of a pastor who
read a newspaper report of my talk linking male emotions and violence.
2
Here I
provide an illustration of the opposite extreme, Hitlers behavior as coming near
to having been an innity of malignity and violence.
Repression of Shame
3
The Swiss psychoanalyst Alice Miller (1983) has suggested on origin of Hitlers
psychopathology, the conjunction of the fathers physical and emotional violence
and his mothers complicity in it. Miller argues that the rage and shame caused
Violent Males: A Theory of Their Emotional/Relational World 129
by his fathers treatment might have been completely repressed because of his
mothers complicity. Although she pampered Hitler and professed to love him, she
didnt protect him from his fathers wrath, or allow Adolf to express his feelings
about it. Klara, as much as Adolf, was tyrannized by her husband, but offered only
obedience and respect in return. Because of his mothers love for him, as a young
child Adolf was required not only to suffer humiliation by his father in silence, but
also to respect him for it, a basic context for repression.
In later years Hitler (1927) was to gloss over his treatment by his parents, which
is congruent with repression. He described his father as stern but respected, his
childhood as that of a mothers darling living in a soft downy bed (Bromberg &
Small, 1983, p. 40). However, Aloiss son, Alois Jr., left home at 14 because of his
fathers harshness. His son, William Patrick, reported that Alois, Sr. beat Alois, Jr.
with a whip. Alois Jr.s rst wife, Brigid, reported that Alois Sr. frequently beat the
children, and on occasion, his wife Klara (Bromberg & Small, 1983, pp. 3233).
It would appear that Hitlers early childhood constituted an external feeling trap
from which there was no escape. This external trap is the analogue to the internal
trap proposed by Lewis (1971): when shame is evoked but goes unacknowledged,
it generates intense symptoms of mental illness and/or violence towards self or
others. Under the conditions of complete repression that seem to have obtained,
Hitlers personality was grossly distorted. His biographies suggest that he was
constantly in a state of anger bound by shame.
One indication of Hitlers continual shame/rage were his temper tantrums.
Although in later life some of them may have been staged, there is no question
that in most of his tantrums he was actually out of control. His older stepbrother
reported that even before he was seven, (Gilbert, 1950, p. 18):
Hitler was imperious and quick to anger . . . If he didnt get his way he got very angry. He would
y into a rage over any triviality.
In his teens, Hitlers rages were frequent and intense, evoking such expressions
as red with rage, exceedingly violent and high-strung, and like a volcano
erupting (Kubizek, 1955).
Hitlers early shame-proneness is suggested by the slightness of the provocation
that triggered rage. Kubizeks memoir provides two examples: one occasion on
learning that he had failed to win a lottery, another when he saw Stephanie with
other men. Stephanie was a girl who Hitler longed to meet, but never did so. He
was infatuated with her, but never introduced himself (Bromberg & Small, 1983,
pp. 5556).
The most obvious manifestations of Hitlers shame-proneness occurred after
he became Chancellor. Although easily the most powerful and admired man in
Germany, he was constantly apprehensive (Bromberg & Small, 1983, p. 183):
130 THOMAS J. SCHEFF
His anxieties lest he appear ridiculous, weak, vulnerable, incompetent, or in any way inferior
are indications of his endless battle with shame.
Further manifestations of chronic shame states occurred in his relationships with
women. In attempting to interest a woman in himself (Bromberg & Small, 1983,
p. 183):
even the presence of other persons would not prevent himfromrepulsive groveling. [He would}
tell a lady that he was unworthy to sit near her or kiss her hand but hoped she would look on
him with favor . . . one woman reported that after all kinds of self-accusations he said that he
was unworthy of being in the same room with her.
These latter descriptions of Hitlers shame states suggest overt, undifferentiated
shame, emotionally painful states involving feelings of inadequacy and inferiority.
How then is one to understand the other side of Hitlers personality, his arrogance,
boldness, and extreme self-condence? How could a man so shame-prone also
be so shameless?
Lewiss conception of the bimodal nature of unacknowledged shame may
provide the answer to this puzzle. In addition to the overt shame states discussed
above, Hitler also had a long history of bypassed shame. Many aspects of his
behavior suggest bypassed shame, but I will reviewonly three: his temper tantrums,
his piercing stare (Bromberg & Small, 1983, p. 309) and his obsessiveness.
As already indicated, shame theory suggests that protracted and destructive
anger is generated by unacknowledged shame. Normal anger, when not intermixed
with shame, is usually brief, moderate, and can even be constructive, serving to
call notice to adjustments needed in a relationship (Retzinger, 1991). Long chains
of shame and anger alternating are experienced as blind rage, hatred or resentment
if the shame component is completely repressed. In this case, the expression of
anger serves as a disguise for the hidden shame, projecting onto the outside world
the feelings that go unacknowledged within. According to Lewis (1971), persons
in whom shame is deeply repressed would rather turn the world upside down
than turn themselves inside out. This idea exactly captures the psychology of
Hitlers life-long history of intense rage states, and his projection of his inner
conict on to scapegoats.
The second indicator of bypassed shame is Hitlers demeanor, especially
the nature of his gaze. As early as 16, it was described as blank or cruel
(Bromberg & Small, 1983, p. 51). On the other hand, there are descriptions at a
later time (21) in which he was said to have an evasive manner, of being shy
and never looking a person in the eye, except when he was talking politics (ibid.,
70). These descriptions suggest that Hitler may have been in a virtually permanent
state of shame, manifested as either bypassed shame (the stare) or overt shame
(avoiding eye contact). As his power increased, the bypassed mode was more and
Violent Males: A Theory of Their Emotional/Relational World 131
more in evidence, in the form of arrogance, extreme self-condence, isolation,
and obsession.
Isolation from Others
The biographies and psychological studies emphasize Hitlers isolation as a child
and adult (Bromberg &Small, 1983; Bullock, 1964; Davidson, 1977; Miller, 1983;
Stierlin, 1976). As an infant and youth, he was pampered by his mother. But even as
young as three, the relationship with his father was charged with violence, ridicule,
and contempt. By the age of 6, he apparently was walled off from everyone,
including his mother (Bromberg & Small, 1983; Miller, 1983; Stierlin, 1976).
The three most likely candidates for a close relationship after the age of 6
are August Kubizek, Eva Braun, and Albert Speer. Hitler and Kubizek were
companions for three years, beginning when they were both sixteen. Kubizeks
memoir of Hitler (1955) shows that his relationship to Hitler was not that of friend
but adoring admirer. Kubizek describes Hitler as a compulsive talker, brooking
no interruptions, let alone any disagreement. Lacking any other listeners at this
age, Hitler used Kubizek as a sounding board.
Speer, an architect-engineer, was closest to Hitler among his ofcials during
the last years of WWII. In an interview after the war, Speer revealed that although
he spent countless hours with Hitler, there was no personal relationship between
them (Bromberg & Small, 1983, p. 112): If Hitler had friends, I would have been
his friend.
Her diary (Bromberg & Small, 1983, pp. 107108) shows that Eva Braun,
Hitlers mistress, came no closer than Kubizek or Speer. For most of the fteen-
year relationship, he attempted to keep it hidden, conning her to her rooms
during meetings with others. A few entries suggest the tone of the whole diary.
In 1935, when she was 23 and Hitler 46, she complained that she felt imprisoned,
that she got nothing from their sexual relationship, and that she felt desperately
insecure: He is only using me for denite purposes (March 11). Most of the
women with whom Hitler had sexual relations either attempted or committed
suicide (Small & Bromberg count seven such relationships, with three of them
attempting, and three completing suicide 1983, p. 125). Eva Braun made two
such attempts.
In 1942, Hitler inadvertently suggested his isolation from Eva. Hearing of the
death of one of his ofcials, Fritz Todt, chief of armaments, he said that he was
now deprived of the only two human beings among all those around me to whom
I have been truly and inwardly attached: Dr. Todt is dead and Hess has own
away from me! (Toland, 1976, p. 666.) As Bromberg and Small (1983) note, this
132 THOMAS J. SCHEFF
statement leaves Eva out entirely, mentioning instead a remote man who could
rarely be induced to sit at Hitlers table and a man he could not bear to converse
with, denounced as crazy, and wished dead (p. 150).
Neither as a soldier nor as a politician did Hitler have close attachments. His
experience as an enlisted man in the Army during WWI is illustrative. Although he
was a dedicated soldier who demonstrated courage in battle, he was a loner; he
had no intimates. This may be one of the reasons that although he was decorated
for bravery, he was little promoted after four years, he left the army at the rank
of lance corporal, the equivalent of a private rst class. In his evaluations, he was
described as lacking in leadership.
After becoming the leader of the Nazi party, he moved no closer to human
relationships. A description of his campaign the year before gaining power is
representative (Small & Bromberg, 1983, p. 108):
[In the campaign, Hitler] had almost no real contact with people, not even with his associates,
who felt they were touring with a performer . . . He remained a lone wolf, now. . . . more distant
from his senior associates, and contemptuous of them.
Although the adored leader of millions of people, Hitler apparently had no secure
bond with anyone after the age of six.
A Single Obsession
According to Lewis, the rapidity of speech and behavior that is the prime
outer indicator of bypassed shame is usually accompanied by a primary inner
manifestation, obsessiveness. Persons in a state of chronic shame may avoid and
deny emotional pain by obsessive preoccupation. Hitlers principle obsession,
the Jewish problem, is particularly indicative of unacknowledged shame. At
the center of Hitlers belief system was the concept of racial superiority, i.e.
that the Aryan race was the superior race, the Jewish race, inferior. His many
obsessions with superiority-inferiority, racial purity, pollution and contamination
can be interpreted as operations for bypassing shame.
One can also make the case that Hitler was obsessed with shame itself. The most
frequent sequence in his writings is the progression from shame to pride. Here is
one example, in his discussion of scientic education (Hitler, 1927, p. 427):
There is ground for pride in our people only if we no longer need be ashamed of any class. But
a people, half of which is wretched and careworn, or even depraved, offers so sorry a picture
that no one should feel any pride in it. Only when a nation is healthy in all its members, in body
and soul, can every mans joy in belonging to it rightfully be magnied to that high sentiment
which we designate as national pride. And this highest pride will only be felt by the man who
knows the greatness of his nation.
Violent Males: A Theory of Their Emotional/Relational World 133
There is a reference to pride in each of the four sentences is this passage, but only
one to shame (the word ashamed in the rst sentence). This pattern is character-
istic: an initial reference to shame followed by repetitive references to pride. This
pattern also tends toward denial of shame, which is mentioned only once, compared
to the repeated references to pride, a more respectable, i.e. less shameful emotion.
The meaning of the passage is also of interest, because it may imply proneness
to shame. The phrase any pride at the end of the second sentence suggests that
if a group has any reason for shame, then all pride is lost. A more normal response
would be that we always have reason for both pride and shame; that is the human
condition. I return to further suggestions of Hitlers shame-proneness below.
In the passage just quoted, the references to pride and shame were explicit. In
the following passage, which has the same structure, the references are indirect
(p. 411):
Particularly our German people which today lies broken and defenseless, exposed to the kicks
of all the world, needs that suggestive force that lies in self-condence.
Again there is a progression from shame to pride, with a single reference to shame
followed by repeated references to pride, but this time both feelings are evoked
obliquely. In the rst sentence, there is an image of the German people exposed
to the kicks of all the world. Although the word shame is not used, the image
is clearly one of gross humiliation, of being subject to a humiliating assault by
anyone and everyone.
As in the rst example, the statement moves very quickly from shame to
pride. This time, however, like the reference to shame, those to pride are indirect,
using the cognate self-condence rather than the word pride itself. Although
there are direct references to pride throughout the book, there are many more
indirect references. In addition to those already mentioned, self-condence,
honor, superiority, and faith in ones invincibility, Hitler also frequently invokes
dignity (and being worthy) as valued characteristics (p. 431):
[The task of the folk-state] is not to preserve the decisive inuence of an existing social class,
but to pick the most capable kinds from the sum of all the national comrades and bring them to
ofce and dignity.
This passage contains both of the key elements in Hitlers appeal, community and
pride. It negates social class in the interest of community, and promises prideful
ofce to the most capable, regardless of background.
Most of the manifestations of pride and shame are disguised, requiring reading
between the lines. The emotional content of the following passage (Hitler, 1927,
p. 390) would be invisible unless one realized that the basic shame context is
seeing ones self negatively in the eyes of the Other (Lewis, 1971; Sartre, 1956):
134 THOMAS J. SCHEFF
How terrible is the damage indirectly done to our Germanism today by the fact that, due to the
ignorance of many Americans, the German-jabbering Jews, when they set foot on American
soil, are booked to our German account.
In this passage, Hitler seems to be seeing himself (and the German people)
negatively in the eyes of the Other, the American people. Because the shameful
Jewish migrants speak German, the Americans denigrate Germans and Germany.
In the second sentence, he goes on to protest the injustice of the situation produced
by his imagination. There is a gratuitous element to this passage that is difcult to
dene, but it captures the kind of emotional aura that is characteristic of Hitlers
prose; it is shame-haunted.
In Mein Kampf, there are many manifestations of shame, but they are virtually
always hidden in encoded terms. Hitler repeatedly refers to disgrace, lack of
self-condence, inferiority, and phrases like bowing and scraping (p. 625)
in describing the German people or their representatives. Frequently, shame
manifestations are even more indirect, as in the passage quoted above about the
German-jabbering Jews. One of Hitlers frequent themes is the lack of respect
for Germany by other nations (p. 621):
Will any [nation ally itself with] a state . . . whose characteristic way of life consists only in
cringing submissiveness without and disgraceful oppression of national virtues within . . .
Not only Hitlers statement but his actions were haunted by the specter of shame.
Bromberg and Small (1983, p. 119) note in passing Hitlers obsession with
giantism, of building bigger than anyone. He explained to the workers on one of
his building projects (Speer, 1970, pp. 69, 107):
Why always the biggest? I do this to restore to each individual German his self-respect . . . I
want to say to the individual: We are not inferior; on the contrary, we are the complete equals
of every other nation.
Because the references to pride and shame are in code language, Bromberg and
Small miss their signicance. A substantial part of his nations resources, even
during wartime, were devoted to the attempt to make Hitler and his followers feel
large (proud) rather than small (ashamed).
The primary manifestation of shame in Hitlers behavior was not in
construction, however, but in destruction. The sequence [unacknowledged
shamerageaggression] can be traced in particular passages in Mein Kampf,
as well as in the thrust of the book as a whole. The following passage is
representative. In one of many attacks on the Treaty of Versailles, after describing
it as an instrument of abject humiliation, he states (p. 632):
How could every single one of these points have been burned into the brain and emotion of this
people, until nally in sixty million heads, in men and women, a common sense of shame and a
Violent Males: A Theory of Their Emotional/Relational World 135
common hatred would have become a single ery sea of ame, from whose heat a will as hard
as steel would have risen and a cry burst forth: Give us arms again! (p. 632).
In this excerpt, the text moves from humiliation to fury to aggression, the latter
step in the form of re-armament for the battle that Hitler prescribes as the destiny
of Germany.
What is the battle for which Hitler wants Germany to prepare? It is a battle
against the external and the internal enemy. At rst sight it appears that France is
the external enemy, since he refers many times to the eternal conict between
the two countries (e.g. p. 674). He also repeatedly refers to the French motive for
destroying Germany, the thirst for vengeance (e.g. p. 624), with great indigna-
tion, quite oblivious of his own vengefulness. Hitler does not aver revenge as his
own motive, but is quick to detect it in others e.g. he ascribes to another hereditary
enemy, Negroes, the perverted sadistic thirst for vengeance (p. 624).
As becomes apparent quite quickly, however, the ultimate enemy that Hitler
sees everywhere is the Jewish people, or as he puts it, the International Jew.
Hitler had a classical idee x, a fanatical and unswerving belief. Behind every
enemy nation, race, occupation or class, the source of every disaster, is the Jewish
conspiracy, whose aim is world conquest. Hitlers rage is directed against Jews,
who he confabulates with all other enemies. In Hitlers discourse, capitalists,
traitors, revolutionaries and Marxists are either Jews themselves or in the pay
of Jews.
In Mein Kampf, Hitlers solution to what he calls the Jewish problem is only
slightly disguised. Hitler repeatedly alludes to the settling of accounts and a
day of reckoning. In the middle of the last chapter, which has the ominous title
The Right of Emergency Defense, Hitler gave a foretaste of what he had in
mind (p. 679):
If at the beginning of the War and during the War twelve or fteen thousand of these Hebrew
corrupters of the people had been held under poison gas, as happened to hundred of thousands
of our very best German workers in the eld, the sacrice of millions at the front would not
have been in vain.
The cycle of shame and rage is focused on a mythical enemy, the Jewish conspiracy
and those Hitler believed in its pay, but his destructive aggression killed millions
of real people. The appeal of Hitler to the German people also suggests the way
that the psychology of shame and anger, en masse, mobilizes the energies of
an entire people for war, in a way overlooked by Hindes (1994) proposal that
warfare is mostly independent of individual tendencies toward aggression.
Hitlers appeal to the German people is otherwise a puzzle. In his person, he
was singularly unprepossessing, to say the least. From a logical point of view,
his speeches were disasters; he rambled incoherently, with little order and less
136 THOMAS J. SCHEFF
substance. His political program was no better; it was disorganized, vague, and
silent on key issues.
Beneath the surface, matters were still worse. From the testimony of his
intimates, Hitlers personality was bizarre to the point of madness. His delusions,
phobias, sadism, sexual aberrations and utter isolation are well documented. All
of the biographies clearly show manifold symptoms of severe mental illness.
The puzzle is that this extraordinarily unattractive madman had charismatic
appeal not only to the masses but also to a large coterie of devoted followers.
These individuals knew most or all of the unsavory details, yet were fanatically
loyal. The theory of shame dynamics offered here suggests a solution: Hitlers
obsession with restoring his lost pride and that of his nation was the key to his
vast appeal to his public and his followers, because they were suffering from
exactly shame as he was, in the aftermath of their defeat in WWI and their ensuing
humiliations.
DISCUSSION
The causal explanation offered here is hypothetical; it needs to be tested in a wide
variety of circumstances. How would one go about such testing? One direction
would be to employ a method similar to the one used by Helen Lewis in her study
of emotional episodes in psychotherapy sessions (1971). As already indicated, she
used the Gottschalk/Glaser method to locate emotion episodes in transcripts of the
sessions, than analyzed the episodes, word by word, in the verbal context in which
they occurred.
Since the time of that study, Gottschalk has further developed his method, to
the point that it is now based on a computer software program (available from
Louis Gottschalk, Psychiatry, UCI). This method has by now been validated in
several different languages (Gottschalk, 1995). It could be used to nd shame
and anger indicators in verbatim transcriptions of discourse, the rst condition
of the theory. To test the second and third conditions, the quality and size of the
individuals interpersonal network could be assessed, and the degree and type of
obsessive preoccupation.
Gottschalk and his associates have found that a ve-minute speech sample
yields sufcient information to allow the measurement of cues to emotion and
other indicators (many studies are summarized in Gottschalk, 1995). To test
the theory proposed here, a rst step would be record brief life histories from
violent and non-violent men and women convicted of crimes. For the violent
group, emphasis might be placed on the recounting of the circumstances of the
aggressive or violent act that led to their imprisonment. For the non-violent group,
Violent Males: A Theory of Their Emotional/Relational World 137
for sake of comparison, emphasis could be place on the recounting of their most
aggressive or violent act, even though it was not the cause of imprisonment.
Such a study might disentangle the correlation between violence and gender.
According to the argument of this article, we would expect to nd the men much
more isolated from affectional attachments than the women (Condition 1). Also
hypothesized is an equally high level of shame cues in the discourse of the men
and women, but much higher anger cues in men. That is, it is hypothesized than
when men fail to acknowledge shame, they are more likely than women to take
the path of shame, anger and aggression (Condition 2). Since the anger response
serves to bypass shame, we would expect more obsessive preoccupation in men
than women (3). Women, on the other hand, are more likely than men to take the
path of overt, undifferentiated shame, silence, and withdrawal, and therefore less
likely to respond with obsession, anger and aggression.
EPILOGUE
In an as yet unpublished paper, Nascar Dads, Arlie Hochschild has argued that
working class men voted for Bush for emotional reasons. Although their economic
interests are not served by Bush, she argues, in a way that parallels the thesis of
this chapter, that they do so out of resentment and because they admire his cowboy
way of managing fear.
NOTES
1. In her study of the verbatim discourse of four marital quarrels, Retzinger (1991) also
found that shame cues always occurred prior to anger cues.
2. The sermon was given by Vicar Margareta Brandby-C oster in Karlstad, Sweden
during my visit there in 2001. Unfortunately, the only text I have is in Swedish.
3. The analysis of Hitlers life and writings is an abbreviated and updated version of
one section of Chapter 5, Hitlers Appeal to the Germans, in Scheff (1994).
REFERENCES
Archer, J. (1994). Male violence in perspective. In: J. Archer (Ed.), Male Violence (Chap. 1). London:
Routledge.
Brockner, J., & Rubin, J. (1985). Entrapment in escalating conict. New York: Springer-Verlag.
Bromberg, N., & Small, V. (1983). Hitlers psychopathology. New York: International Universities
Press.
Bullock, A. (1964). Hitler, A study in tyranny. New York: Harper & Row.
138 THOMAS J. SCHEFF
Cooley, C. H. (1922). Human nature and the social order. New York: Scribners.
Darwin, C. (1872). The expression of emotion in men and animals. London: Murchison.
Davidson, E. (1977). The making of Adolf Hitler. New York: Macmillan.
Feshback, S., Feshback, N., & Jaffe, Y. (1997). Aggressive and depressive tendencies in boys and
girls. In: S. Feshback & J. Zagrodska (Eds), Aggression: Biological, Developmental, and
Social Perspectives (pp. 121138). New York: Plenum.
Gaylin, W. (1984). The rage within: Anger in modern life. New York: Simon & Schuster.
Gilbert, G. (1950). The psychology of dictatorship. New York: Ronald.
Gilbert, P. (1994). Male violence: Toward an integration. In: J. Archer (Ed.), Male Violence
(pp. 352389). London: Routledge.
Gilligan, J. (1996). Violence: Reections on a national epidemic. New York: Vintage.
Gottschalk, L. (1995). Content analysis of verbal behavior. Hillsdale, NJ: Lawrence Erlbaum.
Gottschalk, L., & Gleser, G. (1969). Manual for using the Gottschalk-Gleser content analysis scales.
Berkeley: University of California Press.
Hinde, R. (1994). The institutional basis of violence. In: R. Hinde & H. Watson (Eds), War, a Cruel
Necessity?: The Bases of Institutionalized Violence (Chap. 1). London: I.B. Tauris.
Hitler, A. (1927). Mein kampf. Boston: Houghton-Mifin (1943).
Karli, P. (1997). The study of brain-behavior relationships underlying aggression. In: S. Feshback & J.
Zagrodska (Eds), Aggression: Biological, Developmental, and Social Perspectives (pp. 314).
New York: Plenum.
Kaufman, G. (1989). The psychology of shame. New York: Springer.
Krystal, H. (1988). Integration and self-healing: Affect, trauma, alexithymia. Hillsdale, NJ: Analytic
Press.
Kubizek, A. (1955). The young Hitler I knew. Boston: Houghton-Mifin.
Lansky, M. (1984). Violence, shame, and the family. International Journal of Family Psychiatry, 5,
2140.
Lansky, M. (1987). Shame and domestic violence. In: D. Nathanson (Ed.), The Many Faces of Shame.
New York: Guilford.
Lansky, M. (1989). Murder of a spouse: A family systems viewpoint. International Journal of Family
Psychiatry, 10, 159178.
Lewis, H. B. (1971). Shame and guilt in neurosis. New York: International Universities Press.
Lewis, H. B. (1976). Psychic war in men and women. New York: New York University Press.
McCorkle, L., & Korn, R. (1954). Resocialization within walls. The Annals, 293(88, May).
Miller, A. (1983). For your own good. New York: Farrar, Straus, Giroux.
Retzinger, S. M. (1991). Violent emotions: Shame and rage in marital quarrels. Newbury Park:
Sage.
Rosenbaum, R. (1998). Explaining Hitler. New York: Random House.
Sapolsky, R. (1997). The trouble with testosterone. New York: Scribner.
Sartre, J. (1956). Being and nothingness. New York: Philosophical Library.
Scheff, T. J. (1987). Interminable quarrels: Shame-rage as a social and a psychological spiral. In:
H. B. Lewis (Ed.), The Role of Shame in Symptom Formation. Hillsdale, NJ: LEA.
Scheff, T. J. (1994). Bloody revenge. Boulder, CO: Westview (iUniverse, 2000).
Scheff, T. J. (2002). Working class emotions and relationships: Secondary analysis of Sennett and
Cobb, and Willis. In: B. Phillips & H. Kinkaid (Eds), Beyond Babel: Reconstructing Sociology.
Scheff, T. J. (2003). Male emotions and violence: A case study. Human Relations, 56(6), 727749.
Sennett, R., & Cobb, J. (1972). The hidden injuries of class. New York: Knopf.
Speer, A. (1970). Inside the Third Reich. New York: Macmillan.
Violent Males: A Theory of Their Emotional/Relational World 139
Stierlin, H. (1976). Adolf Hitler: A family perspective. New York: Psychohistory Press.
Taylor, G. J., Bagby, R., & Parker, J. (1997). Disorders of affect regulation: Alexithymia in medical
and psychiatric illness. New York: Cambridge University Press.
Toland, J. (1976). Adolf Hitler. Garden City, NY: Doubleday.
Tomkins, S. (1963). Affect, imagery, consciousness (Vol. II). New York: Springer.
Willis, P. (1977). Learning to labor. New York: Columbia University Press Agg Oct. 2, 9832.
EMOTIONS, SENTIMENTS, AND
PERFORMANCE EXPECTATIONS
Robert K. Shelly
ABSTRACT
Expectation states theories linking status and behavior enhance our
understanding of how social structures organize behavior in a variety of
social settings. Efforts to extend behavioral explanations anchored in state
organizing processes based on emotions and sentiments have proceeded
slowly. This chapter presents a theory of how emotions organize observable
power and prestige orders in groups. Emotions are conceptualized as
transitory, intense expressions of positive and negative affect communicated
from one actor to another by interaction cues. These cues become the basis
of long-lasting sentiments conceptualized as liking and disliking for other
actors. Sentiments become the foundation for differentiated social structures
and hence, performance expectations. This chapter describes how such a
process may occur and develops theoretical principles that link emotions,
sentiments, and performance expectations.
INTRODUCTION
Emotions organize interaction by dening who is to be accorded positive and neg-
ative regard, the intensity of the response to another actor, and role appropriate
actions such as displays of affection and disdain. The fundamental question is
whether such differential expressions of emotion toward others may become the
Theory and Research on Human Emotions
Advances in Group Processes, Volume 21, 141165
Copyright 2004 by Elsevier Ltd.
All rights of reproduction in any form reserved
ISSN: 0882-6145/doi:10.1016/S0882-6145(04)21006-2
141
142 ROBERT K. SHELLY
foundation for state organizing processes reecting inequalities in power and pres-
tige. The use of state organizing processes based on emotions has received limited
attentioninthe studyof power andprestige orders intaskgroups. Treatingemotions
as a basis for such fundamental state organizing processes expands the range of
phenomena encompassedinthe expectationstates researchprogram. It alsoleads to
fuller substantive understanding of general principles to explain social interaction.
Briey, expectation state organizing processes are based on interaction
dynamics associated with actors preexisting circumstances in a social situation.
1
These processes may favor advantaged actors with more chances to contribute
organizational and task solutions, more positive evaluations for their contributions,
and exercise more inuence over task activities and outcomes. Disadvantaged
actors receive fewer chances to contribute organizational and task solutions, more
negative evaluations, and exercise less inuence over task activities and outcomes.
Actors who are not initially advantaged or disadvantaged with respect to others in
the situation evolve inequalities similar to those observed in differentiated groups
through a process described by behavior status theory. Behavior status theory says
that others behaviors are evaluated by the parties in an interaction and that these
evaluations form the foundation for the development of performance expectations
reected in observable behavior. The conception of a state organizing process
(Fararo &Skvo retz, 1986) is adapted to develop a theory of howemotions and their
expression may lead to the formation of interaction inequalities in task groups.
The chapter proceeds in several parts. First, a series of questions is identied
linking emotion to interaction inequalities. Second, fundamental assumptions are
detailed, basic nomenclature claried, and boundary conditions specied. The
theoretical specication of how emotions may be transformed into sentiments
and thence to performance expectations follows the explication of these basic
ideas. This approach parallels the process developed in behavior status theory
and assumes that actors enter a situation as social equals. No actor in the group
has advantages over any other as interaction gets under way. Any differences
observed which emerge during the interaction are the result of interaction dy-
namics governed by a feedback process based on behavior-assessment-behavior
sequences. I conclude with a discussion of possible extensions of the approach to
additional situations and suggest some potential avenues for investigation.
RESEARCH QUESTIONS LINKING
EMOTIONS AND EXPECTATIONS
Several investigators have examined howemotions are linked to social interaction.
The earliest systematic investigations by Bales (1950) and his associate Slater
Emotions, Sentiments, and Performance Expectations 143
(Bales & Slater, 1955; Slater, 1955) integrated the study of inequality and emo-
tional processes and suggested that parallel social structures emerged as a con-
sequence of interaction dynamics in groups of Harvard undergraduates studied.
Subsequent research has found this result to be suspect (Burke, 1967, 1968, 1971).
Disentangling the relationship between status processes and emotional expression
in interaction is a complex problem. A brief review of other approaches that have
been attempted follows.
One approach focuses on ways in which actors employ emotions and emotional
displays to inuence the behavior of others. This work emphasizes how emotions
may have played a role in the social organization of early human societies (Turner,
2000), how emotions are socially constructed (Kemper, 1981), and how emotions
are employed to manage the behavior of self and others (Heise & Calhan, 1995;
Hochscild, 1983). The common concern of these studies is identication of
behaviors given off in emotional display and their interpretation as information
about internal states of self by both self and others. Actors signal to one another
using emotional displays (Turner, 2000), develop meanings for observed behaviors
of others (Kemper, 1981), and learn to control this behavior by manipulating
emotional displays for personal and organizational reasons (Hochscild, 1983).
The number of emotions employed in interaction is relatively limited (Morgan
& Heise, 1988) and described with a limited number of dimensions (evaluations,
potency, and activity). Heise and his associates (Heise & Calhan, 1995; Heise &
Thomas, 1989; Morgan &Heise, 1988; Thomas &Heise, 1995) have demonstrated
that prescriptive emotional expression is relatively invariant across gender and
group setting. In other words, there is widespread agreement about what ought to
be as emotions are enacted. These results vary by circumstance of the actor with
actors embedded in multiple roles more likely than those with limited contacts
to share this consensus. Variation in what actually occurs with emotional reaction
varies by gender, with females more likely to self-blame than males (Heise &
Calhan, 1995).
Research has identied a minimal set of physiological experiences humans
interpret as emotions, ways actors manipulate expressing them, and ways actors
combine fundamental emotions to arrive at derivative emotions. In addition,
research contributes to our understanding of howsocially constructed meanings of
emotional display and interpretation are accepted in the larger society. Actors learn
to interpret and manipulate the relatively small set of innate human emotions.
2
Who is one of us as opposed to one of them is an important part of this learning
process, as situations in which emotions may be displayed are discriminated from
one another. The appropriate interpretation of emotional stimuli presented by self
and others is also learned (Smith-Lovin, 1995). This last point will be critical in
developing our understanding of how emotions organize interaction.
144 ROBERT K. SHELLY
Result I. Emotions are manifestations of physiological reactions of the organ-
ism. Fundamental emotions are few in number, but may combine to produce
derivative emotions.
The second issue links status with interaction by emphasizing how status is
reected in the expression of emotions. Ridgeway and Johnson (1990) hypothesize
that status processes constrain the expression of negative emotion but have no
effect on the expression of positive emotions. This conclusion is based on two
fundamental principles. The rst is that actors in advantaged positions use their
position to constrain the expression of negative emotions by disadvantaged others.
The expression of positive emotions does not depend on the status position of
the person expressing the emotion, and hence is not constrained by advantage
or disadvantage location in the group hierarchy. These conjectures are based
in part on how actors attribute evaluations of task contributions. Advantaged
actors attribute task success to self and failure to others, while disadvantaged
actors attribute success to others and failure to self. These conjectures explain the
paucity of negative emotional expression in task groups and the disproportionate
concentration of negative targeting aimed at disadvantaged actors.
Lucas and Lovaglia (1998) provide evidence that advantaged actors are
beneciaries of the process hypothesized by Ridgeway and Johnson. Advantaged
actors are happier in task groups than are disadvantaged actors, according to their
data. Gender differences are present in this data. Women reported more positive
emotion than did men and were seen as competent as leaders as were men in
these mixed gender groups. However, groups led by women rated their group
performance lower than did groups led by men. Thus, while individual expression
of emotional experience does not vary by gender, aggregated assessments show a
gender difference.
Result II. The expression and experience of emotions are constrained by
structures of social inequality. Advantaged individuals express and experience
more positive emotion than disadvantaged individuals.
A third concern is how affect processes represented by both emotion and
sentiments impact status processes. We know that positive and negative emo-
tions toward another individual lead actors to alter behavior that is otherwise
predictable based on relative advantage and disadvantage. Lovaglia and Houser
(1996) show that advantaged actors who experience positive emotions toward
their disadvantaged partner dedifferentiate their behavior. That is, they are more
likely to accept inuence from the disadvantaged other. Disadvantaged actors
who experience negative emotion toward their advantaged partner dedifferentiate
their behavior. That is, disadvantaged actors who dislike their partner are less
Emotions, Sentiments, and Performance Expectations 145
likely to accept inuence from the partner than if they are neutral or like the
partner. Similar results are reported by Driskell and Webster (1997) who employ
a different, less intense manipulation of disliking for the interaction partner. Their
research manipulates the affective component so that it produces a response closer
to the concept of sentiment employed below.
Precise mechanisms governing the combination of affect and status processes
are not clear (Fisek &Berger, 1998). One possibility is that information processing
is the same for both affect processes and status processes. Hence, a combining
process similar to that which governs combining of status information may be
appropriate (Berger et al., 1992). This is referred to as the constituent model. In
this model, status elements and sentiment elements are simultaneously processed
in the formation of performance expectations. The other possibility is that status
processes are primary and affect processes enter the information processing
at the point at which actors translate performance expectations to behavior.
This is referred to as the translation model. In this model, status elements are
processed rst and performance expectations formed based only on the status
information. Sentiment elements enter the process at the point at which the
actor translates the performance expectation to behavior. As Fisek and Berger
(1998) conclude, choosing between these alternative models is not possible at the
present time. Explanatory parsimony and work in cognitive psychology (Zajonc,
1980) favor the former over the latter principle. While we can favor one of these
models over the other for logical reasons, initial research to test the predic-
tions of the two different models supports the moderator or translation model
(Bianchi, 2000).
Result III. Emotions and status combine with one another in information
processing to alter status related behavior in groups. The combining mechanism
appears to be similar to the process that governs combining of salient status
information.
The nal issue to consider is the way in which emotions and sentiments are related
to one another and subsequently to performance expectations. This question has
been addressed by several investigators. Shelly and his collaborators (Shelly,
2001; Shelly et al., 2001) have presented evidence that actors process information
about emotions and sentiments in similar ways and that this information leads
to differential assessments of qualities and capacities of target individuals. These
differential assessments of qualities and capacities are reected in performance
expectations an actor holds for others differentiated by positive and negative
sentiment ties (Shelly, 2001).
Similarly Tiedens et al. (2000) have shown that actors hold stereotypes for
emotional expression that associate particular responses with high and low status
146 ROBERT K. SHELLY
expectations for performance. The stereotypes apparently function in reciprocal
directions as subjects in these studies ascribed high status to a particular constel-
lation of emotional expression and low status to a different complex. Subjects also
linked emotions to high and low status others in ways that are consistent with the
expression of stereotypes linking status to identied emotions. In other words, ac-
tors impute emotions and status to others in ways that suggest a naive correlation
between high and low status and positive and negative emotion. High status others
are associated with positive emotions and low status others are associated with
negative emotions.
Result IV. Actors associate emotions and sentiments with status positions in
task groups. Individuals infer status from emotional information about others.
Similarly, they infer status information from sentiments associated with others.
The results identied above, together with principles frombehavior status elements
of the expectations research programwill be employed to elaborate howtransitory,
intense emotions may be linked to enduring sentiments. These sentiments in turn
are the basis for the formation of performance expectations based on information
about similarity and difference an actor has about others in the social environment.
The process of linking sentiments to expectations also requires an explication.
Ideas developed to explain how social objects acquire status value will be applied
to solve this problem. Understanding howemotions may be transformed fromtheir
transitory intensity to the enduring lowintensity sentiments is critical to advancing
our understanding of the links in this process.
BASIC ISSUES
A fundamental a set of assumptions about the relationship of emotion, sentiment,
and performance expectations is necessary. First, emotions, sentiments, and
performance expectations are treated as analogs of one another. Each involves
information given off by an actor that reects internal, unobservable states of that
actor. Often, the actor is unable to report on these internal states and we infer
their existence and effects from observed behavior. Similarly, for all three of these
concepts, received information is processed by the observer(s). It is this processed
information that constitutes the basis of action by the observer(s) and hence leads
to the sequences of behavior referred to as observable social interaction.
Cues are behaviors an actor displays, sometimes unintentionally, that provide
information about the actor and his or her states of mind to others around the actor.
A status cue may be an accent, a use of syntax and vocabulary, style of dress, or
object denoting an ofce such as rank insignia (Berger et al., 1986). Such indicative
Emotions, Sentiments, and Performance Expectations 147
cues convey specic information that explicitly labels an actor as belonging to a
valued social category or as possessing a particular state of a status characteristic.
For instance, in the American Army, two gold bars denote the rank of Captain,
which is higher than the rank of Lieutenant (one gold bar), but lower than the
rank of Major (gold oak leaves). Indicative cues are valuable sources of direct
social information only when categories are differentiated from one another and
have recognizable attributes associated with these categories. In expectations states
research how categories acquire status value is explained by status construction
theory (Ridgeway, 1991; Webster & Hysom, 1998).
Expressive cues are given off in the course of normal interaction and do not di-
rectly convey status information about the actor. Such cues may convey emotional
or sentiment information with a tone of voice, facial expression, or posture. I make
use of this concept to develop the theory of how emotions are translated to perfor-
mance expectations. Some cues may have both emotional and status value. Smiles
and gazes are differentially employed by high and low status actors to connote
both deference (Dovidio & Ellyson, 1982; Dovidio et al., 1988) and emotional
reactions to others or their behaviors.
Task cues indicate an actors ability to perform a particular task. They include
explicit information about the skills and abilities of the actor and may be either
expressive, identifying the actor as a member of a skilled category, or indicative,
which explicitly label the actor as possessing the special skill or ability that
provides a task advantage. For instance, a male is labeled as a member of a social
category expected to have superior mechanical ability based on gender stereotypes
regarding the male sex. This expressive categorization imputes status value when
the task is to repair an automobile. He would generally be thought to be a member
of a disadvantaged group if the task were sex typed as female, for instance
repairing a torn pant leg with a sewing machine. Indicative cues for the task set
described above might include certicates hung on the shop wall showing that
Jane Smith is a master mechanic, or a blue ribbon for sewing Fred Jones won at the
local fair.
Categorical cues place an actor in a particular social grouping and may or may
not convey status information. For instance, wearing a green tie on St. Patricks
Day indicates ones sympathy with the Irish, or membership in that ethnic group.
It conveys no information about the ability of the actor to accomplish a task,
nor does it convey status information in the sense that being Irish, or an Irish
sympathizer, is not currently socially evaluated in American society. On the other
hand, placement in some social categories does convey status value. Hence, we
recognize that status value is associated with membership in disadvantaged ethnic
groups, some categories of the socially stigmatized, and socially privileged.
Figure 1 displays a simple categorization of these four types of cues.
148 ROBERT K. SHELLY
Fig. 1. Types of Cues and Associated Status Information.
Mechanisms by which cues are translated into affective and evaluative
information are well understood. Status processes, an application of evaluative
information, have been examined extensively employing the distinctions raised
above. Zajonc and his collaborators (Murphy & Zajonc, 1993; Murphy et al.,
1995; Zajonc, 1998, 2001) have developed a research program which contributes
to our understanding of how cues contribute to the development of emotions
and sentiments expressed in liking for social objects. Their focus has been
understanding how exposure to a source leads to liking for the source. Two results
are of particular note for our concerns. First, the exposure of stimuli often leads
to liking for the stimulus. This effect is present even if the stimulus is presented at
a level below the threshold at which a person can knowingly process information.
Second, the effect is robust across situations and mechanisms of presentation (sub-
liminal, supraliminal, or primed). The effect is also observed when the stimuli are
highly similar to one another (Monahan et al., 2000). We can conclude that cues,
even those that are not actively processed by an actor, produce informed affective
responses such as liking and disliking. How affective information organizes social
activity is our next concern.
Affect information is conceptualized as providing actors with ways to place
others in social categories based on expressive cues. Transformation of these
undifferentiated social categorizations to differentiated states is the central issue
in explicating how actors assign status value to emotions and sentiments. First,
we must delineate the phenomena of interest and the range of situations in which
these transformations may occur.
The various concepts employed in discussions of affect processes must be dis-
tinguished as they provide different bases for action for an individual depending on
Emotions, Sentiments, and Performance Expectations 149
whether they are dened as internal predisposing states of the individual, elements
of social structure, or as embedded in interaction dynamics. Mood, emotion, and
sentiment may be distinguished by the intensity of experience and its duration for
the actor. Mood is generally treated as a low intensity, short duration state for the
actor and has limited organizing effect on social interaction.
3
Emotions are intense
and transitory states signied by intense behavior of actors. They often occur
in response to experienced situations or acts of others and are thus instantiated
in ongoing interaction. The number of emotions and the appropriate catalog of
foundational as opposed to derivative emotions is a matter of some debate (cf.
Turner, 2000; Kemper, 1981, 1987). Their intensity and the interaction associated
with their production provide the foundation for state organization processes.
Sentiments are distinguished from emotions by both intensity and duration.
Sentiments are generally conceptualized as of lesser intensity than emotions, but
of longer duration. Their enduring quality leads to structurally dened properties
of similarity and difference between actors (Shelly, 2001). These differences
may then acquire status value. The task in this chapter is to detail one means by
which this may occur.
Boundary conditions for the study of interaction in groups in the expectation
states research program are well established. Investigations are conditioned for
task groups that share a collective orientation toward a valued task (Wagner &
Berger, 1993). Investigations involvingstatus processes clearlyrequire a taskfocus,
though collective orientation may be less important than task focus. Investigations
focusing on emotions might require the reverse of this pattern with collective
orientation more important than a task focus. This is certainly a possible area of
investigation. For the purposes of this chapter, task groups are the focus of theory
building efforts. Other types of groups in which emotions are generated are beyond
the scope of the present inquiry, though they are important social situations and
deserve investigation.
4
A substantial concern for this research is that structures are dened within pairs
of actors in which members of the pair hold reciprocal, legitimated beliefs about
the relative competence of one another. For instance, one member of the pair may
be thought of as advantaged relative to the other with this knowledge shared by both
parties. In addition, both parties accept this state of affairs as fair, reasonable, and
appropriate if it is seen as legitimate by both parties. Interaction in larger groups
is based on the successive pairing of individuals in a group setting. For instance,
A may interact with B, then with C, while B is a bystander.
This dyadic approach is insufcient for dening structures when analyzing affect
processes, however. Whether emotions or sentiments are the focus of attention,
complementary states for self and other may not be assumed. Culturally, we do
150 ROBERT K. SHELLY
not agree that both of us share the knowledge that one of us likes the other while
the other is disliked. Interactions in such situations are unstable and likely to lead
to conict, i.e. they are not socially legitimate in American culture.
Cultural beliefs determine the extent to which emotions and sentiments are
reciprocated and hence may dene structures. At the present time in American
society, beliefs specify reciprocal treatment of emotions and sentiments. If an
actor says, I like you or I hate you the legitimate cultural norm is for the
other party to reciprocate these feelings. The culturally possible structure is an
equivalence relationship of mutual admiration or mutual distaste. The differential
expression of admiration and regard analogous to differentiated dyadic interaction
is obviated by this pattern of emotionally dened reciprocity.
The solution to this problemis to recognize that effective theoretical treatment of
affective patterns is possible only in larger groups. Hence, the pair wise treatment
of interaction commonly pursued in the expectation states theory programmust be
revised when working with affect patterns in groups. The simplest extension is to
consider triads as the basic unit of analysis rather than the dyad. Thus, it is possible
to dene sentiment structures comparatively when the focal actor has positive
regard for one other actor and negative regard for a second other. Presumably, both
of the other parties recognize this state of affairs and see it as legitimate.
These structures may then be treated as different social settings. Some of
the settings are stable, while others are unstable. Reciprocated positive regard
among all three actors in such a group is regarded as stable, though under some
special conditions such as love triangles such a group may be inherently unstable.
Reciprocated positive regard between one of the three pairs is also considered a
stable structure. A simple rule of thumb is that even numbers of negative pairings
(0, 2, 4, etc.) produce stable social structures, while odd numbers of negative
pairings (1, 3, 5, etc.) produce unstable social structures. The foundational
principles capturing social structural features dened by affective ties and how
they change have been well understood for over a half century as Heiderian
patterns of balance and imbalance (Heider, 1958).
To summarize, emotions, sentiments, and expectations are treated as analogs
of one another. They reect internal states of actor communicated to others. The
communication of emotions and sentiments takes place with expressive cues that
are both indicators of these internal states of self and reect an appraisal of the
behavior of the others in the situation. Our principal interest is the expression
of transitory emotions and their translation by actors to durable sentiments. The
minimal social unit is a triad rather than the dyad associated with pure status
processes.
Next I briey describe behavior status theory and briey assess its evidentiary
status.
Emotions, Sentiments, and Performance Expectations 151
BEHAVIOR STATUS THEORY
Explaining how inequality emerges in groups of social equals is a fundamental
problemfor sociologists. Bales (1950) rst addressed this issue in the study of task
groups by attempting to identify how actors responded to the task contributions
of others with both task and emotional activity (Bales & Slater, 1955). These
early efforts provide the foundation for the development of behavior status theory.
Berger (Berger, 1958) developed the foundation of the theory by identifying a
basic unit sequence that captured patterns of interaction in task groups and related
differences in the behavior of individual actors to the observable power and
prestige order of the group (who talked the most, whose ideas were most often
adopted, and who was regarded by others as the most inuential). Subsequent
elaborations by Berger and Conner (1969, 1974a, b) identied inuence elements,
accepted performances, and the importance of component inequality. Component
inequality recognizes the fact that actors who are initial status equals may possess
differential qualities on some components of their personality. For instance,
some individuals react faster to verbal prompts than do others. This difference
is sufcient to entail formation of performance expectations based on behavioral
inequalities triggered by the shorter reaction times.
The theory is based on two principles. The rst identies a fundamental interac-
tion sequence which describes how actors in groups sequence their behavior with
one another. This sequence begins with an action opportunity that may be a direct
behavioral request for action such as a question or expressive cue such as a glance
or head nod. This is followed by a performance that is either a contribution to solv-
ing the task or an effort to provide a suggestion on howto organize the group or task
activity. A second action opportunity may follow the performance. This second
opportunity may invite an evaluation of the performance. These evaluations may
be either positive or negative. The sequence is repeated until the task is solved.
The sequence may be truncated when actors do not include action opportunities
or evaluations at various points. Turn taking norms apparently contribute to these
truncated sequences in some groups (Shelly & Troyer, 2001a, b; Gibson, 2003).
The second principle describes howinformation contained in performance eval-
uations is transformed into expectations. The process is both subtle and robust.
While we do not have precise information about how this occurs and how many
positive or negative evaluations are necessary for an actor to form high of low ex-
pectations for self and other, we do have substantial reason to believe expectations
are a product of a contingent processing of evaluative information. This processing
is based on prior experience of actors in the situation as well as current experience,
i.e. current assessment of information about how others see you is contingent on
prior interaction experience with these others (Berger & Conner, 1969, 1974a, b).
152 ROBERT K. SHELLY
The theory species that receipt of positive evaluations leads to high expec-
tations for self; it also species that the receipt of negative evaluations leads to
low expectations for self. Positive evaluations of the contributions of the other(s)
in the group lead to high expectations for the other; evaluating the contributions
of the other negatively leads to low evaluations for the other. It is important to
recognize that evaluations of actions of particular others may be carried out by
a third party who thus indirectly acquires information about others relative skills
and abilities. Expectation patterns for self and other(s) formed by combinations of
these evaluation patterns for self and other organize interaction observed in task
groups in systematic ways . Referred to as Behavior Interchange Patterns (BIPS)
(Fisek et al., 1991), these interchange sequences are accepted by actors as appro-
priate patterns of interaction in the group. Fisek, et al. argue that BIPS lead to the
formation of performance expectations since they reect accepted rights to speak
and contribute to the group task.
High expectation for self and other, or low expectations for both, lead
to interaction patterns of approximate equivalence in levels of participation.
Undifferentiated expectation levels lead to undifferentiated inuence as well.
Differential expectations for self and other lead to higher levels of participation for
advantaged actors and lower levels for disadvantaged actors. Similarly, inuence
levels differentiated by expectations with advantaged actors more inuential than
disadvantaged actors. This process is hypothesized for individual acts (Berger &
Conner, 1969, 1974a, b) and for clusters of acts which are linked performance and
evaluation sequences recognized by actors as legitimate patterns of interaction
within the situation (Fisek et al., 1991).
The expression of evaluations may be direct and focused specically on the
task suggestion made by the previous actor, or it may be more subtle. Subtle
communication of evaluations may include body language or facial expressions
as well as simply ignoring the contribution. Another evaluative communication
may be a simple alternative suggestion employing facts or ideas expressed by
the rst actor, but recombined or altered by the new speaker. Such an alteration
may constitute rejection of the contribution, hence a negative evaluation, of the rst
speakers contribution without the direct confrontation required by saying, I think
that is a bad idea. These nonverbal and non-confrontational evaluations have the
qualities associated with expressive cues: actions given off in normal interaction
that do not directly provide status information about the relative standing of actors
in the setting. As such, they provide valuable information to actors in a situation,
but often in ways that are subtle and below the surface.
Several consequences of this formulation may be investigated. For instance, the
sequence of action opportunity, performance, and evaluation implies a structure
of correlation between behaviors initiated and behaviors received by actors in the
course of a task solving session. This has been supported by research reported
Emotions, Sentiments, and Performance Expectations 153
in Shelly (1997). Correlations between action opportunities received and perfor-
mances initiated exceed chance levels as do correlations between performances
initiated and evaluations received. The interaction sequence conceptualization also
receives support as the correlations between activity initiated and activity received
are positive and exceed chance.
The process of expectation formation hypothesized by the theory is supported
by data reported in Fisek and Ofshe (1970) and Shelly and Troyer (2001a, b) for
three person groups. These investigators employ different measures of interaction
activity. Fisek and Ofshe report results that employ social acts as the unit of analy-
sis. Shelly and Troyer employ duration of time talking as their dependent measure.
In both studies, members of three person groups differentiate themselves into a
rank order of participation and recognize the highest ranked participator as the
most inuential individual. Similar results are reported by Skvoretz et al. (1999)
for four person groups.
Of particular note is that Shelly and Troyer (2001b) report data for several
conditions of an experiment which varied the extent to which group members
were advantaged or disadvantaged with respect to one another. In one condition
one member had much greater task skill than did the others. In another condition
members began the session as equals on task skill. The evolution of inequality
occurs at approximately the same rate in these two distinctly different sets of initial
conditions. The communication of evaluations by means of explicit statements of
agreement and disagreement as well as by subtle expressive cues is apparently as
effective an organizer of observable power and prestige activity as the indicative
cues that inform members of groups in which one of them has a distinct skill
advantage.
Behavior status theory is supported by results from various studies. The
conceptualization of interaction sequences postulated by the theory has been
supported. Its hypotheses about how evaluations lead to performance expectations
and thence to the observable power and prestige order have been substantiated. In
addition, there is some evidence that for some groups, evaluations communicated
by expressive cues, are as effective as state organizing information as indicative
task cues. The application of these principles to the transformation of transitory
emotions to enduring sentiments is the next step in linking these emotion and
sentiment processes to performance expectations.
EMOTIONS AND SENTIMENTS
A theory of how transient emotional expressions are transformed to durable
sentiments and then manifested in performance expectations requires the in-
clusion of principles identied above. First, we must recognize a limited set of
154 ROBERT K. SHELLY
emotional expressions so that a relatively simple transformation structure may be
identied. A mechanism that species how actors process information is the next
requirement. This is followed by specication of ways of translating emotional
information to behavior so that interaction between two or more actors is possible.
Finally, sequences of interaction must be linked to the formation of sentiment
structures. Once sentiment structures are specied transformation to expectation
structures requires linking sentiments to qualities and capacities of actors in
various structural positions. These qualities and capacities are the foundation of
performance expectations.
Identifying a small list of emotions to employ in our theory requires some
careful thought. Several investigators have considered the question of how many
emotions are recognized by humans. For instance, Fehr and Russell (1984) asked
respondents to list emotions they recognized and produced a list of more than
one hundred sixty different terms used to describe internal feeling states. The
list has high levels of agreement on only four items. Sixty to seventy-ve percent
of respondents list love, sadness, anger, and happiness as part of the emotion
repertoire. All other emotion concepts were nominated by less than 50% of
respondents. Analytic attempts to develop catalogs of emotions have met with
mixed results with respect to parsimony (cf. Izard, 1992a, b; Plutchik, 1980). These
approaches often identify eight to ten emotions as fundamental with others treated
as derivatives of smaller sets. Both approaches present substantial problems for
our purposes. Any contingent branching posited as a consequence of emotional
sequences in interaction quickly leads to a geometric proliferation of emotional
paths as the number of emotions expands.
A more parsimonious approach is to identify a small number of basic emotions,
posit the development of derivative emotions from this short list to analytically
recognize the shadings and combinations of feeling states reported by actors in
their every day experience, and employ this abbreviated list in our analysis of in-
teraction. Kemper (1987) and Turner (1999a, b) favor this approach. They suggest
a basic emotion set consisting of satisfaction or happiness, fear or aversion, anger
or assertion, and disappointment, sadness or depression. Even four fundamental
emotions lead to large numbers of possible branches based on emotionally pat-
terned interaction
5.
A smaller number yet is necessary for our theoretical purpose.
The common element of each of the four emotions suggests a positive and negative
reaction by the actor as an arousal based on information from the environment.
For the purpose of analyzing how emotions are transformed to sentiments, the
use of the simple positive/negative polarity of emotional experience will be
employed. Hence, an actor may experience either a positive or negative reaction
to information presented by another individual as they interact with one another.
The number of steps is also limited, so that the actor may make transitions from
Emotions, Sentiments, and Performance Expectations 155
positive to negative state with respect to the other, but does not have to keep track
of long strings of experiences.
Two approaches are available to analyze how information given off by an actor
is processed emotionally. One involves the explication of social messages at the
structural and biochemical level of the brain (LeDoux, 1996). The other is based on
the social responses elicitedininteractionandis similar toindicative andexpressive
cues explicated in the discussion of expectations states theory (Turner, 2000).
Development of a detailed theory of howthese processes function in social contexts
is beyond the scope of the present endeavor.
Emotional Parsimony Principle: The basic number of emotions is small
in number; basic emotions may be combined to produce large numbers of
differentiated feeling states for actors. The fundamental dimension capturing
emotions is by a positive/negative continuum.
It is sufcient for our purposes to posit that positive and negative emotions are
communicated between actors via expressive or indicative cues. These cues may
be categorical in the sense that they place actors in socially unranked groupings or
they may be direct emotional assessments of a behavior or preceding sequence of
behavior. For instance, a husband and wife discussing an evenings entertainment
options may employ positive and negative cues that reect preferences to identify
the partners preferred entertainment options. They may employ expressive cues
such as smiles or frowns, nods or head shakes, eye contact or aversion, or private,
nonverbal codes to communicate positive or negative emotional reactions to the
actions of the other that do not categorize the interaction partner. Similarly, the
use of indicative cues place interaction partners in particular social categories
with direct statements that assign the other to the sports nut or chick ick
categories of desired entertainment.
Statements made byanactor withemotional content serve toplace the interaction
partner in the us or them category depending upon whether the emotional
reaction to the partner is positive or negative. Indicative cues may communicate
emotions directly with statements such as I am angry with you, or I am pleased
with you (cf. Lovaglia & Houser, 1996). Emotional actions and reactions are
employed in interaction sequences instead of evaluative comments and nonverbal
cues characteristic of indicative and expressive evaluations.
Emotional Cues Principle: Expressive and indicative cues communicate
emotional information from one actor to another. These cues place an actor in a
particular social category or communicate discrete emotion based information.
Aggregation of communicated cue information to form an enduring affective
assessment of an other in an interaction sequence is based on principles similar
156 ROBERT K. SHELLY
to those of behavior status theory. Specically, an actor accumulates information
about how the other has responded to behaviors given off by self during an
interaction sequence. Positive responses such as satisfaction or happiness lead
to formation of long lasting relations such as liking or love for the other party.
Negative responses by the other such as anger, disappointment, and expressed
fear lead to enduring affective states such as disliking or hate for the other
party. This simple conditioning model is well known in the work of Zajonc and
his collaborators in psychology who study processes governing how emotional
reactions are associated with cognitive objects (Monahan et al., 2000; Murphy
& Zajonc, 1993; Murphy et al., 1995; Zajonc, 1998, 2001). Briey, repetitive
exposure to a source is sufcient to induce positive affect toward an object
associated with the source, but the effect can be interrupted or reversed with the
provision of negative information about the source. This principle is well known
in advertising and accounts for the proliferation of product endorsements by
celebrity spokespersons.
Foschi and Foschi (1976, 1979; Foschi, 1986) have developed a model of
expectation formation based on Bayesian decision principles that allows us to
rene our understanding of how actors form expectations from evaluation experi-
ences. The analog of this process is posited as the mechanism by which emotion
experiences are transformed to sentiments by actors. Their work shows that actors
process information about evaluations, and by extension, emotional cues, based
on an interaction history in which prior experience conditions the chance an actor
will develop an expectation for performance and produce a particular behavior
in the current interaction. According to this approach, an actor is more likely to
arrive at positive assessments of the other party in an interaction if interaction
experience with the other has been positive in the recent past. Similarly, an actor is
more likely to arrive at negative assessments of the other in an interaction if their
interaction experience has been negative in the recent past. These instantaneous
positive and negative assessments become the foundation for enduring sentiments.
The model has been tested and an important result for our purposes uncovered:
while rare in observed interactions, negative information provides distinct and
important information for members of groups (Foschi, 1986). It as if actors expect
to experience positive events (praise for their contributions, emotionally positive
reactions fromthe other). If the interaction is unpleasant (criticismof contributions,
or angry interactions with the other), two outcomes occur. First, actors attend to this
negative experience more than to the positive experiences they have had. Second,
they are more likely to employ this information in future interaction with the other.
The implication for the transformation of emotions to sentiment is that actors
positive expectations regarding emotions will govern the interaction sequences
under most circumstances. Negative emotional expressions are more consequential
Emotions, Sentiments, and Performance Expectations 157
in providing information because they stand out from the expected patterns of
interaction.
Application of these information processing principles in conjunction with
analogs of those of behavior status theory (cf. Berger & Conner, 1969,
1974a, b) lead to the transformation of transitory emotional information to en-
during sentiments. This transformation is based on the aggregate experiences of
actors with interaction partners. Positive experiences lead to liking for the partner
while negative experiences lead to disliking for the partner. Negative information,
because it contradicts cultural expectations about interaction partners, is more
powerful than positive information in its effects on the formation of enduring
sentiments.
Emotional Transformation Principle: The exchange of emotional information
between pairs of actors may be both positive and negative in valence. Aggregated
emotional information is transformed to enduring, reciprocated sentiments in the
same way performance expectations are based on repeated evaluations of task
contributions.
Once formed between pairs of actors, sentiments are sufcient to dene social
structures based on solidarity. The reciprocal commitments implied by these
pairings provide a foundation for more complex structures as larger groups
are composed of multiple pairings of actors, some of which are based on positive
relationships and others on negative relationships. The denition of structure
arrived at in this fashion does not entail performance expectations. The sufcient
condition for expectation formation is a socially valued distinction between
attributes of actors (e.g. gender, race, skill, ability).
Two possible mechanisms may be invoked to explain how emotions are
associated with performance expectations by way of sentiment relations between
self and others. One process that leads to the translation of sentiments and
sentiment structures to performance expectations and their subsequent expression
as power and prestige orders is based on interaction patterns biased by sentiment
preferences (Shelly, 1993; Shelly & Webster, 1997). This mechanism is similar
to the process captured by behavior status theory (Berger & Conner, 1969,
1974a, b). Briey, actors develop expectations for others based on experience in
an interaction sequence in which opportunities, performances, and evaluations
lead to expectation formation and subsequent observable power and prestige
order. In this approach, sentiment patterns bias interaction patterns so liked
others are given more chances to perform and are more likely to receive positive
evaluations in comparison to disliked others. This biasing of interaction by
sentiments that are in turn a result of interactions based on emotions provides a
mechanism by which emotions can lead to performance expectations. It implies
158 ROBERT K. SHELLY
Fig. 2. Emotions, Sentiment, Interaction Sequence.
that the mechanism linking emotions to performance expectations is a weak one
as it is mediated rst by the transformation of emotions to sentiments, and then
by the biased interaction present in groups organized by sentiments.
Figure 2 presents a conceptual model of how emotions may organize interac-
tion by structuring the expression of sentiments. The arrows represent positive
associations between the conceptual elements of the model. Emotional expression
is transformed, over time, to sentiments. These sentiments in turn organize the
interaction that takes place in the task group. Since the concepts are positively
associated with one another, a baseline linear model is assumed for purposes of
hypothesizing empirical relationships. The actual functional form may vary from
this, however.
The second mechanism by which sentiments may be linked to performance
expectations is based on a direct tie of sentiments to expectations. In this approach,
the construction of a sentiment structure is sufcient to create a set of linked infor-
mation about the qualities and capacities of self and others such that liked others
are assigned to high states of performance expectations and disliked others are
assigned to low states. This association of liking with performance expectations,
demonstrated by Shelly (2001), is similar to the concept of a valued personal
characteristic demonstrated Driskell (1982) and employed by Webster and Hysom
(1998) in status construction theory. Briey, valued personal characteristics are
attributes of an actor that are valuable in the social context, but they do not
have status value attached to them. Because of their value, they easily acquire
performance expectations when made salient through legitimated interaction.
For instance, honesty is a valued personal characteristic which may acquire
status value during interaction if an actor is perceived as being exceptionally
honest or dishonest.
Sentiment Transformed to Expectations Principle: Sentiments organize a struc-
ture of solidarity and alienation among actors in groups. These sentiment struc-
tures serve as foundations for biased interaction that leads to the formation of
performance expectations subsequently reected in power and prestige orders
in groups.
Emotions, Sentiments, and Performance Expectations 159
Fig. 3. Emotions/Sentiments/Personal Characteristic.
Figure 3 presents a conceptual model of the process by which emotions may be
transformed to sentiments and subsequently to information about valued personal
characteristics of the actor. Emotions are transformed to sentiments and sentiments
form the basis for the assignment of states of valued personal characteristics for
the actor. The links between these concepts are presumed to be positive and linear.
Hence, we expect that when activated, the valued personal characteristics will be
transformed to status characteristics. Whether, and under what conditions, they
acquire the importance of diffuse or specic status characteristics is an empirical
question of some interest. As with the earlier discussion of how sentiments or-
ganize interaction, the appropriate functional form of the relationship is open for
investigation.
Implications
Several implications of this analysis and suggestions for further studies may be de-
veloped fromthe above principles. First, empirical work that focuses on repetitions
of emotional sequences between actors, and the behaviors employed to commu-
nicate emotions from one actor to another are important situations to investigate.
Identifying a small set of emotions to examine in interaction will provide a founda-
tion for developing the transformation of transient emotion to enduring sentiments.
Many of the emotions identied by investigators who catalog emotions (cf. Fehr &
Russell, 1984) do not distinguish between emotion and sentiment. The distinction
is crucial to an understanding of how emotion contributes to the development
of interpersonal inequality. For the present purposes, I have suggested that it is
sufcient to differentiate positive and negative emotion states. This may prove
to be insufcient when tested empirically. It may be necessary to differentiate
intensity as well as direction of emotion in order to develop this research
appropriately.
Identifying cues is both a methodological and theoretical problem. Methodolog-
ically, we must be able to distinguish indicative from expressive cues, verbal from
nonverbal cues, and associate them with the appropriate emotion given and given
off. Theoretically, it will be important to be able to distinguish the relative impact
of emotional categorization as a consequence of the use of indicative cues from
emotional assignment based on expressive cues. Do these two different outcomes
of emotional expression have the same or different consequences for the actor and
other? If the consequences are the same, it may be appropriate to reduce the concept
set in this case to only one type of cue in the interests of theoretical parsimony. If
160 ROBERT K. SHELLY
the consequences are different, maintaining the distinction between indicative and
expressive cues preserves the analog with the expectation states theory tradition.
Research that explores mechanisms by which transient emotions are converted
to enduring sentiments is needed. We know it is possible to induce strong positive
and negative emotions by providing actors with positive or negative feedback
from partners, but the extent to which these reactions are enduring is not known
(Lovaglia & Houser, 1996).
Bakeman and Gottman (1997) have been able to develop techniques to study
the formation of enduring sentiments between married couples. These techniques
provide models for our purposes. In particular, dynamic analysis of positive and
negative messages, as both explicit, indicative cues and as implicit, expressive
cues, within dyads based on their techniques can be used to inform the patterns
of transformation of emotions to sentiments. Techniques of inducing emotional
reactions of varying levels of intensity will be needed, as well as calibration of
numbers of positive and negative emotion cues received from other to induce the
transformation.
Finally, identifying the process by which expectations emerge from emotion-
ally based features of interaction requires that investigators observe interaction
sequences that begin with emotional cues that may evolve into hierarchically dif-
ferentiated social groups. Identifying the particular empirically observable steps in
this sequence will provide answers toquestions about processes we donot presently
understand. For instance, we can posit that status evolution develops out of gener-
alization of the enduring sentiment patterns of a group through a process similar
to the process observed when hierarchies are imposed by investigators in groups.
Some evidence exists to support this claim (cf. Driskell & Webster, 1997). Two
fundamental questions are of concern in this area. Is the differentiation in task
groups organized by emotions an outgrowth of an interaction process governed
by legitimated rules of who talks and what is communicated? Alternatively, do
emotions and their durable counter parts, sentiments, create qualitatively different
valued personal characteristics similar to honesty, which are easily associated with
performance expectations? Developing experimental tests to answer these ques-
tions will be a particularly challenging problem as the processes are unobservable
and hypothesized to have similar empirical consequences.
CONCLUSION
I began this chapter with a question about the relationship between inequality
processes, with their differences in observable power and prestige orders, and
emotional processes with their observable integrative and disintegrative outcomes.
The extant literature of both areas has limited overlap. We know that one of the
Emotions, Sentiments, and Performance Expectations 161
processes can affect the observable behavior of the other. Positive emotions are
an outgrowth of elevated status positions in groups (Lucas & Lovaglia, 1998)
while emotions (Lovaglia & Houser, 1996) and sentiments (Driskell & Webster,
1997) affect observable power and prestige outcomes. Sentiments apparently
contribute to the hierarchical organization of interaction by biasing who receives
opportunities to talk and positive evaluations (Shelly, 1993; Shelly & Webster,
1997). A number of possible research questions may be developed with this
limited empirical guidance by simply expanding scope and initial conditions of
experimental tests.
By adopting some principles from the expectations states research program,
I develop a theoretical account of how emotions may become organizers of
hierarchical differentiation in groups. First, it is necessary to limit the instances
of emotions considered to a small number. The solution here is to work with only
two emotions: a positive one that is pleasurable for the actor, and a negative one
that is painful for the actor. The transformation of transient emotion experiences
of the actor into enduring social sentiments for interaction partners is the central
concern of the theory development effort. This transformation occurs as a result
of repeated exposure to interaction cues that communicate emotional information
between actors. Once the transformation has occurred, the biasing of interaction
that leads to hierarchy formation is likely to occur. This process is analogous
to the emergence of inequality in groups of individuals who are initially status
equals as explained by behavior status theory.
In addition to this process, it is also possible that in assigning other actors to en-
during sentiment positions, a person may also develop performance expectations
for these individuals. This association of performance expectations with senti-
ments has been reported by Shelly (2001). The data for this study are restricted to
a specic situation and are thus limited in scope, but they do suggest that actors
impute qualities and capacities, the information building blocks of performance
expectations, to others based on sentiments. Well-liked others are assigned
advantaged positions while disliked persons are assigned disadvantaged positions.
This outcome opens additional avenues for investigating the relationship between
emotion, sentiment, and status processes. An intriguing question for future
work is whether these expectations are direct consequences of sentiment ties, or
whether they are an outgrowth of a two step sequence of information processing in
which valued personal characteristics are rst formed and then become salient as
status characteristics.
NOTES
1. In addition to expectation states, identity creation (Johnston, 1988) and role behaviors
(Talley & Berger, 1983, cited in Berger, 1988) have been analyzed as state organizing
processes.
162 ROBERT K. SHELLY
2. There is substantial disagreement about how many emotions are innate in humans.
Exploring this disagreement would take us too far aeld. The interested reader is advised
to examine Kemper (1987) or Scheff (1990).
3. Moods as elements of social interaction are not analyzed in this chapter because of
their low level of intensity and brief duration.
4. I wish to thank Murray Webster for identifying this point and making the suggestion
regarding differential application of scope conditions across domains of group type.
5. The basic principle is n
x
where n is the number of emotions and x is the number of
interchanges in the interaction sequence.
ACKNOWLEDGMENTS
I wish to thank Ann Converse Shelly, Murray A. Webster, Jr., and Joseph Berger
for reading and commenting on drafts of this chapter as it evolved. Their support
and contributions are gratefully acknowledged.
REFERENCES
Bakeman, R., & Gottman, J. M. (1997). Observing interaction: An introduction to sequential analysis
(2nd ed.). New York: Cambridge University Press.
Bales, R. F. (1950). Interaction process analysis. Cambridge, MA: Addison-Wesley.
Bales, R. F., &Slater, P. E. (1955). Role differentiation in small decision-making groups. In: T. Parsons
& P. E. Slater (Eds), The Family, Socialization, and Interaction Processes (pp. 259306).
Glencoe, IL: Free Press.
Berger, J. (1958). Relations between performance, rewards, and action-opportunities in small groups.
Ph.D. dissertation, Harvard University.
Berger, J., & Conner, T. L. (1974a). Performance expectations and behavior in small groups. Acta
Sociologica, 12, 186198.
Berger, J., & Conner, T. L. (1974b). Performance expectations and behavior in small groups: A revised
formulation. In: J. Berger, T. L. Conner & M. H. Fisek (Eds), Expectation States Theory: A
Theoretical Research Program (pp. 85109). Cambridge, MA: Winthrop.
Berger, J., Norman, R. Z., Balkwell, J. W., &Smith, R. F. (1992). Status inconsistency in task situations:
A test of four status processing principles. American Sociological Review, 57, 843855.
Berger, J., Webster, M., Jr., Ridgeway, C., & Rosenholtz, S. J. (1986). Status cues, expectations, and
behavior. In: E. J. Lawler (Ed.), Advances in Group Processes (Vol. 3, pp. 122). Greenwich,
CT: JAI Press.
Bianchi, A. (2000). Examining sentiments effect on status: A test between the constitutive and
mediator models. Paper presented at the Annual Meeting of the American Sociological
Association, Washington, DC.
Burke, P. J. (1967). The development of task and social-emotional role differentiation. Sociometry,
30, 379392.
Burke, P. J. (1968). Role differentiation and the legitimation of task activity. Sociometry, 31, 404411.
Burke, P. J. (1971). Task and social-emotional leadership role performance. Sociometry, 34, 2240.
Emotions, Sentiments, and Performance Expectations 163
Dovidio, J. F., & Ellyson, S. L. (1982). Decoding visual dominance: Attributions of power based
on percentages of looking while speaking and looking while listening. Social Psychology
Quarterly, 45, 106113.
Dovidio, J. F., Brown, C. F., Heltman, K., Ellyson, S. L. et al. (1988). Power displays between men
and women in discussions of gender-linked tasks: A multichannel study. Journal of Personality
and Social Psychology, 55, 580587.
Driskell, J. E. (1982). Personal characteristics and performance expectations. Social Psychology
Quarterly, 45, 229237.
Driskell, J. E., & Webster, M. A. (1997). Status and sentiment in task groups. In: J. Smatzka,
J. Skvoretz & J. Berger (Eds), Status, Network, and Structure: Theory Development in Group
Processes (pp. 179221). Stanford, CA: Stanford University Press.
Fararo, T. J., &Skvo retz, J. (1986). E-state structuralism: Atheoretical method. American Sociological
Review, 51, 591602.
Fehr, B., & Russell, J. A. (1984). Concept of emotion viewed from a prototype perspective. Journal
of Experimental Psychology, 113, 464486.
Fisek, M. H., & Berger, J. (1998). Sentiment and task performance expectations. In: J. Skvoretz &
J. Szmatka (Eds), Advances in Group Processes (Vol. 10, pp. 2339). Stamford, CT: JAI Press.
Fisek, M. H., Berger, J., & Norman, R. Z. (1991). Participation in heterogeneous and homogeneous
groups: A theoretical integration. American Journal of Sociology, 97, 114142.
Foschi, M., & Foschi, R. (1976). Evaluations and expectations: A Bayesian model. Journal of
Mathematical Sociology, 4, 279293.
Foschi, M., & Foschi, R. (1979). A Bayesian model for performance expectations: A Bayesian model.
Social Psychology Quarterly, 42, 232241.
Gibson, D. R. (2003). Participation shifts: Order and differentiation in group conversation. Social
Forces, 81, 13351381.
Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.
Heise, D. R., & Calhan, C. (1995). Emotion norms in interpersonal events. Social Psychology
Quarterly, 58, 223240.
Heise, D. R., & Thomas, L. (1989). Predicting impressions created by combinations of emotion and
social identity. Social Psychology Quarterly, 52, 141148.
Hochscild, A. (1983). The managed heart: The commercialization of human feeling. Berkeley, CA:
University of California Press.
Izard, C. (1992a). Basic emotions, relations among emotions, and emotion-cognition relations.
Psychological Review, 99, 561565.
Izard, C. (1992b). Four systems for emotion activation: Cognitive and noncognitive. Psychological
Review, 100, 6890.
Johnston, J. R. (1988). The structure of ex-spousal relations: An exercise in theoretical integration and
application. In: M. Foschi & M. A. Webster, Jr. (Eds), Status Generalization: New Theory and
Research (pp. 309336). Stanford, CA: Stanford University Press.
Kemper, T. D. (1981). Social constructionist and positivistic approaches to the study of emotions.
American Journal of Sociology, 87, 336362.
Kemper, T. D. (1987). How many emotions are there? Wedding the social and autonomic components.
American Journal of Sociology, 93, 263289.
LeDoux, L. (1996). The emotional brain: The mysterious underpinnings of emotional life. New York:
Simon & Schuster.
Lovaglia, M. J., & Houser, J. A. (1996). Emotional reactions and status in groups. American
Sociological Review, 61, 867883.
164 ROBERT K. SHELLY
Lucas, J. W., & Lovaglia, M. J. (1998). Leadership status, gender, group size, and emotion in
face-to-face groups. Sociological Perspectives, 41, 617637.
Monahan, J. L., Murphy, S. T., & Zajonc, R. B. (2000). Subliminal mere exposure: Specic, general,
and diffuse effects. Psychological Science, 11, 462466.
Morgan, R. L., & Heise, D. (1988). Structure of emotions. Social Psychology Quarterly, 51, 1931.
Murphy, S. T., Monahan, J. L., & Zajonc, R. B. (1995). Additivity of nonconscious affect: Combined
effects of priming and exposure. Journal of Personality and Social Psychology, 69, 580602.
Murphy, S. T., &Zajonc, R. B. (1993). Affect, cognition, and awareness: Affective priming with optimal
and suboptimal stimulus exposures. Journal of Personality and Social Psychology, 64, 723739.
Plutchik, R. (1980). Emotion: A psychoevolutionary synthesis. New York: Harper & Row.
Ridgeway, C. (1991). The social construction of status value: Gender and other nominal characteristics.
Social Forces, 70, 367386.
Ridgeway, C., & Johnson, C. (1990). What is the relationship between socioemotional behavior and
status in task groups. American Journal of Sociology, 95, 11891212.
Slater, P. E. (1955). Role differentiation in small groups. American Sociological Review, 20, 300310.
Scheff, T. J. (1990). Microsociology: Discourse and social structure. Chicago, IL: University of
Chicago Press.
Shelly, R. K. (1997). Sequences and cycles in social interaction. Small Group Research, 28, 333356.
Shelly, R. K. (2001). How performance expectations arise from sentiments. Social Psychology
Quarterly, 64, 7287.
Shelly, R. K., Handley, I. M., Baer, J., & Watson, S. (2001). Groups and affect: Sentiments, emotions,
and performance expectations. Current Research in Social Psychology, 6, 135150.
Shelly, R. K., & Troyer, L. (2001a). Speech duration and dependencies in initially structured and
unstructured task groups. Sociological Perspective, 44, 419444.
Shelly, R. K., & Troyer, L. (2001b). Emergence and completion of structure in partially dened
groups. Social Psychology Quarterly, 64, 7287.
Skvoretz, J., Webster, M., Jr., & Whitmeyer, J. (1999). Status orders in task discussion groups. In:
S. R. Thye, E. J. Lawler, M. W. Macy & H. Walker (Eds), Advances in Group Processes
(Vol. 16, pp. 199218). Stamford, CT: JAI Press.
Smith-Lovin, L. (1995). The sociology of affect and emotion. In: K. S. Cook, G. A. Fine & J. S. House
(Eds), Sociological Perspectives on Social Psychology (pp. 118148). Needham Heights, MA:
Allyn & Bacon.
Talley, J., & Berger, J. (1983). Social control as a state organizing process. Unpublished Manuscript.
Stanford University.
Thomas, L., & Heise, D. R. (1995). Mining error variance and hitting pay-dirt: Discovering systematic
variation in social sentiments. Sociological Quarterly, 36, 425439.
Tiedens, L. Z., Ellsworth, P. C., & Mesquita, B. (2000). Stereotypes about sentiments and status:
Emotional expectations for high- and low-status group members. Personality and Social
Psychology Bulletin, 26, 560574.
Turner, J. (1999a). Toward a general sociological theory of emotions. Journal for the Theory of Social
Behavior, 29, 133162.
Turner, J. (1999b). The neurology of emotions: Implications for sociological theories. In: D. Franks
(Ed.), The Sociology of Emotions (pp. ??). Greenwich, CT: JAI Press.
Turner, J. (2000). On the origins of human emotions. Stanford, CA: Stanford University Press.
Wagner, D. G., & Berger, J. (1993). Status characteristics theory: The growth of a program. In:
J. Berger & M. Zelditch, Jr. (Eds), Theoretical Research Programs: Studies in Theory Growth
(pp. 2563). Stanford, CA: Stanford University Press.
Emotions, Sentiments, and Performance Expectations 165
Webster, M., & Hysom, S. J. (1998). Creating status characteristics. American Sociological Review,
63, 351378.
Zajonc, R. B. (1980). Feeling and thinking: Preferences need no inferences. American Psychologist,
35, 151175.
Zajonc, R. B. (1998). Emotions. In: D. Gilbert, S. T. Fiske & G. Lindzey (Eds), Handbook of Social
Psychology (4th ed., Vol. 1, pp. 591632). New York: McGraw-Hill.
Zajonc, R. B. (2001). Mere exposure: A gateway to the subliminal. Current Direction in Psychological
Science, 10, 224228.
THE ENHANCEMENT IMPERATIVE
AND GROUP DYNAMICS IN THE
EMERGENCE OF RELIGION AND
ASCRIPTIVE INEQUALITY
Michael Hammond
ABSTRACT
In order to deepen our understanding of contemporary social structures, we
must often trace their distant origins in our evolutionary past. The origins
of two structures are analyzed here. Both religion and collective ascription
are shaped in part by a common imperative to access rewarding emotional
arousal release protected by a special set of arouser screening rules. Only
certain enhanced arousers with an attractive ratio of contrast values to
access costs can regularly tap these emotional reservoirs. In these two cases,
it is the larger social group context that must supply the enhancements.
The result is that some group processes are marked by emotional dynamics
deeply rooted in the pursuit of these extraordinary arousers.
INTRODUCTION
In an earlier volume of this series, Satoshi Kanezawa (2001) emphasized that
most of the powerful theoretical models in sociology, such as structuralism and
its extension, social network theory, have little to say about the origins of the
Theory and Research on Human Emotions
Advances in Group Processes, Volume 21, 167188
2004 Published by Elsevier Ltd.
ISSN: 0882-6145/doi:10.1016/S0882-6145(04)21007-4
167
168 MICHAEL HAMMOND
structures they analyze. He argued that an understanding of the evolutionary
emergence of our species can sometimes add a crucial perspective on the origin
question and therefore sharpen our analysis of contemporary social structures. In
this article, the rewarding emotional release dynamics of two macro structures,
religion and ascriptive inequality, will be considered from an evolutionary per-
spective. It will be argued that these structures can be seen as group constructions
offering a very special arouser package unavailable at a more micro level, such
as direct personal ties or close kin networks. The key to this emotional release
is enhanced arousers offering an increase in attractive contrasts without a parallel
increase in access costs. It is imperative that some part of group interaction offer
these special arousers. Only a small range of social structures can offer such
enhancements, and these creations played a key role both in our evolutionary
emergence and in our later history.
For certain human interests, preconscious emotional arousal release schedules
are set such that only group constructions can provide the special stimuli necessary
to trigger rewarding arousal at a reasonable access cost (Hammond, 2001, 2003).
Humans always seek some such enhanced arousers in their wider social life,
and this pursuit gives human groups a very distinct quality in comparison to
other species in our family tree. The article begins with a consideration of this
enhancement imperative model and religion. It concludes with an application of
the same model to the analysis of ascriptive inequality. Both mark all of human
history, and both have part of their selective origins in their role as vehicles to
expand the web of reliable networks and social cooperation in humans. Much of
this article focuses on our early human history, but as Douglas Massey stressed
in his presidential address to the American Sociological Association (2002), it
is crucial that sociologists studying modern history ground their analysis in our
most distant past and in the possible impact of that past on the present.
RELIGION AND EXPANDING NETWORKS
The revolutionary technological advances in our tools to investigate the workings
of our brain are raising a number of interesting sociological questions. For in-
stance, neurophysiological studies of the arousal release associated with religious
arousers demonstrate that there does not appear to be any separate reward wiring
for such stimuli (Hammond, 2003). Instead, religion seems to piggyback on other
reward systems, especially in regard to rewarding emotional release associated
in one way or another with personal attachments (Joseph, 1996; Newberg et al.,
2001, pp. 4, 4245; Saver & Rubin, 1997). What is striking sociologically about
this piggybacking is that it usually creates little confusion for individuals in regard
Emergence of Religion and Ascriptive Inequality 169
to mixing personal ties and religious attachments using much the same reward
systems. Why are there not more difculties in such a dual use? If the same
reward systems are being tapped, why dont we nd individuals with no religious
commitments consistently having more extended networks of emotionally moving
personal ties? After all, at rst glance, it appears logical that networks should
be larger when arousal reward systems are not being tapped by two different
arouser packages. And vice versa, why dont we nd more individuals with heavy
religious bonds having far fewer close personal ties due to the heavy utilization
of arousal reservoirs by religion? Empirically, it does not appear as if the two
pursuits get in each others way. Individuals who are highly religious do not
have on the average fewer or less intense personal ties than others without such
commitments, and non-believers do not have on the average a greater number or
more intense personal ties (see for instance, Fischer, 1982, pp. 208214). This
is only possible if the shared arousal release is screened or buffered in some
manner. As we shall see, the dynamics of this regulation point to the indirect
contributions made by social constructions like religion to shaping the formation
of human groups.
As Maryanski and Turner (1992) stress, one of the key necessities in the origin
of our species was to expand network size and reliability in comparison to the
patterns of our closest evolutionary relatives and to the presumed pattern of our
last common ancestor. Without this expansion, humans would be conned to
much the same range of ecological circumstances, and this connement would
in turn limit much of the reproductive success of our species. The enlargement
of our cognitive capabilities in language and communication provided one basis
for network expansion. The same appears to be true for the enlargement of our
emotional release capacities. Turner (2000, p. 92) has demonstrated both trends
in terms of comparative evidence for the relative size of brain structures in great
apes such as the gorilla and the chimpanzee, humans, and a primitive mammal.
The neocortex is associated with cognitive skills, and shows the greatest increase
in relative size. There are also lesser but still substantial increases in subcortical
areas associated with emotional release, such the diencephalon, the amygdala, the
septum, and the hippocampus. These increases suggest that humans had greater
affective capacities to apply to their social interaction.
With additional cognitive and affective capabilities, individuals could have a
greater range of ties and cooperation. These ties could have a greater or lesser
degree of reliability based in part on the rewarding arousal release in such bonds.
Other valuable social resources could ow along the connections established
by this emotional loading. In our evolutionary context of origin, the selective
advantage was not in xing the precise number or direction of these linkages.
Affective loading was ideal in promoting the reliability of bonds with a minimum
170 MICHAEL HAMMOND
sacrice in exibility. This is part allowed humans to exploit a greater range
of ecological niches. For instance, humans could move into areas where only
highly cooperative forms of food provisioning through hunting and gathering
could support a population. This economy required networks of male-male,
female-female, and male-female cooperation and supportive bonds that were not
possible among our evolutionary relatives.
Such networks could be expanded on a direct personal basis, with additional
resources put into additional social bonds on a one-to-one basis. This occurred
to some extent, but it is a very expensive way to expand social ties. Direct ties
can be costly in terms of energy and time demands. However, there is another
means to expand ties in a less costly manner per additional tie created. Ties can
also be formed indirectly through a common bond to a third arouser package.
The creation of this third point has some costs, but it can be the basis for a
vastly expanded network in comparison to what is possible with a primary
reliance on direct personal ties. How to get humans to be interested in creating
these bridges to extended networks? They could be hard-wired, but this would
sacrice exibility and would actually require more adaptive changes than a more
economical alternative in terms of how many selective changes would have to
occur. The key to this alternative is to use arouser screening rules to send humans
in the pursuit of stimuli that only larger social groups can provide on a cost
effective basis.
The human brain has a wide range of stimulus appraisal mechanisms, both
conscious and preconscious (Hammond, 2001, 2002; LeDoux, 1996). These
mechanisms can alter the response to arousers based on a number of criteria. They
play a central role in converting an arouser into arousal release. The most well
known of these preconscious screening rules is habituation (Koukouas & Over,
1993; Meuwissen & Over, 1990). In the face of the repetition of even the most
attractive stimuli, arousal release eventually begins to fade. The same arouser
produces less and less rewarding arousal. The erosion of rewarding arousal
produces a reservoir ready for use if another attractive arouser becomes available.
There is a signicant preconscious factor in this screening pattern. Even a species
with our awesome cognitive capacities does not have the means to be directly
aware of these screening rules and does not have the means to consciously prevent
such arouser screening from occurring.
The preconscious arouser screening that is the most important sociologically
involves responses to a variety of attractive stimuli for any particular interest
(Hammond, 1999). In response to multiple arousers, there is a point at which it
takes stronger and stronger contrast values to trigger the same emotional arousal
release produced before by arousers with lesser attractive contrasts. At some
further point, even the strongest of arousers only produces a reduced arousal
Emergence of Religion and Ascriptive Inequality 171
release; and eventually, no stimuli will produce anything signicant in the way
of rewarding release. These rules for converting arousers into arousal are at the
root of diminishing marginal utility. In this case, utility is measured in terms of
emotional release rewards. If natural selection favors individuals who pursue only
a limited variety of arousers in regard to an interest, the depreciation of the arousal
impact of additional stimuli is severe and unstoppable. If there is a selective
advantage to individuals pursuing more extended variety, the rates of discounting
are eased. To limit interest, the principle is to accelerate preconscious arousal
release dampening responses. To extend interest, ease off on the depreciation.
The second twist on the use of arousal release rules to preconsciously shape
behavior concerns the relationship between arousal release and costs an individual
will be likely to bear in accessing such release. If extremely heavy rewarding
arousal is given for the rst set of appropriate arousers, individuals will be ready
to pay heavy costs for those stimuli. Calculations of effort, time, and risk are often
minimal if the alternative is to have no access to even the most basic arousers
for an interest. As more arousers become regularly available, the costs that will
be paid for yet more variety begin to decline. In the face of preconscious arouser
screening, the calculation of access costs can be very rational, but rationality
cannot play the same role in certain aspects of the control of arousal release.
For instance, a food interest is a common example of this pattern (Carlson, 1995,
pp. 338351; Rolls et al., 1986). Deny food to an individual for a long enough
time, and they will pay virtually any cost to get even the most basic food arousers.
They will also receive extremely heavy rewarding arousal for those stimuli.
However, if food is regularly available, then individuals will be unwilling normally
to pay great costs for some additional food. They might nd the additional food
attractive in terms of the contrasts to the food they already have, but they are likely
to pursue those additions only if the price is right. If the access costs are as great
as the attractive arouser contrasts, then most individuals will not bother to extend
their food interest. That is, there appears a point in interest extension at which a
proportionate relationship between additional arouser contrasts and access costs is
no longer motivating. Interest can be extended only with a disproportionate rela-
tionship in which more attractive contrasts are available without a parallel increase
in access costs. Finally, there is a point at which even a disproportionate ratio
of attractive contrasts to access costs cannot produce additional arousal release
to fuel more interest.
The special arouser packages with additional attractive contrasts without a
parallel increase in access costs are called enhancements (Hammond, 2001,
2003). In many circumstances, if enhancements are available, interest can be
extended. If unavailable, emotionally rewarding release is dampened, and before
too long, individuals look to other interests. Different settings for preconscious
172 MICHAEL HAMMOND
arouser screening shape when these points of extension occur. Enhancements can
be required early on, or they can become necessary only later in an interest curve.
Natural selection can work on synaptic alteration in the electrochemical transmis-
sion of neural messages, as well as on hormone and other neurochemical release,
to shape the general rules for converting arousers into rewarding arousal (Carlson,
1995). In many species, the role of enhanced arousers is comparatively small.
They are an appealing option, but not a necessity. In the human case, the situation
is quite different. Enhancements are an imperative. We all require at least one,
and preferably more, of these special packages. The origin of this imperative is
related to the evolutionary economics of natural selection. In at least two cases in
our context of origin, better results could be achieved with fewer adaptive changes
by using this special arouser package. To understand this logic, we must consider
in terms of natural selection how to get the most benets from the least number
of changes in predisposing individuals to embrace wider social networks than
our evolutionary cousins.
One path was to provide major additions to cognitive and affective capacities.
Then, make parallel changes to arouser screening rules so that the same contrast
values as before can trigger more rewarding arousal release. Screening stays
focused on more ties similar to other direct personal ties. Ties are then added on
a proportionate ratio in terms of contrast values and costs. Some of the additional
bonds could provide heavy arousal release, and require more time and effort.
Given such rewarding arousal and general commitment, these ties could have a
high degree of reliability. Many resources could be exchanged along the lines of
such bonds. Others additional ties could be more or less demanding, and provide
more or less rewarding release. These proportionate ties are greater in humans
than in our evolutionary relatives; but as noted earlier, this is a very expensive
way of adding ties.
Equally important is the human case was a second strategy for adding reliable
ties without having to make as many balanced changes in arousal release
capacities and arouser screening rules. Indirect ties between individuals through
a common and emotionally laden third point such as a macro social structure can
produce a major increase in the range of reliable ties without a parallel increase
in more costly changes to affective reservoirs and to the rules governing access to
those rewards. A triangulation to a common macro attachment is an ideal bridge
between individuals. It can link a greater number of individuals with less total
effort than a direct tie-by-tie creation. However, this is only possible if the bridge
is an enhancement offering a combination of increased attractive arouser contrasts
and some cost savings. The macro structure must offer a package that cannot be
matched by adding more direct personal ties. Otherwise, individuals will stay
with direct personal ties in pursuing an attachment interest.
Emergence of Religion and Ascriptive Inequality 173
The key to nudging individuals to create these macro enhancements is found in
the preconscious arouser screening rules. Do not make balanced changes between
the expanding emotional capacities and the screening rules for arousal release.
Make uneven changes. For personal ties with heavy rewarding release, set the
screen early with an attachment interest. Then, after a relatively small number
of such ties, only enhanced arousers can access the ample arousal reservoirs
still remaining behind the arousal release screen. Additional heavy ties to other
humans therefore become less appealing, for these bonds simply will not have the
contrast values sufcient to get beyond the screening. Early screening discourages
us from going after more of the same, even if those initial arousers were deeply
moving. Something extra-special in terms of contrast values to other such ties is
required for additional heavy release from the rewards systems for this interest.
With an early screen and deep rewarding reservoirs still available, humans will
consistently seek out the means to tap these reservoirs.
EXTRAORDINARY BEINGS AT
LESS-THAN-EXTRAORDINARY COSTS
In our evolutionary context of origin with dispersed and nomadic populations
having limited technologies, what might provide such extra-special contrasts to
extend an attachment interest to other arouser packages? The additional stimuli
must have some anthropomorphic characteristics in order to piggyback smoothly
on reward systems rooted in interpersonal ties. At the same time, they must
have characteristics that set them apart from normal human beings. Otherwise,
their contrast value will not be great enough to get around the arouser screening
blocking the release of more emotional arousal for this interest. More of the same
will not do, but additional packages must be framed with some of the qualities
of the initial arousers. How about the addition of sacred beings and spirits who
have some secular human qualities useful for piggybacking, but who also have
some extraordinary qualities that humans can only dream of? These sacred beings
are ideal in terms of providing attractive contrasts to the alternative of adding
more direct personal ties to other individuals. The emergence of the sacred
does not require what popularly might be referred to as a religion gene. Just
provide early screening for an attachment interest, and provide the cognitive and
affective tools to imagine such creations. If few other alternative enhancements
are available, as is the case in early human history, these special beings will
always emerge.
However appealing these sacred creations might appear at rst glance, they
do raise a second problem. How to make these extraordinary beings available
174 MICHAEL HAMMOND
without have to pay proportionate and hence extraordinary costs? Remember
that preconscious arouser screening for additional bonds is only looking for
disproportionate packages offering a special ratio of contrasts to costs. The
screening works against additions with proportionate ratios of contrasts and costs.
What could provide the cost savings necessary to get around arouser screening
by providing extraordinary sacred arousers at the regular costs similar to those
humans are willing to pay for their closest personal bonds? With appropriate
arouser screening, it is the extended group that could provide those savings.
In a two step process, use early arouser screening, and set the enhancement
barrier high enough that only larger social groupings can provide the necessary
combination of increased contrast values without a parallel increase in costs.
The selective advantages go to individuals with arouser screening rules set such
that the extraordinary arouser contrasts that an individual or a small network can
regularly create through the sacred will not be strong enough to trigger rewarding
emotional release except with an equally extraordinary set of costs. The arouser
screening will normally discourage such a contrast-cost ratio, and only a larger
group creation of disproportionate arouser packages will be appealing in the
long run. For example, the collective can support specialists in the creation
and maintenance of the sacred, thereby taking some of the load off many other
individuals. The collective can provide exciting ceremonial rituals that are hard
to match with more restricted networks. Only the larger social group can provide
extraordinary beings at less-than-extraordinary costs.
In pursuing those collective enhancements, individuals are drawn to a wider
network of social ties and cooperation than our evolutionary relatives. Those
extended ties can become the scaffolding for the exchange of a wide variety of
social resources among a population with a common commitment to a sacred
enhancement package. This applies to both within-group exchanges framed by
a common religious symbol such as a totem, and to between-group exchanges
framed in terms of different but related totems. Since the technological demands
for these anthropomorphic creations are minimal at best, they represent the most
widely available enhancement package in our evolutionary context of origin
where technologies with alternative enhancement possibilities do not exist.
Metaphorically speaking, natural selection has no direct interest in sacred
enhancements. There is little advantage to directly wiring the human brain for
something like religion; but as we have seen, there are selective advantages that
these constructions produce indirectly through a wiring that makes it imperative
for humans to pursue enhanced arousers. Given the reliance of humans on macro
social structures as the primary source of enhancements, humans are going to be
social in a manner than other related species are not.
Emergence of Religion and Ascriptive Inequality 175
It is important to note that these group enhancement dynamics are not
emotional contagion in the classic model of the individual being swept up and
acting differently in a group context than they otherwise would act (Hateld et al.,
1994). Individuals are acting in their own interests in the pursuit of enhancements,
but they require a group context to carry off this quest. Similarly, the enhancement
model does not argue in a Durkheimian manner that the group is actually the source
of these special arousers (Hammond, 2003). Since we do not have the means to be
directly aware of preconscious processes like arouser screening, the piggybacking
of emotional rewards means that strong personal ties and constructions like
religion are going to feel equally real. Some reication of the group is virtually
inevitable in such a mix, but the basic source of the reication is still rooted in
the individual.
A sociological account of the dynamics behind the emergence of religion is
a crucial supporting argument to the classic alternative psychological account
focusing on the existential problems created by the increase in cognitive capacities
in humans. A smart species is going to have concerns that other species will
not (Hinde, 1999). Religion can certainly reduce somewhat the psychological
anxiety that humans experience, but anxiety reduction is not by itself a strong
enough selective factor to create the conditions for the emergence of a macro
social construction like religion. The network extension model of collective
enhancements adds another important selective advantage. This model of arouser
screening also ts well with the evidence noted above on the general lack of
confusion between personal and religious arouser packages using much the same
reward systems to provide the emotional release fueling these pursuits. The
selective logic was to reduce such possible confusion to a minimum by adjusting
the rules for converting additional arousers into more rewarding arousal.
The task that shaped much of these group dynamics, how to spread the web of
social interaction, is not a problem today; but we must live with the consequences
of natural selections solution to that original problem. With the exodus from
our evolutionary context of origin, the small social scale capping the expansion
of collective anthropomorphic enhancements would gradually disappear. The
pursuit of these enhancements would then proliferate. Ethnic and racial group
formation, hero worship, nationalism, and a host of other macro anthropomorphic
constructions would draw upon these dynamics of early arouser screening. These
social constructions would produce a vast range of additional problems rooted in
our ecological release, problems we must struggle with every day. If this were not
difcult enough, we must also live with another set of social inequality dynamics
originally rooted in much the same task of network extension and having many
of the same problematical consequences with our exodus.
176 MICHAEL HAMMOND
THE PUZZLE OF COLLECTIVE
ASCRIPTIVE INEQUALITY
The puzzle in regard to piggybacking of attachment and religious arousal release
was why not more confusion? The puzzle in regard to collective ascriptive
inequality in humans is why bother? With collective ascription, an entire group
is imputed to be superior in way or another over another group. This group does
not have to be closely related biologically, and it is not always necessary that
individual members of the group actually demonstrate their superiority in an open,
merit competition. For instance, all males could be imputed to be superior to
females in regard to more qualities than all females are imputed to be superior to
males. Individuals in both groups are then discouraged or prevented from actually
demonstrating their possible merit in activities assigned to the other group. There
is only moderate sexual dimorphism in humans compared to our evolutionary
cousins. Thus, the result of such collective differentiation is that some males get
some status that they would not merit; and to a greater extent, some females do
not get some status they would merit in a fully open status competition.
This collective inequality is not obviously necessary in terms of natural
or sexual selection. In our family tree, both forms of selection normally give
precedence to demonstrated performance and/or to kin relatedness. What would
be the selective advantages for individuals who engage in this additional arena of
status differentiation into gender solidarity groups? One advantage would be in
terms of the partial reduction of sexual tension. For instance, in our evolutionary
context of origin as hunters and gatherings, if some females were a regular part
of male hunting parties, and some males were regular participants in gathering or
in certain domestic activities, there could be increased problems related to sexual
activities. As Morris noted (1967) in his analysis of male solidarity groups, there
must means to reduce these problems. However, he did not suggest any specic
mechanisms to shape the behavior of men in groups or women in groups. Early
arouser screening leading to collective ascription is one mechanism to reduce
those potential problems. There might be some loss in functionalist efciency with
this group level ascriptive differentiation, but the ultimate selective benets could
be greater. Individuals who used collective ascription to reduce potential sexual
problems in the division of labor could have better survival and reproduction
rates. However, this analysis leaves open the question of the best means to achieve
this goal. How can natural selection produce this behavior pattern with the least
number of changes? As in the case of religion, early arouser screening provides
one mechanism to this ultimate evolutionary advantage. Before considering
these changes, let us consider another possible selective advantage in this group
level differentiation.
Emergence of Religion and Ascriptive Inequality 177
The enhancement perspective would suggest that there is a second advantage
to collective ascription, and that it is much the same as religion, namely a
signicant extension of reliable social ties and cooperation far beyond that of our
evolutionary cousins. Once again, much the same set of adaptations can be used
to provide such an extension in a most economical form in terms of the number of
changes to an already existent behavioral repertoire. In terms of natural selection,
the same proximate mechanism can be used to reap the same ultimate benets.
Let us examine collective ascription in terms of the enhancement model.
First of all, why can this type of ascription be seen as an enhanced arouser
package? Ascriptive inequality means that some individuals do not have to
pay a proportionate cost for some attractive status arousers. Ascription imputes
that some are worthy of status differentiation, and the others are prevented or
discouraged from competing for those status rewards. In terms of the distinction
between proportionate and disproportionate arouser packages, by not insisting
fully on demonstrated performance and by limiting merit competition, ascription
allows some individuals to get a disproportionate package with attractive contrasts
not matched with an equal amount of access costs. Thus, ascriptive inequality is
similar to religion in terms of its value as an enhancement.
Of course, the calculation of costs is somewhat different than in the case
of religion. This is because there is always some zero-sum element in status
differentiation. The gains of one set of status arousers are in part possible only
with losses on another set. Thus, there is always likely to be some resistance on
the part of individuals in the second set to the gains of the rst set. Countering
or defeating such resistance is part of the cost of accessing inequality rewards. As
we shall see, this aspect of resistance can be used to push individuals into more
collective forms of social inequality.
Ascriptive inequality is similar to religion in that its arousal screening can be
set such that in some cases, only a wider collective structure is able to tap the
rewarding emotional reservoirs protected by such screening. But once again, what
mechanismmight be used to nudge individuals into such collective differentiation?
As with religion, there could be hard wiring, but a more economical change
would be once again to use arouser screening rules that maintain a higher degree
of exibility than more direct wiring, and that piggyback the new interest on an
already established reward system. This piggybacking would be an indirect means
to encourage individuals to create certain shared macro-social constructions
providing rewarding arousal release that was not regularly available in more
micro-level situations.
This is actually not hard to do with a smart species like ours. As with religion,
there is no need to use in popular terms an ascription gene and to wire directly for
this kind of inequality. Once again, use two steps in preconscious arouser screening
178 MICHAEL HAMMOND
in order to redirect a classic interest with ample arousal reservoirs protected by
the screening. In this case, the interest is in status differentiation. First, set early
screening rules for rewarding emotional release so that individuals will soon be
looking for enhanced arousers to tap arousal reservoirs protected by these rules.
With early screening, more of the same will not do, even if those initial arousers
were very attractive. Then set the enhancement barrier high enough so that
only macro structural factors can regularly provide the special enhanced arouser
packages necessary to get around the preconscious screening without paying too
great an access cost. The result is that only extended social networks or conditions
that affect all individuals in the population will be the most reliable basis for these
additions. The neurophysiological wiring is not for collective ascription itself.
Instead, what we label as collective ascription is most often the only widely avail-
able means to offer a large percentage of a population some enhancements to their
status interest. As discussed in Hammond (2003), scientists are not as advanced
in their studies of the neurophysiology of status rewards as they are in the analysis
of rewarding arousal release in attachment. However, there is no reason to assume
that the basic mechanisms are different, or that natural selection cannot operate on
status reward systems to shape behavior in much the same way that occurs with the
shaping of extended attachments.
With the enhancement requirement set high enough, if some try for enhanced
status markers using too small a network or in conditions that only affect a small
part of the population, then other individuals who would lose out in these additions
will resist in one way or another. In small and dispersed nomadic populations of
hunters and gatherers, this means that the base line for enhanced additions must
be something like a gender solidarity group embracing a larger percentage of the
population. Or as we shall see below, the requirement becomes the existence of
conditions that temporarily or permanently expand group size, such as the periodic
gathering of normally dispersed groups, or an especially rich and non-seasonal
ecological niche that increases resource accumulation and average population size.
To understand these wider group dynamics, it is important to recognize that
our status interest is framed somewhat differently than the attachment interest.
Attachment is set so that most all individuals end up with some bonds having
a great deal of rewarding emotional release. For the initial set of arousers, even
small contrast values can lead to heavy emotional bonds. The status differentiation
interest is set up such that there are many circumstances in which there is only
moderate or small rewarding arousal release because the only status arousers
available at a reasonable cost are moderate or small in terms of their contrast
value. That is, the response to an initial set of status arousers is wired to be
proportionate. If those arousers offer heavy contrasts, provide heavy emotional
release. For moderate and small contrasts, provide moderate and small release.
Emergence of Religion and Ascriptive Inequality 179
Proportionate rewarding arousal release means that individuals will normally
not pay more than proportionate costs to access these arousers. Costs in terms of
time, effort, and risk apply both to obtaining status arousers and resisting activities
by others that might erode or even remove those status markers. Unlike the case
of religion, there is always, as noted above, some zero-sum aspects to inequality
acquisition. Any status arousers acquired by some individuals or groups mean less
status available to others. The proportionality rules apply to inequality resistance
as well. Do not expend high resistance costs to obtain or keep arousers with only
moderate contrasts values. Expend moderate resistance costs to defend moderate
contrast values. Or even better, try to nd some favorable disproportionality, and
expend small costs for moderately strong arousers.
As Maryanski and Turner (1992, pp. 8485) demonstrate, most hunting and
gathering social structures are quite uid with comparatively easy movement of in-
dividuals in and out of specic groupings. Also, the variety of ecological niches that
humans could occupy meant that in many situations, there was an option of ssion-
ing in which for one reason or another, a part of a larger group picks up and moves
away. Furthermore, with limited technologies, the acquisition of material goods as
status markers is severely hampered. These circumstances do not provide much op-
portunity for individuals or a small sub-group to use coercion or some other means
to inate inequality. In the face of an attempt at really strong escalation, there are
many avenues of resistance. At worst, some will simply move away, leaving a great
hole in any attempt to pump up inequality signicantly. Resistance that is moderate
in costs and does not require taking too many physical risks is most effective in
these circumstances. With a proportionality rule for initial status arousers, indi-
viduals are not going to expend elevated costs to protect moderate status arousers.
In our evolutionary context of origin, moderate resistance was normally sufcient
to check the ination of status arousers by some individuals or a sub-group
within a population.
In these circumstances, the proportionality for initial status responses com-
bined with early arouser screening had the selective advantage of decreasing
competition, and especially male-male competition, for key status positions in
most hunting and gathering conditions. Elevated male-male differentiation based
on demonstrated performance qualities, such as seen in chimpanzees and gorillas,
also means limited male-male cooperation. This is not a useful strategy for early
humans moving into ecological niches requiring a great deal of low technology
hunting and gathering. As noted earlier, male-male cooperation is at a premium in
such conditions (Morris, 1967). At the same time, male-male competition is still
a key part of sexual selection. How then to balance these two conditions? Use a
proportionate response pattern with early screening. This will provide a moderate
amount of individual performance based status differentiation useful in sexual
180 MICHAEL HAMMOND
selection, but also limit status competition in other pursuits. The mix is of course
never perfect. However, metaphorically speaking once again, natural selection is
not interested in perfection, but only in what is useful.
If performance based differentiation is likely to be moderate in these conditions,
so too will enhanced packages offering arouser additions for status differentiation.
Enhanced moderate ascriptive inequality is ideal. It provides moderate additions
at low costs, and is therefore an attractive enhancement package. At the same
time, by demanding only moderate differentiation and offering it to a large part of
a population, such as all males, it becomes very difcult to oppose consistently.
This is all the more the case if the excluded group, in this case, females, are
given some ascriptive inequality of their own, albeit less than that offered males.
How to promote these results? Once again, use early screening so that only
enhanced arousers will be able to tap the emotional reservoirs protected by
preconscious screening rules. Individuals successful at differentiation in terms
of demonstrated performance will soon be looking for such enhanced additions.
Individuals not as successful will be looking for any useful status arousers. The
interests of the two groups coincide around collective ascriptive inequality based
on gender.
For example, consider what early screening means for a male with a demon-
strated high performance in hunting. There can be status rewards for such a
skill valued among a wider group, but with some status arousers in hand, early
preconscious arouser screening will only permit additional rewarding emotional
release with enhanced arousers. These additions must offer a disproportionate ratio
of increased arouser contrasts to increased access costs. Where is an individual
going to nd such arousers in a hunting and gathering context? Individual level
enhancement packages for a skill like hunting would certainly be very moving,
but why would others support such an individual elevation that could only erode
their own relative status? If pressed or coerced by an attempt at status ination,
others could ultimately resist in the most effective manner by simply moving on.
This was the resistance luxury of our evolutionary context of origin. In the face
of this option, what was a high performance individual to do? One option was to
seek some moderate enhanced additions for other qualities that the individual was
not as strong as in terms of actual performance demonstration, or for qualities
that the individual simply did not have the time and energy to demonstrate an
unequal share of a valued quality. How about status additions for imputed skills in
political or domestic leadership, religious activities such as supernatural contact
and healing, warfare, or any other social role?
Higher performance individuals in one role might dream that ascriptive
additions for other roles are focused only on themselves. However, this is most
difcult in the normal hunting and gathering context. Once more, why should other
Emergence of Religion and Ascriptive Inequality 181
males support this extension of status to only a few males? Why not resist such an
extension unless they too are offered something? Faced with resistance, in order
to access these additions, higher performance individuals would have to cooperate
with many other individuals, and some of them would not be totally deserving
of these imputed status evaluations. In regard to any particular status role, higher
performance individuals would have to cooperate on a regular basis with lower
performance individuals. Otherwise, the specially priced status additions from
collective ascription will not be available. Obviously, lower performance males
would nd such ascription very attractive. Unable to garner a basic package of
status arousers based on their own actual merit, an alternative means becomes
available to access some status stimuli. The fact that these ascriptive arousers
are so reasonably priced is an added bonus for this second group. The net result
of this collusion of interests between higher and lower performers is a degree of
male solidarity unequalled in our evolutionary family tree. Altogether, the most
important result of proportionate status differentiation release rules and early
screening of additions is a major extension of male-male cooperation.
The same logic applies on a more modest scale to the emergence of female-
solidarity groups, and a more modest but still important extension of female-female
cooperation. The number of superior qualities imputed to females might be smaller
than the package for males, but as long as there are some ascriptive status arousers,
females can use these to increase their arouser package. Some females might
acquire status arouser based on performance and kin relationships, but with early
arouser screening, they too will soon be looking for some enhanced additions.
Ascription offers just such a package; and as in the male case, it is likely to be the
only enhancement readily available since the pursuit of more individually based
or kin centered status escalations will initiate a very effective resistance in most
cases. Individual merit based disproportionate rewards would be very appealing
in terms of arouser screening rules; but as noted above, these kinds of additions
are not likely to be available in these dispersed and uid social groupings with
limited technologies.
It should be noted that it is not sufcient to account for these gender solidarity
groupings by arguing that on the average males are more interested than females
in status differentiation or at least in its aggressive promotion (for instance, see
Sanderson, 2001, pp. 199200). This may or may not be the case, but there
is still a necessity to account for this peculiar extension of collective status in
comparison to other social species in our family tree. Gender differences might
limit resistance to ascriptive inequality. However, these differences cannot explain
why rather than regularly increasing male-male competition in the quest for
additional status markers, higher performance males should ally themselves with
lower performance males in gender solidarity groups. The enhancement model
182 MICHAEL HAMMOND
tries to ll in this explanatory problem by looking at the consequences of early
arouser screening on any status differentiation interest in humans.
Given this arouser screening, the greatest total release will normally come
from combining some individual achievement status differentiation with some
collective ascriptive additions. The mix is the maximization strategy. If individuals
rely solely on ascriptive arousers, there will be some rewarding emotional release
without too much effort, but the total release will be capped. Early screening
applies to both proportionate and disproportionate arousers. If the only arousers
available are ascriptive, then before long, only more disproportionate arousers are
going to have any signicant additional arousal release, and they are rarely going to
be available in our evolutionary context of origin. The best strategy is if necessary
to pay proportionately for the rst arousers, and then add on disproportionate
arousers whenever possible. Thus, early screening keeps individual performance
differentiation going, but also sets up individuals to look for some short-cuts for
arouser additions. Individuals might take the ascriptive arousers and relax, but
then they will lose out to those able to dip into additional emotional reservoirs.
Altogether, in the most commonly recurring conditions of our context of
origin, the effect of early arouser screening is to dampen inequality based on
the magnication of individual differences and to promote inequality based on
collective solidarity. Naturally, it will also be most common to combine enhance-
ment packages. For example, religion can provide excellent rationalizations of
ascriptive inequality. It becomes the will of the gods and thereby even harder
to resist. In turn, ascriptive divisions can be used to frame religious activities
specially reserved for a particular group. However, this conation should not be
seen as some kind of more or less conscious self-deception. Remember that we
have no means to be directly aware of processes like arouser screening. Arouser
packages able to get behind this screening are going to produce emotional
release that does not feel anything different than in the case where the arousers
are real. Ascriptive inequality is going to feel every bit as real as merit based
differentiation, just as religious bonds and personal attachments are going to feel
equally real. Reication will once again occur on a regular basis.
ENHANCEMENTS AND THE
ESCALATION OF INEQUALITY
Another advantage of early screening rules is that it gives humans a great deal
of exibility in expanding or contracting inequality as circumstances change. In
regard to our evolutionary context of origin, there are three good examples of the
advantage of early screening rules for increased status differentiation. All involve
Emergence of Religion and Ascriptive Inequality 183
changes in resistance costs due to other changes in macro social circumstances.
As we have seen, under the most common conditions, early arouser screening
dampens inequality enhancements focused on the magnication of differences
among individuals; but under a few special conditions, early arouser screening
fuels additional inequality in other conditions. Both enhancement packages are
useful in their appropriate circumstances. Change the context and early arouser
screening has different consequences, but there is a pattern in these differences.
Early arouser screening means that individuals will always be on the look-out
for opportunities involving enhanced arousers. Some of these opportunities are
tied to circumstances that change the arouser access costs involving resistance to
expanded status differentiation. For example, take the case of a crises like group
warfare in a hunting and gathering population. In such a situation, there is a
premium on demonstrated performance involving certain special skills (Johnson
& Earle, 2000, p. 130; Wrangham & Peterson, 1996). To encourage individuals
to play a role such as war chief or to increase solidarity among a war party,
increased status rewards magnifying differences are most useful. However, with
the disproportionality requirement of early screening for enhancements, these
expanded rewards must be paralleled by a decrease in resistance to such status
ination. The special circumstances of the war crises promote such a decrease.
Many individuals will be willing to trade temporarily some relative status to
fuel certain behaviors in others, especially if those actions involve life and death
issues. These crises conditions will trump the moderate arousal release from
normal status differentiation. Once again, there is no need to hard wire these
patterns. Early arouser screening makes it comparatively easy for individuals
and groups to expand, and later to contract status differentiation in relation to
different situations.
Another good example of the exibility accompanying early screening is in
regard to the periodic temporary assemblies that are so often a part of hunting
and gathering cultures. The assemblies are useful for the exchange of a range of
resources and for the accomplishment of special tasks, such as expanded group
hunts or religious ceremonies (Johnson & Earle, 2000, pp. 6263, 71, 78; Lee,
1979, pp. 103104). However, these gatherings do create a potential problem in
terms of decision making and organization. Smaller groupings can engage many
individuals in these processes, and this provides a variety of status arousers to these
individuals. If there are attempts to maintain this same consultative pattern in the
temporarily expanded group, many difculties are likely to arise. Increased status
differentiation could reduce some of these difculties. If a smaller percentage of
individuals take on increased roles, then some efciency can be maintained.
How to encourage some individuals to take on these roles, and how to
discourage others from resisting such a change due to the fact that they suffer a
184 MICHAEL HAMMOND
temporary decrease in relative status? Early arouser screening helps to accomplish
both goals. This screening has individuals on the look out for enhanced arouser
packages. These assemblies offer just such a package for the escalation of
differences among individuals, if only on a temporary basis. There are added
opportunities for status differentiation, and also for religious activities, sexual
pursuits, or whatever. The result is that these gatherings increase the cost of
resistance to expanded inequality by limiting the appeal of ssioning. Individuals
might choose not to partake in these assemblies, and thereby avoid some tempo-
rary status loss, but they are surrendering many other attractive arousal release
opportunities. Most will accept the short-termloss of some relative status, and reap
other benets.
There is little need for coercion here. Individuals can accept this as a trade-off
because of the proportionality setting for initial arousers. As long as the loss of
some status arousal release is compensated for with gains in rewarding emotional
release for other interests, most individuals will go along with this trade off.
Indeed, a proportionality setting for an interest like status differentiation invites
some compromise and trade-offs. At the same time, some individuals can pursue
performance based enhanced status arouser packages. They must put out more
efforts to compete for elevated status, but if the rewards are appropriately inated,
there will always be candidates for such positions. Once again, only the expanded
group context, in this case temporarily expanded, can provide the right circum-
stances for more enhanced arousers. The resulting mix of ascribed status arousers
and enhanced personal differentiation based on demonstrated performance is
not always an ideal mix, but natural selection is not interested in perfection,
only in what is useful. With early arouser screening, another form of enhanced
status differentiation emerges in these special conditions, and there are many
benets. Even though some individuals benet more than others from this status
extension, most all individuals who are part of populations that periodically gather
together for the exchanges of resources are more likely to have better survival
and reproduction rates. Factors that facilitate such gatherings, such as increased
status differentiation, could still be in general benecial even to those who
temporarily lose some relative status.
A third example of early arouser screening and changing macro contexts
concerns the rare but still important cases of permanent settlements in special
ecological niches that are particularly rich in food sources and thereby limit
or eliminate a nomadic aspect to year to year life. These conditions make
possible larger populations and larger resource accumulations on a long-term
basis (Johnson & Earle, 2000). As in the case of the temporary assembly, these
circumstances limit the appeal of ssioning as a means to resist expanded status
differentiation. The costs of moving to other less rich ecological niches are
Emergence of Religion and Ascriptive Inequality 185
much greater, and that makes possible much greater risk taking on the part of
those seeking enhanced individual differentiation. Make physical risks a part of
resistance costs, and given the screening rules that discourage paying high costs
to protect moderate status arousers, many will not resist as long as other interests
such as food provisioning are not compromised by the expansion of inequality.
The big man or chief who as part of his inated role redistributes basic goods and
services is a good example of the likely outcome of early arouser screening with
permanently settled populations taking advantage of special ecological niches in
our evolutionary context of origin. In the pursuit of enhanced arousers, the big man
and his closest supporters might work more than others in the pursuit of inequality,
but they also receive status rewards that are of even greater contrast value. These
enhancements can be seen as benecial in terms of both functionalist efciency
and sexual selection. With a special ecological niche, sexual selection could lead
to increased competition through status differentiation nowthat the group does not
have to maintain the same kind of social cooperation required for a pure hunting
and gathering context. Similarly, the increased population size and resource use
could create organizational and decision making problems for the larger group, and
these problems would fuel individuals turning to increased status differentiation
as one means to cope.
The side by side combination of inequality dampening and expanding dynamics
is not always a comfortable t. However, the selective advantage in having both
patterns available was far greater than the difculties created by the combination.
This has often led to some confusion in the analysis of inequality. Some theorists
have stressed the dampening aspects and used early human history as a template
for social organization with a minimumof inequality (for instance, see Lee, 1979).
Others have correctly noted that an interest in inated inequality always seem to
be bubbling just below the surface, even in the least stratied cultures, and that as
soon as conditions are right, inequality expands (Sanderson, 2001, pp. 288294).
The enhancement model tries to demonstrate how these different patterns can
co-exist, and how there are selective advantages in this co-existence. Once again,
metaphorically speaking, natural selections interest is not in things that t
together perfectly, but rather in things that work. Early arouser screening worked
well in both dampening certain kinds of inequality while promoting other forms
in one set of circumstances, and then expanding the originally dampened forms
in other conditions.
What were exceptional conditions in our evolutionary context of origin become
common conditions with our exodus, and inequality expands as humans settle
down and social scale grows. Like all experiments in natural selection, the use
of collective ascriptive inequality was context specic. Change the context, such
as with the spread of permanent settlements and general population growth, and
186 MICHAEL HAMMOND
the consequences are very different. The checks and balances of our evolutionary
context of origin began to disappear, but the imperative to nd enhanced inequality
opportunities continued.
The core proportionality rule for a status interest handicaps resistance because
it closely limits the costs individuals will bear in opposing rising inequality that
does not benet them. Individuals will generally not pay high costs to resist
elevated inequality in order to protect moderate or small status differences. This
is not too important in conditions similar to our context of origin because those
circumstances have by themselves a dampening effect on the possibility of inated
status differentiation. With our exodus, the proportionality handicap became
increasingly severe. More and more individuals and groups were rewarded for the
pursuit of disproportionality enhanced status packages, while resistance could not
offer a counter-set of equally moving arousers for more limited differentiation.
It would be thousands of years before another enhancement alternative emerged
with high technology mass production. This economy meant that even moderate
status differences can be rewarded with a variety of technologically enhanced
goods and services. With a third set of enhancements widely available, increased
opposition to the hyper-ination of ascriptive inequality can be indirectly fueled by
large groups of individuals in the pursuit of these enhanced economic additions
(Hammond, 1999).
Most often, the post-exodus magnied collective inequalities were tied to
equally inated anthropomorphic enhancements like religion, race, ethnicity, and
later, nationalism. With a status interest proportionality rule handicapping resis-
tance, this was a toxic combination in which some individuals could double-dip
with the same enhancement package. They could reach into emotional arousal
reservoirs through both attachment and status reward systems. In contemporary
high technology societies, once again, we can partially deate these classic
enhancements by offering technologically inated arousers for a wide range
of human interests, with or without early arouser screening. Similarly, ascrip-
tive inequality can be replaced more and more by demonstrated performance
differentiation. Nonetheless, the enhancement imperative model highlights the
probability that classic packages like religion and ascriptive inequality will not
simply disappear. They are like sugar to a species with a sweet tooth.
CONCLUSION
The enhancement imperative has had an impact on human social groups since
our very origins. Conditions change, but the imperative continues. Larger groups
remain the focus of these processes. No other context can provide so regularly the
Emergence of Religion and Ascriptive Inequality 187
special arouser packages required by this imperative. Even though today religion
has to compete more often with alternative anthropomorphic social constructions,
many cannot resist embracing one or another of these special group creations.
Even though collective ascription is used less than in the past, dominant inequality
structures still have disproportionately magnied rewards, and this too suggests
that early arouser screening is in play again. Tracing the continuing inuence of
these enhancement dynamics in social groups is the next step in the development
of the enhancement model.
REFERENCES
Carlson, N. (1995). Foundations of physiological psychology. Boston: Allyn & Bacon.
Fischer, C. (1982). To dwell among friends. Berkeley: University of California Press.
Hammond, M. (1999). Arouser depreciation and the expansion of social inequality. In: D. Franks &
T. Smith (Eds), Mind, Brain, and Society: Toward a Neurosociology of Emotion (pp. 339358).
Stamford, CT: JAI Press.
Hammond, M. (2001). The enhancement imperative and uneven neurophysiological change: The case
of anthropomorphic religion. American Sociological Association, Anaheim, CA.
Hammond, M. (2002). A new model of uneven development in human neurophysiology. Human
Behavior and Evolution Society. Rutgers University.
Hammond, M. (2003). The enhancement imperative: The evolutionary neurophysiology of
Durkheimian solidarity. Sociological Theory, 21(4), 359374.
Hateld, E., Cacioppo, J. T., & Rapson, R. L. (1994). Emotional contagion. Paris: Cambridge
University Press.
Hinde, R. (1999). Why gods persist: A scientic approach to religion. New York: Routledge.
Johnson, A. W., & Earle, T. (2000). The evolution of human societies. Stanford: Stanford University
Press.
Joseph, R. (1996). Neurophychiatry, neuropsychology, and clinical neuroscience. Baltimore: Williams
& Wilkins.
Kanezawa, S. (2001). Where do social structures come from? In: H. Walker, S. Thye, E. Lawler &
M. Macy (Eds), Advances in Group Processes (Vol. 18, pp. 161183). Elsevier.
Koukouas, E., & Over, R. (1993). Habituation and dishabituation in male sexual arousal. Archives of
Sexual Behavior, 31(6), 575585.
LeDoux, J. (1996). The emotional brain. New York: Simon & Shuster.
Lee, R. B. (1979). The !Kung San: Men, women, and work in a foraging society. Cambridge:
Cambridge University Press.
Maryanski, A., & Turner, J. (1992). The social cage: Human nature and the evolution of society.
Stanford: Stanford University Press.
Massey, D. (2002). A brief history of human society: The origin and role of emotion in social life.
American Sociological Review, 67, 129.
Meuwissen, L., & Over, R. (1990). Habituation and dishabituation of female sexual arousal. Behavior
Research and Therapy, 28(3), 217226.
Newberg, A., dAquili, E., & Rause, V. (2001). Why god wont go away: Brain science and the biology
of belief. New York: Ballantine Books.
188 MICHAEL HAMMOND
Rolls, E. T., Murzi, E., Yaxley, S., Thorpe, S. J., & Simpson, S. J. (1986). Sensory-specic satiety:
Food-specic reduction in responsiveness to ventral forebrain neurons after feeding in the
monkey. Brain Research, 368, 7986.
Sanderson, S. (2001). The evolution of human sociality. Lanham, MD: Rowman & Littleeld.
Saver, J., & Rubin, J. (1997). The neural substrates of religious experience. The Journal of
Neuropsychiatry, 9, 495499.
Turner, J. (2000). On the origins of human emotions. Stanford: Stanford University Press.
Wrangham, R., & Peterson, D. (1996). Demonic males. Boston: Houghton Mifin.
TOWARDS A UNIVERSAL POWER
AND STATUS THEORY OF EMOTION
Robert Thamm
ABSTRACT
It is the general purpose of this chapter to introduce assumptions, postulates
and hypotheses concerning the social nature of human emotions. I will
propose some universal social causes of emotion categories by integrating
Kempers (1978) power and status dimensions in dyadic relations to universal
structures of human groups. These structures, of Self and Other meeting or not
meeting expectations and receiving rewards or not, predict specic emotion
categories. Power and status dimensions are added to the model and dened
in terms of expectation/sanction (E/S) states, and are proposed to be univer-
sal as well. Furthermore, changing E/S conditions produce corresponding
changes in power/status relations, and changes in emotion categories. These
changing social structural conditions cause individual anxieties to emerge.
Extending Kempers theoretical conceptualizations, gaining or losing power-
advantage or status-advantage predicts syndromes of universal anxiety
emotions.
INTRODUCTION
In this paper I will propose a theoretical framework for predicting emotion
categories using universal social structural variations. Specically, the purpose
is to introduce some propositions suggesting that the structure of human groups
and emotions are universal. I will then argue that there is a direct link between
Theory and Research on Human Emotions
Advances in Group Processes, Volume 21, 189222
2004 Published by Elsevier Ltd.
ISSN: 0882-6145/doi:10.1016/S0882-6145(04)21008-6
189
190 ROBERT THAMM
specic universal social substructures and specic universal emotions. Finally,
I will integrate these substructure dimensions into one general model explaining
how actors gain and lose power or status in changing social relations, and how
universal anxiety emotions emerge from them.
I will not be concerned with physiological manifestations of emotions or with
culture-specic explanations of how emotions are expressed or managed. The
discussion will be limited to the relations between social structural dimensions
and how they predict emotion categories. I will begin with some assumptions and
postulates concerning the general biological nature of humans and the universal
nature of human groups.
My rst assumption is that all humans have basically the same biological
structures (anatomy and physiology) that enable them to experience the whole
range of emotions, just as they have the capacity to express any behavior, or any
belief. Behaviors, attitudes and emotions are all personality attributes generally
triggered by social conditions. The manifestation of most emotions at any time
and place is a function of the social structural conditions that effect individuals in
social situations (Kemper, 1978).
In a larger sense, emotions are the effects of social relationships, institutions
and processes (Barbalet, 1998). This does not exclude the possibility that some
emotions (e.g. fear and pleasure) can come from none-social situations when
assessments of emotion-relevant conditions do not include others.
Secondly, most human emotions originate in human groups. As the structures
of these groups change, emotions tend to change. Without variations in social
situations, individuals would not normally experience changes in their emotions.
These of course are assumptions from a sociological perspective.
UNIVERSALITY OF EMOTION
Whether human emotions are universal or culture-specic has been debated
extensively in the literature. Sociologists, the most recent participants in this
controversy, feature human emotion as a function of social environmental
conditions (see, e.g. Barbalet, 1998; Collins, 1975; Heise, 1979; Hochschild,
1979; Kemper, 1978; Scheff, 1979; Shott, 1979; Thamm, 1975). Conversely,
many psychologists, sociologists and anthropologists have provided a rich variety
of studies illustrating that emotions are socially and cognitively constructed,
and are therefore culture-specic (see, e.g. Averill, 1980; Harre, 1986; Heelas,
1986; Hochschild, 1975; Lazarus, 1984; Levy, 1984; Schachter & Singer, 1962).
Scheff (1983) reviewed the arguments and evidence surrounding this contro-
versy and attempted to reconcile the conict between the universalistic and the
Towards a Universal Power and Status Theory of Emotion 191
culture-specic position. He judged that evidence for universality to be at least
as strong and probably stronger than the evidence for cultural specicity. The
discussion will continue favoring the universal assumption.
SOCIAL CONTENT VS. SOCIAL STRUCTURE
Next I am concerned with the distinction between social content and social
structure. My assumption is that there are universal group structures, but only
the social content of groups is culture-specic. The quest for understanding
this distinction might have begun with the proposal that emotion is a social
construction, and therefore culture-specic.
I agree that the content of human groups is culture-specic, and varies from
one culture to another. Group content contains much information about expected
behaviors and what appropriate sanctions should be, but is generally not concerned
with structural variations. The value of content analysis is that it effectively predicts
how actors subsequently manage and express their emotions, but it is of limited
value in explaining how emotions originate in the rst place.
Although the Constructionists are somewhat concerned with causal conditions,
the events preceding emotions are generally viewed as culture-specic rather
than universal. Of course a theory of content is possible but it would seem to
be limited to the consequences of emotion in terms of how emotions are subse-
quently expressed, but would be somewhat limited in offering antecedent causal
explanations.
To illustrate the distinction between culture-specic content and universal social
structure, lets briey examine three culture-specic (middle-class American) con-
tent situations where actors behaved shamefully. The rst actor took some money
from his roommate. The second actor cheated on his income taxes and the third
actor lied to his friend. The content in each case is different and culture-specic,
but in each case the actor failed to meet the social expectations of the culture.
It is the appraised failure to meet expectations, not the content, that predicts the
shame. Such structural denitions of shame then transcends all culture-specic
content examples, regardless of time and place, and therefore takes on a more
universal character.
Thus what we need to know to explain the universal nature of emotion is the
common conditions that occur in each emotion-producing situation, at any time,
at any place, for any person. Only then can we begin to understand what causes
specic emotions to occur. What for example is common in all guilt-producing
situations, or in all jealousy-producing situations, or in all any-other-emotion-
producing situations?
192 ROBERT THAMM
These common factors, if we can uncover them, can be dened as necessary
universal social structural conditions in the emotion causal chain. Such indepen-
dent conditions must then be appraised for a given emotion to emerge. Collectively
they provide sufcient conditions to predict specic universal emotion categories.
It would then be possible to make these kinds of predictions independent of social
content in any culture-specic situation.
Proposition1. Groupstructure is universal but groupcontent is culture-specic.
Universal group structures predict emotion categories while cultural content
predicts emotion management and expression.
In everyday application however, we would hardly use structural denitions of
emotions without rst citing the specic content of a situation. We thus tend to
overly elaborate the content and fail to acknowledge the underlying structures.
This contributes to the mystication of emotions since the underlying structures,
not the content, are the necessary components in dening and predicting universal
emotion categories.
If universal social forms do exist, and if we universally have the capacity to
experience any human emotion, then I suggest that there may be a link between
the two. If this is the case then two important questions must be asked. What are
the universal social forms that are emotion-relevant, and how are they linked to
specic human emotions?
UNIVERSAL EXPECTATION AND
SANCTION DIMENSIONS
Now I am going to attempt to elaborate the primary universal structural dimen-
sions that are linked to universal human emotions. I will rst propose that the
two universal social structural dimensions include expectation and sanctioning
substructures.
These dimensions have been central in social theoretical and research literature
for many years. Hochschild (1983, p. 85), for example, contended that emotion
deals with expecting and wanting. Jonathan Turner (2002, p. 83) believed
that there are two critical dimensions of any interaction that constrain and
circumscribe the valence and amplitude of emotions, (1) sanctions and (2)
expectations. Thoits (1985, p. 223) emphasized the importance of conformity
to social expectations and obtaining social rewards to the self-labeling processes
in understanding emotions. Abelson (1983) featured non-conrmed expectations
producing affective processes as central in his analyses. And Kelleys review of
four affect theorists mentioned above (1984, p. 93) notes that one feature common
Towards a Universal Power and Status Theory of Emotion 193
to all of the theories is the positive or negative, pleasant or unpleasant, wanted or
unwanted, nature of the experience, obviously sanctioning dimensions.
Scheff (1990) has emphasized the importance of attunement in understanding
the social bond. He asserts that attunment results in pride (an emotion that will
be identied with conformity) and its lack with shame (an emotion that will be
identied with deviance). He maintains that pride and shame are the primary
motives in human conduct and are signals of the state of the bond, and are
fundamental (and perhaps universal) motives. Turner (2002, pp. 7577) also
linked meeting expectations with pride, and not meeting expectations with shame.
Research supporting the universality of sanctioning systems also derive from
the works of behavioral psychology and exchange theory. The behaviorists in
the tradition of Skinner (1938) explicitly state that emotions are the effects of
sanctions (reinforcement patterns). Traditional research from an exchange theory
perspective (see, e.g. Blau, 1967; Homans, 1961; Thibaut & Kelley, 1959) also
focused on the assumption that sanctioning patterns (rewards and punishments)
are central in predicting emotion responses.
Ortony et al.s analysis (1988) of these two dimensions has provided a seminal
model for predicting emotions. They argue, for example, that disapproving of
ones own action (not meeting expectations) plus an undesirable event (receiv-
ing punishment) equals the emotion of remorse. It is apparent that expectations
and sanctioning are extensively discussed dimensions in the literature. But are
they universal and do they predict emotion categories?
Parsons and Shils (1951) combined the expectation and sanctioning dimensions
in describing (assumedly universal) dyadic interaction patterns. Kemper (1978)
after an extensive review of early theorists and empirical studies did not abandon
rewards and punishments as universal emotion-relevant dimensions. He concluded
that the manifestation of any specic emotion at any time and place is a function
of the social structural conditions that effect individuals in real, imagined or
anticipated social situations.
The following expectation-sanction postulates are introduced as universal social
forms found in all human primary groups.
Postulate 1: The rst universal social substructures are expectation systems.
All institutions and all social groups have specic expectations of their members.
They are implicit in roles, norms, morals, customs, laws, rules, etiquette, fads, and
other folkways which dene the accepted ways of doing things in social settings.
They are built into the role and normative structures of institutions, but are less
formalized in loosely structured groups.
Postulate 2: Specic expected behaviors in any social group are not universal,
but appraisals as to whether or not any expectation is met, are. All humans in
all human groups cognitively evaluate compliance to expectations or deviation
194 ROBERT THAMM
from expectations. Thus it is not specically what the behavior is (the content)
that is emotion-relevant, but whether or not the expectations were met. Who
performed well, and who failed to perform are the rst universal emotion-relevant
dimensions.
Postulate 3: The second social substructures that are universal and found in
all human primary groups are sanctioning systems. All institutions and all social
groups have specic sanctions for members who comply or deviate. Sanctions
reect the cultural value systems, and prescribe how rewards and punishments are
to be distributed among members in all social groups.
Postulate 4: The specic sanction attached to behaviors in any social group is
not universal, but appraisals as to whether or not the situation is rewarding or not,
are. All humans in all human groups cognitively evaluate being rewarded or being
punished. It is not the content of the sanction that is emotion-relevant, but whether
or not it is appraised as being rewarding or punishing. Who received rewards and
who received punishment are also universal emotion-relevant dimensions.
In any primary group, seven universal structures can be elaborated: (1) basic
structures (how each actor performed or was sanctioned); (2) distribution of
performances (comparing each actors performance); (3) distribution of rewards
(comparing each actors sanction); (4) attributions of actors (comparing how an
actors performance effected their sanction); (5) interactions of actors (comparing
how one actors performance effected the other actors sanction); (6) power and
status relations (comparing each actors attributions); and (7) structural transitions
(the social process of each of these variations changing to another. Some of
each of these variations along with their hypothesized emotion labels will be
elaborated below.
Before continuing, we might ask whether or not congurations of these
universal variations constitute all emotion-relevant structures. Does this exhaust
the list of universal structural variations in human groups? Are there any additional
structures found in all human groups? Perhaps there are, but none that I am aware
of. Age, ethnicity, gender, and the like, are not properly structural but are social
constructions of biological traits. They are culture-specic, and thus not universal
in form. For the remainder of the discussion, I will assume that this list of seven
social substructure forms is complete.
THE EXPECTATION/SANCTION
(E/S) PARADIGM
The immediate task is to examine how these universal expectation/sanction
dimensions apply to human emotions.
Towards a Universal Power and Status Theory of Emotion 195
Proposition 2. Expectations and sanctioning as universal substructures occur
in all human groups and as they are appraised predict universal emotion
categories. The link involves actors previously internalizing norms and values
through the socialization process. They are thus able to detect when they
and others are meeting (or not meeting) internalized social expectations and
are receiving (or not receiving) what has been internalized as rewarding.
Of course what is expected and what is rewarding is content-specic to the
culture in which any actor has been socialized. In this sense having emotions
is dependent upon culture-specic socialization, whether in the form of formal
institutionalization or informal group interaction. Generally, when actors have
no expectations of themselves and others, and cannot detect what is rewarding
or not, they will have no emotion response.
Proposition 3. Each universal substructure conguration when appraised
predicts a specic universal emotion category. The assumption here is that
after an actor appraises one of these universal substructures, they will have
the emotion linked to it. Each of these social structural variations can then be
used to predict a corresponding emotion category, and each emotion category
is dened by its structural conguration. How actors subsequently manage and
express their emotions is culture-specic, and not of interest here.
If universal group structures and substructures are emotion-relevant, then we can
argue that as specic appraisals of these structural conditions are made in dyadic
interaction, specic emotions can be predicted. The social dyad is composed of two
Actors (Self and Other) in interaction and is the smallest human group structure.
The universal structures found in dyadic situations are found in all human groups,
and these structures along with corresponding emotions will be the focus in the
following discussion.
I have introduced the expectation/sanction paradigm (1992), a tool in delin-
eating the substructure components of emotion systems in dyadic interaction.
The paradigm suggests that the structure of emotion is determined primarily by
resolving two fundamental structural questions; rst, Did each of the actors meet
or not meet expectations? and second, Did each of the actors receive rewards
or not?
In any specic social situation Self and Other can be assigned values reecting
the answers to these questions. The answers are represented below in quadrant
form. They are always asked from the perspective of Self-appraising the emotion
structure.
Permutations of these appraised responses predict specic emotion categories.
Emotions linked to the eight one-conditional substructures and the 24 two-con-
ditional substructures have been elaborated, tested and reported (Thamm, 1992).
196 ROBERT THAMM
In any focused social situation, actors may appraise any number of these
conditions. When a condition is neither appraised positively or negatively,
emotion neutrality or affective neutrality for that condition exists. Most of our
ongoing interaction is unemotional as emotion-relevant conditions are unknown,
undetermined or not focused upon, and are therefore considered affectively
neutral. Only when the emotion-relevant conditions are focused on and appraised
positively or negatively could we have an emotional reaction. Thus emotions can
be more or less complex ranging from one condition to four conditions. If, for
example only one condition is appraised as relevant, the other three conditions are
neutral. Neutrality in appraisals will be symbolized by leaving the condition blank
in the paradigm.
For example, even though only the Self-conditions are appraised as positive or
negative while the appraisals for Other are neutral, a social relation still exists
between the two. Thus an actor can have positive or negative attributions while the
other has attribution neutrality. In this manner the whole social situation is always
relevant but at any given time and place, some conditions may be more relevant or
more neutral. This will be apparent later as we explore power and status attributes
of an actor when the attributes of the other actor are neutral.
In the E/S paradigm outlined above, it is also assumed that there is a consensus
on a common culture, that both actors share expectations and value the same
rewards. Both actors agree on the prevailing norms and values and appraise the
emotion structural states the same. For example, if Self appraised each of the four
conditions as , Other, in reversing the column signs, would appraise the same
situation as . The two actors would then have opposite or complementary
emotions but would agree with each others appraisals.
Towards a Universal Power and Status Theory of Emotion 197
It is also possible for an actor to evaluate himself/herself as meeting expectations
in a particular situation while another actor evaluates the rst as not meeting
expectations. In this case one actors evaluation may be inconsistent with how the
larger group or society would make the evaluation. We can assume that at least one
of them must have identied with alternative or deviant group expectations and
values. Another exception is when an actor escapes social reality and appraises
imaginary conditions uniquely and abnormally. In either case, a lack of consensus
on norms and values can contribute to conict in relations. These inconsistent and
idiosyncratic evaluations between actors can be handled perfectly well within the
theoretical structure since an actors evaluations can also be elaborated independent
of another in applying the E/S paradigm.
In proposing this model, however, I will take the functionalist position and
assume consensus and normality, remembering that the emotion appraisal is
always from the perspective of Self-evaluating the situation. Assuming this, there
is still conict within the model as some actors may not comply with their own
internalized social expectations, and re-dene their own actions as not meeting
expectations, as in the emotions of Self-shame or guilt.
Proposition 4. Social structures including power, status, and social processes
are variations of universal expectation/sanction dimensions, and in combination
predict more complex universal emotion categories.
How power and status dimensions t into this paradigm is our next concern.
POWER AND STATUS ATTRIBUTES
I have shown how universal group substructures, including expectation and
sanctioning systems, predict emotion categories. But there are two additional
primary universal substructures that predict more complex emotions. Kemper and
Collins (1990, pp. 3268) argued that there is converging evidence from diverse
perspectives for the application of a two-dimensional model of status and power
to the domain of emotions. Kemper rst introduced a social interaction theory
of emotion proposing that power and status are universal dimensions of social
outcomes that predict emotion categories. As actors interact, emotions ow from
the outcomes of the different status and power relations (Kemper, 1978, 1984). He
believed that the power and status concepts were so fundamental for the analysis of
emotions that they would be retained by whatever other name was applied to them.
To include these universal emotion-relevant dimensions in the theoretical
model, these attributes must be expressed in terms of the expectation/sanction
198 ROBERT THAMM
(E/S) conditions and states outlined above. Thus I propose that power and status
dimensions can be dened in terms of universal E/S conditions, and are therefore
universal as well.
Attribution Power vs. Attribution Status
Status and power dimensions have been examined in such diverse research areas
as small groups, cross-cultural behavior, personality theory, clinical psychology,
neurophysiology, and social and cultural macroanalysis.
These social dimensions as they relate to emotion labeling have been inves-
tigated by a number of social psychologists. Osgood, May and Miron (1975)
presented evidence for the cross-cultural universality of power and status. Osgood
and Tannenbaum (1954), Heise (1979) and Smith-Lovin and Heise (1988) have
explored the dimensions of goodness (status) and potency (power) in their
theories of affect control. Kemper (1978) in his theory and research argues for the
centrality of power and status dimensions, and also suggests that these dimensions
in social interaction have universal properties.
Many theorists have offered denitions of power and status and have suggested
some distinctions between them. This rather extensive literature was summarized
by Kemper (1978). He reviewed theorists including Weber (1947), Parsons (1951),
Homans (1961), Blau (1964), Osgood et al. (1954), Thibaut and Kelley (1958),
Heise (1969), Kemper and Collins (1990), Scheff (1990), and many others. The
table belowlists some of the major concepts used by these scholars to dene power
and status. In some cases I have taken the liberty to negate concepts and place their
opposites in the respective powerless and statusless cells.
Towards a Universal Power and Status Theory of Emotion 199
I suggest that the concepts in each cell can be summarized in terms of their
positive or negative representations. A consensus of any set of judges would
probably group each concept listed into either a negative or positive category.
Below is my interpretation of the action/sanction valences reected in each cell.
After translating these valences into the expectation/sanction paradigm, I
propose the following simple denitions of an actors power or status attributes.
They are dened rst in terms of valences outlined in the paradigm and then in
terms of the expectation and sanction dimensions. The corresponding graphic
sets, reecting the conditions and states in the E/S Paradigm, follow the
denitions
Power: When Self did something negative but received something positive, or
Self did not meet expectations but received rewards .
Powerless: When Self did something positive but received something negative, or
Self met expectations but did not receive rewards .
Status: When Self did something positive and received something positive, or
Self met expectations and received rewards .
200 ROBERT THAMM
Statusless: When Self did something negative and received something negative,
or Self did not meet expectations and did not receive rewards .
In further discussion I will not use positive and negative valences in dening
power and status. These concepts will be elaborated only in terms of meeting
or not meeting expectations and receiving or not receiving rewards. Also, I will
attempt to write during the remainder of the text in a manner that the content can
be understood independent of the accompanied graphics. The graphics are used
as a supplement to the text so that the reader may refer back to the E/S paradigm
for a more formal elaboration.
Given these power and status denitions, how then are power concepts
generally differentiated from status concepts? It is apparent, as Kemper (1978,
p. 35) suggested, that rewards are not the differentia. I would agree and propose
that the condition that distinguishes attribution power from status is not the
sanctioning dimension but whether or not expectations were met in the social
situation. Both power and status suggest that actors are rewarded but power
derives reward as a result of not meeting expectations (deviance), including
coercion, as Kemper argues. Status however requires compliance in meeting
expectations (conformity), also argued by Kemper.
When compliance fails to yield rewards (unsuccessful conformity) we have
powerlessness. There is some evidence that this emotion is associated with
meeting expectations but not being able to gain rewards (Thamm, 1992). Thus
powerlessness implies meeting expectations, but to no avail. Powerlessness in
relations may result in forced compliance by another. But whether compliance
is forced or voluntary, it is still compliance, and the feeling of powerlessness
emerges whenever compliance fails to yield rewards.
On the other hand, if actors deviate, dont get caught but get what they want
anyway, they have power. If they get caught their rewards may be terminated, their
deviance is unsuccessful, and they become statusless. When compliance is within
the control of an actor, and rewards follow (successful conformity), status exists.
Actors may prefer status to power but when conformity fails, power (successful
deviance) may be the only means of obtaining rewards.
I have argued that E/S dimensions are universal substructures of human groups.
If power and status can each be dened in terms of E/S conditions then power and
status are also universal substructures of human groups.
Proposition 5. Each power or status variation when appraised in terms of ex-
pectation/sanction dimensions predicts a specic universal attribution-emotion
category.
Towards a Universal Power and Status Theory of Emotion 201
Power and Status Emotions in the Dyad
The attribution emotions, comparing an actors meeting or not meeting expec-
tations with the resulting sanction, reect the power and status dimensions of
universal group substructures. If each power/status variation expressed in terms
of E/S dimensions is a specic universal attribution emotion category, then
each of them can be elaborated in terms of their attribution properties, graphic
representations, and hypothesized emotion labels. Power and status variations
and emotions are then attributed to Self and Other using E/S dimensions in dyadic
interaction.
Power Attribution and Emotion. Universal power attributes include getting away
with something. It involves receiving gratication by circumventing the rules.
It implies that the actor has been somewhat deceptive and devious and that the
rewards that follow are undeserved. It may take the form of successful dishonesty,
successful crime, successful immorality, or shameful joy. but whatever the form
of deviance, it must always result in some kind of overt or covert success (rewards).
Its dyadic variations and emotion categories include:
Self has high power: After Self did not meet expectations but was rewarded ,
Self feels manipulative.
Self has low power: After Self met expectations but was not rewarded , Self
feels powerless.
Other has high power: After Other did not meet expectations but was rewarded
, Self feels disgusted with Other.
Other has low power: After Other met expectations but was not rewarded ,
Self feels compassion for Other.
Status Attribution and Emotion. Universal status attributes includes getting what
you earned. It involves receiving gratication by following the rules. It implies
that the actor is to some degree complying and cooperative and that the rewards
that follow are deserved. It may take the form of successful honesty or successful
morality, but whatever the form of compliance it must always result in some kind
of overt success. Its dyadic variations include:
Self has high status: After Self met expectations and was rewarded , Self feels
honored.
Self has low status: After Self did not meet expectations and was not rewarded
, Self feels disgraced.
Other has high status: After Other met expectations and was rewarded , Self
feels impressed with Other.
202 ROBERT THAMM
Other has low status: After Other did not meet expectations and was not rewarded
, Self feels disillusioned with Other.
These eight labels were hypothesized and tested successfully as best-t power and
status emotion labels (Thamm, 1992).
Attribution vs. Distribution Emotions. Thus far power and status as attributes
of actors have been examined somewhat independent of another actor within
the group. But more complex emotions emerge as a result of power and status
distributions in relations. In fact, a very large class of human emotions results
from such outcomes of these relations (Kemper, 1978, p. 43).
In any social situation actors may appraise none, one, two, three or four of
the conditions. As suggested above, when none of the conditions are appraised
positively or negatively, emotion neutrality exists. That is, the actor does not
have an emotion or an emotional response. Most of our ongoing interactions are
unemotional since most social conditions are not relevant to our emotions. When
one or more of the conditions become relevant and are appraised positively or
negatively, an emotion results. Actors can then selectively perceive, focus and
evaluate the conditions they deemrelevant, and ignore or repress others. For exam-
ple, Self in a specic interaction with Other might be unconcerned with appraising
Others attributes as Other neither meets nor fails to meet expectations and is
neither rewarded nor punished. Selfs appraisal of Other is neutral and therefore
unemotional. In this way Self may appraise only the self-attribution conditions
while still in a social situation involving Other. Thus Self may have an attribution
emotion while Other attributes are assessed neutrally. Here I am suggesting that
when no attribution valences are assessed, neutrality in power and status relations
is assured.
In this manner, power assessment may be made of one actor and not of the
other. That is, Self may have power while Other is neutral. However, this assertion
violates the assumption of how power and status is used by Kemper and most
sociological social psychologists, including many that have already been cited.
They would argue that an actor could have power only in the context of a relation-
ship with a specic other, as when an actor has the power to punish another if they
do not comply.
I certainly agree with this popular use of power, but I hope there is room to
consider an extension of this thinking, and allow power to exist as only one actors
attributes, as in the example when an actor deviates successfully while the Other
actors conditions are neutral.
The problem with demanding the valence evaluations of both actors is that it
negates the possibilitythat anactor canhave anemotionindependent of a positive or
negative evaluation of another actor. This seems a rather narrow conceptualization
Towards a Universal Power and Status Theory of Emotion 203
of emotions in that it would suggest that emotions only exist if both actors are
somehow evaluated and compared in interaction.
I suggest that if Self has power, Other may also have power, have no power or
have power neutrality. Sopower andstatus emotions mayeither reect the attributes
of only one actor or they can be combined forming a power distribution emotion
structure where all four conditions are determined. In these latter cases: (1) both
actors can have power or status; (2) one can have power or status advantage; or (3)
both actors can have powerlessness or statuslessness. In this manner, attributions
for Self and Other can be combined dening how power and status are distributed
in relations.
POWER/STATUS DISTRIBUTIONS
AND EMOTION
Four types of relations can be categorized simply in terms of how power or
status is disproportionately distributed between actors in a dyadic system. Such
distributions will be discussed next along with corresponding hypothesized
emotion labels and graphics.
Power-conict Relations and Emotion
Power conict relations exist when one actor has excessive power, or more power
then the other. It includes situations where one actor met expectations and provided
rewards to the other, but the other actor failed to meet expectations and return
rewards to the rst actor. Such imbalances in obligations incurring in transactions
produce differences in power relations (Blau, 1967, p. 140). Two types of power-
conict relations include situations where an actor has either power-advantage or
power-disadvantage.
A power-advantaged relation exists when power is attributed to Self (Self feels
manipulative) , but power is not attributed to Other (Self feels compassion for
Other) . Both of these substructures suggest that actors are undeserving, that is,
the actors sanctions are inconsistent with their performances. The hypothesized
emotion label for this syndrome is guilt .
Homans (1961, pp. 7476) has characterized guilt much the same as I have
hypothesized. He argued that when Self receives more rewards than Other , Self
is indebted and guilt results. Self has done better for himself then he ought to
have done and nds himself in a happy situation. Self has gotten more from
Other than Self has given . Kemper (1978, p. 51) also proposed that having
204 ROBERT THAMM
excess power in interaction predicts the emotion of guilt. His analysis of guilt
includes a negative self-evaluation , hurt to Other , wining too much at the
expense of Other , remorse and regret . The two interaction substructures are
symbolized in the diagonals of the paradigm . They represent an imbalance
in exchange. This imbalance is central in understanding the power-advantaged
emotion structure.
The compliment of guilt is when Self is power-disadvantaged. This is resolved
byaskinghowOther wouldfeel whenSelf feels guilt? It is achievedbyreversingthe
attributes for Self and Other. Self would then have met expectations but not received
rewards (Self feels powerless) while Other would have not met expectations
but received rewards (Self feels disgusted with Other) . When Selfs guilt is
projected onto Other, the complimentary hypothesized emotion is resentment .
In interaction, when one actor feels guilt the other actor would feel resentment.
This again assumes that both actors are appraising the situation the same.
The interaction substructures in the power-disadvantaged resentment syndrome
include, Other did not meet expectations and as a result Self was not rewarded,
while Self met expectations and as a result Other was rewarded. Resentment like
guilt has been widely discussed in the literature, but many times the anger com-
ponent is treated in more detail.
Homans (1961, pp. 7476) saw anger as the opposite of guilt and maintained
that when one actor receives less rewards than the other , and his rewards are
not proportionate to his costs, anger results . The actor has done worse for
himself then he ought to have done and nds himself in an unhappy situation.
He has gotten less than he has given . Anger in the paradigm is important in
understanding resentment, but equally important is the rewards Self provided to
Other, the generosity component .
Resentment can be characterized in terms of both generosity and anger as Self
gives too much to Other and receives too little. It implies that Other is indebted to
Self. It implies that Self didnt deserve being punished, just as guilt implies that Self
didnt deserve the rewards. In interaction, when one actor feels guilt (having power-
advantage), the other actor would feel resentment (having power-disadvantage).
Power-advantage and power-disadvantage are the two universal power-conict
substructures. As the four E/S outcomes are appraised in power-conict relations,
the universal emotions of guilt and resentment emerge.
It is rather apparent that the social exchange interaction substructures are of
central concern in understanding these universal power-conict syndromes and
emotions. They include the diagonals of regret and gratitude in the guilt
syndrome , and anger and generosity in the resentment syndrome . In
power-conict, exchange seems to be relevant and central because the interaction
diagonal signs are both consistent and effective.
Towards a Universal Power and Status Theory of Emotion 205
Status-conict Relations and Emotion
Next I would like to switch to status-conict relations and their corresponding
emotions. These syndromes exist when one actor has excessive status and the other
actor has none. This denition is close to what Homans (1961, p. 248) described
as status-incongruence. When such status inconsistencies exist, substructures of
deferring and demeaning prevail. Actors view each other as inferior or superior,
as having more pride or shame than the other, and as having more or less honor or
disgrace.
Scheff (1990, pp. 7195) introduced a deference-emotion system and argued
that shame and conformity as well as the resulting sanctions, were imperative to
its understanding. He maintained that conformity is encouraged by a system of
sanctions and that actors usually conform because they are likely to be rewarded,
andpunishedwhentheydonot. Durkheims analysis of social inuence (see Scheff,
1990, p. 95) also incorporates a deference-emotion system in which conformity to
norms is rewarded by deference and the feeling of pride, while non-conformity is
punished by lack of deference and feelings of shame.
These denitions closely parallel the attribution structures summarized above
in the E/S paradigm. Conformity-rewarded is dened as status , and deviance-
punished as statusless . Both of these substructures indicate that the actor is
deserving. That is, the actors sanctions are consistent with their performances
(Turner, 1998, p. 458). This is the case in the graphic sets where two actors have
different statuses. Self has higher status then Other or Other has higher status
then Self .
Here the interaction patterns are of particular interest because of their incon-
sistent signs, and . In the process of exchange, one actor meeting or
not meeting expectations is somewhat independent of the other actors sanction.
Thus one actors sanctions do not seem to follow from anothers performance.
In each of these cases an actor meeting expectations does not effect the other
actors punishment, or conversely, an actor not meeting expectations does not
effect the others reward. The implication of these graphic interactions (or lack
of interactions) is interesting in that they suggest a somewhat un-involvement
or detachment between actors of different statuses. Perhaps this is because a
certain social distance exists between actors in status-conict relations, preventing
effective giving and receiving of rewards.
Homans (1961, p. 254) held a similar position in arguing that actors tend to
avoid interaction with incongruent others. I came to a similar conclusion and
assume that status differences in the structure of social action tend to produce
emotions that are distancing or alienating in nature. I propose that in the course of
status-conict, actors become more independent and uncommitted to one another,
206 ROBERT THAMM
and that meaningful exchange (intimacy, attachment, attunement, etc.) between
them is diminished.
Assuming a reduced importance of interaction in understanding status-conict
relations, the performance distributions and reward distributions
become more central. But the most important overall comparison is contrasting
Selfs attributes to Others in assessing which actor has status-advantage in the
situation. Thus distribution and attribution structures seemto have more centrality
than interaction in evaluating the nature of universal status-conict relations
and emotions.
Conversely, in status-equality relations where mutual respect exists (feeling
contentment) , no conict is present and interaction is positive, and . And
when there is mutual disrespect (feeling resignation) , interaction is negative,
and . In both cases involvement is still occurring. The latter however is an
example of negative attachment. There is little conict in this kind of exchange
because of its balanced negative form of intimacy.
Status-advantaged Relations and Emotion. Two universal status-conict substruc-
tures include situations where Self has gained status-advantage (by demeaning
Other) or status-disadvantage (by deferring to Other). The demeaning substructure
is represented in the sequence when Self has honor and Other does not. Self looks
down on Other in a degrading manner. This kind of relation exists when status is
attributed to Self (Self feels honored) , but not attributed to Other (Self feels dis-
illusioned with Other) . The distribution substructure emotions include feeling
superior and feeling selsh . The emotion label for this status-advantaged
syndrome where Other is demeaned is Self feeling self-righteous .
Status-disadvantagedRelations andEmotion. The deferringstructure compliments
the demeaning structure and is represented in the sequence when Self does not have
honor but Other does. Self looks up to Other in a deferring manner. This emotion
structure is derived in reversing the self-righteous conditions. Other becomes Self
and Self becomes Other. More subjectively we might ask how Other would feel if
Self feels self-righteous. In reversing the attributes we can describe the deference
substructure as, Self did not meet expectations and was not rewarded while other
met expectations and received rewards. This kind of relation exists when status is
not attributed to Self (Self feels disgraced) , but is attributed to Other (Self feels
impressed with Other) . The distribution emotion components include feeling
inferior and excluded . The emotion category for this status-disadvantaged
deference syndrome is feeling intimidated .
In status relations, when one actor feels self-righteous the other actor should
feel intimidated. When one actor is status-advantaged, the other would be status-
disadvantaged. Status-advantage and status-disadvantage are the two universal
Towards a Universal Power and Status Theory of Emotion 207
status-conict substructures. Thus emotion categories vary as power and status
relations vary.
Proposition 6. Power-status attributions, distributions, and interactions are
components of universal social structures, and when appraised predict universal
complex emotions syndromes.
SOCIAL CHANGE AND EMOTION
I agree with Kemper (1978, p. 25) that any comprehensive theory of emotions
must have a model that embraces all the social interaction conditions that give
rise to emotions. He argued that power and status as structural attributes do not
usually come into being without power or status as process (Kemper, 1978, p. 172).
I hope to meet these challenges by adding a process dimension to the theoretical
paradigm. Specifying the universal conditions that predict emotions in the E/S
model fails to account for emotions that occur when social conditions are in the
process of change. I now turn to investigating the impact of social change on a
whole new grouping of emotions.
When a particular set of social structural conditions exist, actors in that social
system appraise and exhibit emotions that reect these states. If the substructure
conditions are stable, known, and certain, the appraisals and emotions of the actors
also tend to be stable and certain. Unchanging substructure conditions produce
unchanging emotion states.
But the transition of moving from one set of certain social structural conditions
and states to another is a social change process. This process involves moving from
old stability, to instability, and then to a newstability; fromcertainty, to uncertainty,
to certainty again. But it is this transition of instability and uncertainty that is the
essence of social change and process emotions.
As social change takes place and the structural conditions are being altered,
the emotion categories of actors are impacted. If we assume that stable conditions
predict stable kinds of emotions for actors then it follows that changes in the
conditions will produce corresponding changes in actors emotions. Changing
conditions produce emotions that are unstable because the anticipated outcomes,
either positive or negative, are still unknown. As actors interact, emotions ow
from these anticipated outcomes in different status and power relations (Kemper,
1978, 1984).
As social change produces emotion change, the unstable and uncertain
transitions inherent in the process itself can be dened and elaborated. They are
illustrated in the following temporal sequence of social events:
208 ROBERT THAMM
Although the transitions have some different properties than the more stationary
or structural emotions I have already discussed, they can be differentiated using
the same logic and procedures applied to the E/S paradigm.
Kemper (1978, p. 49) proposed three kinds of emotion categories that seem
to parallel the social change sequence illustrated above. They include structural,
anticipatory and consequent. The anticipatory emotions occur in the middle stage
when the social change states are uncertain. This anticipatory class of emotion
substructures will be referred to as transitional.
I propose that when social change is taking place the substructure conditions are
largely unknown or uncertain, and actors in appraising the changing conditions
have the emotion generally known as anxiety. The anxiety types include positive
(hopes), negative (fears), and neutral (where the conditions are changing but no
direction is anticipated).
Proposition 7. Appraisals of changing expectation/sanction conditions in
power and status relations predict corresponding universal power and status
anxieties.
It is important to note the distinction between changing conditions that
produce anxieties and unchanging or static outcomes that produce station-
ary emotions. Thus gaining or losing status or power advantage results in
different emotions than having power or status advantage. For example a status-
advantaged actor might have the feeling of self-righteousness and is self-centered
complacent and inactive. It is far-fetched to call this a form of change and
anxiety.
Power and status concepts can thus be conceived in three different modes
including, (1) an attribution of an actor (e.g. having power or status ),
(2) the distribution between two actors (e.g. having power-advantage or
having status-advantage ), or (3) a social change anxiety transition from
one distribution to another (e.g. an actor gaining power-advantage or gaining
status-advantage in the relation). The concept of coercion, for example is only
relevant in the third mode as one actor is using coercion as a mechanism to gain
power-advantage over the other. The third anxiety mode of reversing power and
status in relations will be addressed next.
Towards a Universal Power and Status Theory of Emotion 209
REVERSING POWER AND STATUS RELATIONS
The most complex anxiety emotions exist when all four conditions are changing at
the same time. Theyare the social change conditions that effect complete transitions
of power or status. In these cases power or status attributes for Self and Other are
reversed as the four-condition social change transitions are made.
The social change graphics are also more complex in that they represent
the transitions between pre-change and post-change conditions and states.
The transition graphics in each cell move either from plus-to-minus (+) or
minus-to-plus (+). The states in each E/S quadrant are now expressed with two
valences instead of one. In each graphic set the signs on the left in each quadrant
match the pre-change signs and the signs on the right match the post-change signs.
The four universal conict anxiety transitions representing power and status
reversals include: (1) Self gaining power-advantage ; (2) Self losing power-
advantage ; (3) Self gaining status-advantage ; and (4) Self losing
status-advantage . These transitions are competitive processes where actors
are gaining or losing power or status. Each will be examined as it intervenes
between the pre-change emotion set where all four conditions are known, and the
post-change set where the opposite four conditions are known.
REVERSING POWER IN RELATIONS
Theorists have provided different explanations of how actors gain and lose power
in social relations. Blau, for example, suggested that power can be obtained by
threatening to deprive others of benets they currently enjoy, unless they submit
(Blau, 1967, p. 118). Other theorists including Tawney (1931, p. 229), Weber
210 ROBERT THAMM
(1947, p. 152), and Kemper (1978, pp. 2629) argue that power is necessarily an
interaction concept that involves Self forcing or coercing Other to provide rewards
for Self. Kemper (1978, pp. 2829) makes a case for the necessity of the concept
of involuntary compliance in understanding power. For him, and other theorists,
it is important to recognize how Others compliance is gained through Selfs
coercive actions.
Using only the E/S dimensions, involuntary compliance could be interpreted
simplyas Other complyingas a result of Self not complying. Selfs non-compliance
may take on many forms but the assumption here is that whatever means is used it
will result in Other providing some reward to Self. Thus the power-disadvantaged
Other reluctantly and involuntarily gives up rewards while providing rewards for
the power-advantaged Self. Thus gaining power-advantage is, Self not meeting
expectations but receiving rewards as a result of imposing ones will on Other to
gain the Others compliance and rewards. Others involuntary compliance is seen
as a direct function of Selfs non-conforming or deviant actions that causes Other
to unwillingly provide rewards for Self. Theorists dene these kinds of actions as
inuential or coercive or persuasive power.
But power refers to all kinds of controlling inuences (Blau, 1967, p. 115).
Examples of power range from Self persuading Other to cooperate in some
activity to a more extreme case of Self using violent force. It might be instructive
to suggest a continuum of control ranging from the most passive-aggressive
mechanisms to overt violence. The scale might include hinting, suggesting,
complaining, nagging, persuading, convincing, inducing, tricking, mandating,
demanding, threatening, coercing, forcing, and terrorizing, just to name a
few. Distinctions among these kinds of power-gaining mechanisms have been
widely discussed in the literature (Blau, 1967, p. 116; French & Raven, 1959,
pp. 150167; Parsons, 1963, pp. 4345).
When the threat of negative sanctions becomes so weak that it can be interpreted
as neutral or positive, power ceases to exist. Once there is no resistance on the
part of Other to comply with Selfs demands, Self no longer is deviant and
Selfs non-compliance changes to compliance, and thus to status. To promise
punishment for failing to comply is a power mechanism. However, to promise
rewards for voluntary compliance becomes a status process.
Reversing power positions in the dyad involves actors changing their appraisals
of who is undeserving of the rewards and who is undeserving of the punishments. It
involves changing evaluations of injustice in the changing social situation. If actors
are about to better their performances, they are also about to decrease their rewards.
If their performances are poorer, their rewards will be increased proportionately.
Thus changing power-conict attribution structures are both unjust and undeserv-
ing. As power attributes of actors are reversed, universal power anxieties emerge.
Towards a Universal Power and Status Theory of Emotion 211
These universal anxieties exist in the process of an actor gaining or losing power-
advantage. They involve competitive processes in that one actor is either gaining or
losing control over the other. Gaining control over another is an aggressive activity
producing an offensive anxiety, while losing control is a passive activity producing
a defensive anxiety.
Proposition 8. Offensive and defensive anxieties emerge as actors gain and
lose advantage in universal power-conict relations.
Gaining Power-advantage in Relations:
The Offensive Anxiety
During this anxiety process, Self acts offensively to gain (or regain) power-
advantage in the relation.
Pre-change emotion states: Prior to having the anxiety, Self had a power-
disadvantage. Self had met expectations and provided rewards to Other while
Other failed to meet expectations and return rewards to Self. Self had high costs
(contributions) and low prots (retributions). Under these conditions Self had the
emotion resentment .
Process motive: Selfs offensive motive in this process involves a hope to gain
power-advantage by controlling Other so that Other reluctantly (unwillingly) trans-
fers rewards to Self. Selfs intent is to use aggressive means to reverse the power
distribution in the relation. Perhaps the ultimate intent of gaining power is not to
get caught perpetuating an injustice for fear of retribution.
Formal transition: The power-gaining process is illustrated below:
As the transition begins, Self develops an offensive anxiety. It is formally
interpreted as, Self is about to not meet expectations (condition 1), controlling
Others reluctant compliance in meeting Selfs expectations of Other (condition
3). Other begins to provide rewards to Self (condition 2) and in the process Other
loses rewards (condition 4).
Process dynamics: More descriptively, the interaction is about Self applying
devious tactics in threatening harm to Other if Other fails to comply with Selfs
expectations of Other. Other then reluctantly complies providing Self with rewards.
Also in the process, Other is losing rewards as a result of Selfs original threat.
212 ROBERT THAMM
Post-change emotion outcomes: The process is completed and the anxiety
is diminished when the outcomes on all four conditions are reversed. Selfs
power-disadvantaged relation changes to Self having power-advantage. In terms
of the paradigm, Self did not meet expectations and failed to reward Other,
while Other met expectations and provided rewards to Self. Self now has low
costs (contributions) and high prots (retributions). In reversing these conditions
Self would then have the power-advantaged emotion of guilt. , and Other the
power-disadvantaged emotion of resentment .
Social content: The content of this offensive anxiety may include Self using
devious activities ranging from subtle hinting, to coercive threats, to outright
physical attacks; from passive aggression, to manipulation involving lying,
cheating and stealing, to giving ultimatums, to hostile uses of force. Any of these
tactics could be used to gain Others compliance as a result of Selfs direct or
indirect threatening intentions. Examples could range from subtle game-playing
to violent rape where Self is the perpetrator and Other the victim.
Losing Power-advantage in Relations:
The Defensive Anxiety
During this anxiety process, Self acts defensively in resisting Others attempt to
gain (or regain) power-advantage in the relation.
Pre-change emotion states: Prior to having this anxiety Self had a power-
advantage. Self had not met expectations and failed to provide rewards to Other
while Other met expectations and returned rewards to Self. Self had low costs
(contributions) and high prots (retributions). Under these conditions Self had the
emotion guilt .
Process motive: Selfs defensive motive in this process involves a fear of losing
power-advantage by resisting Others controlling actions so that Self reluctantly
(unwillingly) transfers rewards to Other. Selfs intent is to use passive resistance
in the reversal of power distribution in the relation.
Formal transition: The power-losing process is illustrated below:
As the transition begins, Self develops a defensive anxiety. It is formally
interpreted as, Other is about to not meet expectations (condition 3), controlling
Selfs reluctant (unwilling) compliance in meeting Others expectations of Self
Towards a Universal Power and Status Theory of Emotion 213
(condition 1). Self begins to provide rewards to Other (condition 4) and in the
process Self loses rewards (condition 2).
Process dynamics: More descriptively, the interaction is about Other applying
devious tactics in threatening harm to Self, if Self fails to comply with Others
expectations of Self. Self then reluctantly begins to comply providing Other with
rewards. Also in the process, Self is losing rewards as a result of Others original
threat.
Post-change emotion outcomes: The process is completed and the anxiety
is diminished when the outcomes on all four conditions are reversed. Selfs
power-advantage changes to Self having power-disadvantage. In terms of the
paradigm, Other did not meet expectations and failed to reward Self, while
Self met expectations and provided rewards to Other. Self now has high costs
(contributions) and low prots (retributions). In reversing these conditions Self
would then have the power-disadvantaged emotion of resentment , and Other
the power-advantaged emotion of guilt .
Social content: The content of this defensive anxiety may include Self resisting
devious aggressions by Other, ranging from subtle hinting, coercive threats, to
outright physical attacks; frompassive aggression, tomanipulationinvolvinglying,
cheating and stealing, to receiving ultimatums, to hostile uses of force. Any of
these tactics could be used to gain Selfs compliance as a result of Others direct
or indirect threatening intentions.
Now lets briey examine the case where there is a power struggle between two
actors, and neither of themis wining the competition. Both actors are using deviant
tactics to gain the others compliance. As the process continues, the losing actor
feels defensive and the winning actor feels offensive. The one that eventually wins
is successfully deviant, and the loser is unsuccessfully compliant. Following the
struggle, the winner feels guilt and the loser feels resentment.
The anxiety emotions of feeling offensive and defensive are familiar to people
in all cultures, but a cycle of gaining or losing control in social relations can
potentially become chronic and abnormal. In any case, such processes are common
and universal, and the changing power-conict relations when appraised predict
universal offensive and defensive anxiety emotions.
REVERSING STATUS RELATIONS
Status traditionally has been dened as the relative importance of the position an
actor has in the group. In dyadic interaction it might be dened more in terms of
howmuch respect an actor accords another. Actors assign levels of respect for both
themselves and others and communicate this assessment to the group. As these
214 ROBERT THAMM
assessments change, more communication is necessary to establish the new status
rankings.
Exchange of sanctions is not central in these processes because status is
conferred more formally through the assignment of respect. When an actor raises
anothers status, the actor does not offer rewards or expect rewards in return,
but simply acknowledges and announces the performance ranking and deserving
rewards that the other has previously obtained.
Status is sometimes seen as one actor conferring by giving rewards to the
other and the other receiving the rewards (e.g. Kemper, 1978, p. 35). I would
disagree with this but only in status-conict relations. Conferral in these cases of
status-incongruent relations pertains only to the assessment of each actors level of
respect, and not to what each is giving to the other in interaction. In status-conict
relations, exchange is secondary to the different assignments of worth or respect
for past performances and past deserving sanctions. The attributions dominate
the interactions.
The difference between giving rewards and conferring rewards is that
giving implies that the receiver does not already have the rewards but conferring
implies that the conferee already has rewards. The conferral process involves just
acknowledging and announcing them. For example, a boss might provide a contin-
gency where a worker would receive rewards if they performed in a certain manner.
Fromthe bosss perspective, it is not relevant howthe worker received the rewards,
only that they were forthcoming. Only after the worker met the expectations and
received rewards would the boss confer higher status on the worker.
Voluntary vs. Involuntary Compliance. I however agree with Kemper (1978,
pp. 2830) in making a case for the necessity of the concept of voluntary com-
pliance in understanding status-conict relations. He argued that in a dyad, if
Self has higher status than Other, this means that Self receives more in the way of
voluntary compliance from Other than Other receives from Self. Kemper seems to
be dening status relations in terms of how much compliance one actor receives
from another. The key word here is receives.
If receiving more in the way of voluntary compliance means actors receive
more because they complied, or actors receive less because they failed to comply,
this is an attribution or status concern. When an actor complies voluntarily, rewards
will follow and when an actor fails to comply voluntarily, rewards will not follow.
What is voluntary in status relations is each actor dening and communicating
status placements in the group.
To clarify this, all voluntary acts are complying but not all complying acts
are voluntary. In power-conict relations, one actor voluntarily does not comply
while the other involuntarily complies, while in status-conict relations, one actor
Towards a Universal Power and Status Theory of Emotion 215
voluntarily complies and the other actor voluntarily does not comply. It is more
clarifying to suggest that the higher-status actor voluntarily complies and the lower-
status actor voluntarily fails to comply with expectations. Unlike power, status
attributes are always voluntarily assigned.
But voluntary compliance is not equated with high performance. What is
voluntary is the acknowledgement and announcement that Self has lesser respect
than Other (had lower performances and deserved lesser rewards), or has greater
respect than Other (had higher performances and deserved more rewards). In all
status relations, the deserving sanction for an actor therefore consistently follows
from that actors performance.
Status placement is thus a function of each actors assessment of the situation
somewhat independent of what the other actor wants or desires. What is more
relevant in distinguishing power from status is how Self appraised the extent to
which Self controlled Others compliance.
Think of a boss and a worker. The former holds out the promise of a raise if the
worker does something he doesnt want to do. It is the compliance that is appraised
by the boss regardless of whether the worker wants to comply voluntarily or not.
If the boss appraises the worker as having complied voluntarily and received the
raise, the boss accords status to the worker. If on the other hand, the boss appraises
the worker as complying involuntarily, the boss could attribute this to his own
deviance and control over the workers compliance, a power mechanism. Whether
the situation is a power or status concern is determined not by the Others appraisal
of their voluntary or involuntary compliance, but by Selfs interpretation of the
extent to which Others compliance was controlled.
Deferring and Demeaning. Status refers to all kinds of respecting and disre-
specting inuences (Blau, 1967, p. 115). In status relations, actors confer high
or low status attributes on themselves and others. Self conferring higher status
on Other is a deferring process. A deference-emotion substructure involves
communications of respect and honor, and enables actors to make known to each
other their evaluations of each others status (Scheff, 1988).
On the other hand, the demeaning-emotion substructure involves communi-
cations of disrespect and dishonor. And actors are extremely sensitive to the
exact amount of deference they receive as to gestures of respect and disrespect
(Goffman, 1967). So status honor is not a xed attribute but a uid process,
continuously being revised and contested (Scheff, 1990, p. 28). These distinctions
are important in understanding the processes that take place in reversing status
positions in dyadic relations, and in the resulting anxieties.
Reversing Status-conict Relations. Reversing status positions in the dyad involves
actors changing their appraisals of who deserves the rewards and who deserves the
216 ROBERT THAMM
punishments. They involve changing evaluations of justice in the changing social
situation. As status attributes of actors are reversed, just and deserving universal
status-conict anxieties emerge.
Universal anxieties exist in the process of an actor either gaining status-
advantage over another, or losing status-advantage. These are competitive status
processes in that each actor is either gaining or losing respect for the other. Gaining
respect over another is an aggressive activity producing an oppressive anxiety,
while losing respect is a more passive activity producing a depressive anxiety.
Proposition 9. Oppressive and depressive anxieties emerge as actors gain and
lose advantage in universal status-conict relations.
Gaining Status-advantage in Relations: The Oppressive Anxiety
During this anxiety process, Self acts oppressively to gain (or regain) status-
advantage in the relation.
Pre-change emotion states: Prior to having the anxiety, Self had a status-
disadvantage. Self had not met expectations and was not provided with rewards
while Other met expectations and received rewards. Self had lower performance
and lower rewards. Under these conditions Self had the emotion intimidation .
Process motive: Selfs oppressive motive in this process involves a hope to
gain status-advantage by dominating Other so that Other willingly concedes to
inferiority and defeat. Selfs intent is to use condescending means to reverse the
status distribution in the relation. Perhaps the ultimate intent of status seeking is
to get recognition for past contributions.
Formal transition: The status-gaining process is illustrated below:
As the transition begins, Self develops an oppressive anxiety. It is formally
interpretedas, Self is about tomeet expectations (condition1), andbegins toreceive
rewards (condition 2), while Other is not meeting Selfs expectations (condition
3) and, in the process, is losing rewards (condition 4).
Process dynamics: More descriptively, the relation is about Self using deferring
tactics in appreciating Self and demeaning tactics in depreciating Other. Self is
complying and receiving rewards while Other is not complying and is not being
rewarded. In the process, Other is losing respect and Self is gaining respect.
Post-change emotion outcomes: The process is completed and the anxiety
is diminished when the outcomes on all four conditions are reversed. Selfs
Towards a Universal Power and Status Theory of Emotion 217
status-disadvantaged relation changes to Self having status-advantage. In terms of
the paradigm, Self now has meet expectations and received rewards, while Other
has not met expectations and did not receive rewards. Self has high respect (Self
attributions) and Other has low respect (Other attributions). In reversing these
conditions Self then has the status-advantaged emotion of self-righteous , and
Other the status-disadvantaged emotion of intimidation .
Social content: The content of this oppressive anxiety may include activities
ranging from Self shaming, belittling, demoting, demeaning, depreciating, or
rejecting Other, to an outright use of humiliation. All of these tactics are designed
to conrm Selfs emerging victorious superiority in the relation. Although there
may be a degrading sympathy for Other, it is secondary to Selfs motive to gain
honor, prestige or esteem, and to dominate the status relationship.
Losing Status-advantage in Relations: The Depressive Anxiety
During this anxiety process, Self acts depressively in losing status-advantage in
the relation.
Pre-change emotion states: Prior to having the anxiety, Self had a status-
advantage. Self had met expectations and was provided with rewards while Other
did not meet expectations and did not receive rewards. Self had higher performance
and higher rewards. Under these conditions Self had the emotion self-righteous .
Process motive: Selfs depressive motive in this process involves a fear of losing
status-advantage by subordinating to Other, so that Self willingly concedes to
inferiority and defeat. Selfs intent is to use Self-demeaning evaluations to reverse
the status distribution in the relation.
Formal transition: The status-losing process is illustrated below:
As the transition begins, Self develops a depressive anxiety. It is formally in-
terpreted as, Self is about to not meet expectations (condition 1), and begins to
lose rewards (condition 2), while Other is meeting Selfs expectations (condition
3) and, in the process, is gaining rewards (condition 4).
Process dynamics: More descriptively, the interactions are about Self using
demeaning tactics in depreciating Self and deferring tactics in appreciating Other.
Self is not complying and not receiving rewards while Other is complying and
is being rewarded. In the process, Other is gaining respect and Self is losing
respect.
218 ROBERT THAMM
Post-change emotion outcomes: The process is completed and the anxiety
is diminished when the outcomes on all four conditions are reversed. Selfs
status-advantaged relation changes to Self having status-disadvantage. In terms
of the paradigm, Self has not met expectations and has lost rewards, while Other
has met expectations and received rewards. Self has low respect (Self attribution)
and Other has high respect (Other attribution). In reversing these conditions Self
would then have the status-disadvantaged emotion of intimidation , and Other
the status-advantaged emotion of self-righteous .
Social content: The content of this depressive anxiety may include activities
ranging from self-shaming, self-belittling, self-demoting, self-demeaning, self-
depreciating, and self-rejecting to outright self-humiliation. All of these tactics
are designed to conrm Selfs emerging defeat and inferiority in the relation.
Although there may be a subtle admiration for Other, it is secondary to Selfs
intention to dishonor oneself, to lose the prestige or esteem from Other, and to be
subordinate in the status relationship.
Now lets briey examine the case where there is a status struggle between
two actors and neither of them is wining the competition. Both actors are
using demeaning tactics to diminish the others compliance and rewards. In the
process, the losing actor is feeling depressive and the winning actor is feeling
oppressive. The one that eventually wins is successfully compliant, and the loser
is unsuccessfully deviant. Following the struggle, the winner feels self-righteous
and the loser feels intimidated.
The anxiety emotions of feeling oppressive and depressive are familiar to people
in all cultures, but a cycle of gaining or losing respect in social relations can
also potentially become chronic and abnormal. In any case, such processes are
common and universal, and changing status-conict relations when appraised
predict universal oppressive and depressive anxiety emotions.
DISCUSSION
Further elaboration of additional universal anxiety emotions would require more
space than alloted here. It was my intent only to outline the logic and rationale
for integrating power and status dimensions into the E/S Paradigm in predicting
universal human emotions. The framework presented is limited to conict
situations. Consensus relations and partial transitions changing power relations
to status, or visa versa, are all but excluded.
For example, lets briey examine the case between two actors in a status conict
relation. Bill has higher status and Monica has lower status. Bill would feel self-
righteous and Monica would feel intimidated . Now consider two scenarios
Towards a Universal Power and Status Theory of Emotion 219
where only two of the four conditions are being altered. In the rst case Bill asks
or requests Monica to provide sexual favors for him. If she voluntarily complies
and begins to meet Bills expectations, it brings rewards to both Bill and Monica.
They are both into the exchange. All four conditions become positive and both
actors would feel contentment , as they each make positive contributions to the
other and receive positive retributions from the other. The status-conict relation
changes to a status-consensus relation.
Asecond scenario begins with the same status-conict relation between Bill and
Monica. In this case, however, Bills request for sexual favors is rejected. Bill then
exceeds legitimate authority inherent in his higher status position, and turns to
illegitimate or devious means to gain Monicas compliance. He might use threat,
coercion, or even physical force to gain her sexual favors. His previous compliance
changes to deviance and he begins to gain rewards for himself, but certainly not
for her. The sanctioning signs remain unchanged as he increases his rewards and
she continues to lose rewards.
In this case the expectations signs are reversed as his deviant actions force
her compliance. This would change Bills status-advantage in the relation to a
power-advantage and Monica would move from a status-disadvantage to a power-
disadvantage. Bill would then feel guilt and Monica would feel resentment .
Similar scenarios could be elaborated between any two actors of different statuses,
such as a military ofcer over a soldier, a guard over a prisoner, or a boss over an
employee.
Thus relations could change from conict to concensus, from concensus to
conict, from power to staus, or status to power by altering only two of the four
conditions. Numerous other universal transitions and corresponding emotion
categories require further examination and elaboration.
There is also the possibiity of changing normal structures and anxieties into
abnormal structures and anxieties. This would entail, for example, a formal
conversion of the guilt-to-resentment transition to a guilt-to-persecution transition,
and of the intimidation-to-self-righteous transition to intimidation-to-granduer.
Dilusions of persecution (abnormal power anxiety) and granduer (abnormal
status anxiety) would theoretically occur when actors escape social reality and
subjectively create and appraise imaginary rather than real social situations.
This would include situations where they imagine another is threatening them
(defensive persecution dilusions) or where they imagine that they have gained
superioriity over others (oppressive grandiosity dilusions).
Power disorders inclluding schitzophenic paranoia (panic-offensive), status
disorders including bi-polar (manic-depressive), and defensive mechanisms
including denial, displacement, suppression, repression, projection, introjection
and reaction-formation can all be elaborated in terms of the substructures outlined
220 ROBERT THAMM
above. The expectation/sanction paradigm is potentially productive in formalizing
many such abnormal anxieties. Turner (2002, p. 89) has suggested that negative
sanctions and failures to meet expectations activate defense mechanisms, but just
how this defense regime works is still a great mystery. The application of the E/S
model to psychoanalitic theory is challenging.
Another future task involves a question asked by Kemper (1987). How many
emotions are there? The model presented suggests that there are as many emotions
as there are universal structural variations. The formal mapping of these emotion
structures and labels into a comprehensive classigication system is possible.
Perhaps the taxonomy could resemble the periodic chart of elements, using E/S
states instead of protons, neutrons and electrons.
Much research is needed. Empirical support of another kind is suggested by
Kemper, where precisely any emotion labeled is shown to be the result of a given
expectation-sanction outcome. Until then, the labeling results and hypotheses
presented above are supported weakly and certainly not persuasively.
Cross-cultural comparisons are another priority. I am in the process of
analyzing data comparing Vietnamese, Russian and American samples in testing
cultural differences in the emotion labeling of many substructure variations
discussed above. Other research could examine the link between universal social
structures and emotion facial expressions or between universal social structures
and physiological responses. In any case, cross-cultural data will ultimately
provide evidence testing the validity or invalidity of the universal assumptions and
postulates proposed.
I have tried to provide a parsimonious theoretical model containing the least
amount of dimensions to explain the greatest number of emotion categories. But
any proposal must be tentative, be further tested by the methods of science, and
provide satisfying explanations of the social nature of human emotions.
REFERENCES
Abelson, R. P. (1983). Whatever became of consistency theory? Personality and Social Psychology
Bulletin, 9, 3754.
Averill, J. R. (1980). Aconstructivist viewof emotions. In: R. Plutchick &H. Kellerman (Eds), Theories
of Emotion (pp. 305339). New York: Academic Press.
Barbalet, J. M. (1998). Emotion, social theory, and social structure. Cambridge: Cambridge University
Press.
Blau, P. (1964). Exchange and power in social life. New York: Wiley.
Collins, R. (1975). Conict sociology: Toward an explanatory science. New York: Academic Press.
French, J. R. P., & Raven, B. H. (1959). The bases of social power. In: D. Cartright (Ed.), Studies in
Social Power. Ann Arbor: University of Michigan Press.
Towards a Universal Power and Status Theory of Emotion 221
Goffman, E. (1967). Interaction ritual. Garden City: Anchor Doubleday.
Harre, R. (Ed.) (1986). The social construction of emotions. Oxford: Blackwell.
Heelas, P. (1986). Emotion talk across cultures. In: R. Harre (Ed.), The Social Construction of Emotions.
Oxford: Blackwell.
Heise, D. (1979). Understanding events: Affect and the construction of social action. Cambridge:
Cambridge University Press.
Hochschild, A. R. (1975). The sociology of feeling and emotion. In: M. Millman &R. M. Kanter (Eds),
Another Voice (pp. 280307). New York: Doubleday.
Hochschild, A. R. (1979). Emotion work, feeling rules, and social structure. American Journal of
Sociology, 85, 551575.
Hochschild, A. R. (1983). The managed heart. Chicago: University of Chicago Press.
Homans, G. (1961). Social behavior: Its elementary forms. New York: Harcourt.
Kelley, H. H. (1984). Affect in interpersonal relations. In: P. Shaver (Ed.), Review of Personality and
Social Psychology. Beverly Hills: Sage.
Kemper, T. D. (1978). A social interactional theory of emotions. New York: Wiley.
Kemper, T. D. (1987). How many emotions are there? Wedding the social and autonomic components.
American Journal of Sociology, 93, 263289.
Kemper, T. D., & Collins, R. (1990). Dimensions of microinteraction. American Journal of Sociology,
96, 3268.
Lazarus, R. (1984). In the primacy of cognition. American Psychologist, 39, 124129.
Levy, R. I. (1984). The emotions in comparative perspective. In: K. R. Scherer & P. Ekman (Eds),
Approaches to Emotions. Hillsdale: Lawrence Erlbaum.
Ortony, A., Clore, G. L., & Collins, A. (1988). The cognitive structure of emotions. Cambridge: Cam-
bridge University Press.
Osgood, C., Suci, G., & Tannenbaum, P. (1954). The measurement of meaning. University of Illinois
Press.
Osgood, C. E., May, W. H., & Miron, M. S. (1975). Cross-cultural universals of affective meaning.
Urbana: University of Illinois Press.
Parsons, T., & Shils, E. A. (1951). Towards a general theory of action. New York: Harper &
Row.
Schachter, S., & Singer, D. (1962). Cognitive, social and physiological determinants of emotional
states. Psychological Review, 69, 379399.
Scheff, T. J. (1979). Catharsis in healing, ritual, and drama. Berkeley and Los Angeles: University of
California Press.
Scheff, T. J. (1988). Shame and conformity: The deference-emotion system. American Sociological
Review, 53, 395406.
Scheff, T. J. (1990). Microsociology. Chicago: University of Chicago Press.
Shott, S. (1979). Emotion and social life: A symbolic interactionist analysis. American Journal of
Sociology, 84, 13171334.
Skinner, B. F. (1938). The behavior of organisms. New York: Appleton-Century-Crofts.
Smith-Lovin, L., & Heise, D. (1988). Analyzing social interaction: Advances in affect control theory.
New York: Gorden and Breach Science Publishers.
Tawney, R. H. (1931). Equality. London: Allen & Unwin.
Thamm, R. (1975). The social foundations of emotion. In: Beyond Marriage and the Nuclear Family.
San Francisco: Caneld Press (Harper & Row).
Thamm, R. (1992). Social structure and emotion. Sociological Perspectives, 35, 649671.
Thibaut, J. W., & Kelley, H. H. (1959). The social psychology of groups. New York: Wiley.
222 ROBERT THAMM
Thoits, P. A. (1985). Self-labeling processes in mental illness: The role of emotional deviance. American
Journal of Sociology, 91, 221249.
Turner, J. H. (1998). The structure of sociological theory. Belmont: Wadsworth.
Turner, J. H. (2002). Face to face: Towards a sociological theory of interpersonal behavior. Stanford:
Stanford University Press.
Weber, M. (1947). The theory of social and economic organization. New York: Oxford.
THE DIFFERENTIAL IMPACT OF
EMOTIONS ON RATIONAL SCHEMES
OF SOCIAL ORGANIZATION:
READING WEBER AND COLEMAN
Theodore D. Kemper
ABSTRACT
Although the rational and the emotional are often thought to be in conict,
this is not always the case. Here I examine two instances, Max Webers ideal-
typical depiction of bureaucracy and James Colemans proposal for a rational
reconstruction of societal institutions. In the case of Weber, it is clear that the
disciplined, steady and affectless performance of ofcial duties by bureau-
crats can only be possible because of an underlying foundation of emotions,
both positive and negative. In the case of Colemans proposal, which is based
on money bounties as incentives for performing important societal tasks, a
multitude of deleterious and defeating emotions inhere in this ultra-rational
scheme.
INTRODUCTION
The struggle for social theory is to a signicant extent a contest over how to
dene human nature and what that nature enables us to achieve in social life. On
one important aspect of this question we are of two minds in the West, and this
gives rise to periodic pendulum swings of thought: from Apollonian to Dionysian,
Theory and Research on Human Emotions
Advances in Group Processes, Volume 21, 223244
2004 Published by Elsevier Ltd.
ISSN: 0882-6145/doi:10.1016/S0882-6145(04)21009-8
223
224 THEODORE D. KEMPER
Classical to Romantic, Ideational to Sensate in essence, the contest between the
rational and the emotional. History pushes us toward one or another doctrine by
recording the follies and exhaustion of the previous age, allowing the alternative
theory to begin its course to apogee, succeeded inevitably by decline. In the present
period, history has roared decisively on the question of how to organize economic
and political life and this has a particular resonance for the old sparring match
between rationality and emotion.
In effect, the collapse of the Marxist vision in the early 1990s has given
tremendous new impetus to the concept of the market, within which producers
seek their own interest. According to the principles propounded by Adam Smith,
and, as the evidence substantially conrms, the market model is vastly superior in
productivity, innovation, and rationality to competing command and control ap-
proaches. Not surprisingly, market theory, which assumes rational, self-interested
behavior in the economic context, has invaded institutional domains ostensibly
far removed from the marketplace. The rational calculation of incentives and
advantages has been applied to explain love, marriage and family planning
(Becker, 1981), educational attainment (Coleman, 1993), criminal behavior and
its deterrence (Tittle, 1980), military planning, both strategic and tactical (Brams,
1985), political choice (Overbye, 1995), the protection of the environment
(Coleman, 1986) and other so-called public goods, and more (see Hechter &
Kanazawa, 1997).
For the most part, rational models of social organization omit emotions. They are
in fact considered noise or disturbance. Weber (1968) was virtually archetypical
here in his distinction between the path of instrumentally rational action (Zweck-
rational) and the path action would take if its rationality were not diverted by
emotion:
For the purpose of a typological analysis it is convenient to treat all irrational, affectually
determined elements of behavior as factors or deviations from a conceptually pure type of
rational action. For example, a panic on the stock exchange can be most conveniently analyzed
by attempting to determine rst what the course of action would have been had it not been
inuenced by irrational affects; it is then possible to introduce the irrational components as
accounting for the observed deviations from this hypothetical course. Similarly, in analyzing a
political or military campaign it is convenient to determine in the rst place what would have
been a rational course, given the ends of the participants and adequate knowledge of all the
circumstances. Only in this way is it possible to assess the causal signicance of irrational
factors as accounting for the deviations from this type (p. 6).
Here Weber takes the decisive plunge into the camp that declares that emotions
the irrational factors are antagonistic to rational action. Of course, they sometimes
are exactly that (as we will see in the case of Coleman examined below), but Weber
seems to rule out any possibility that emotions may make rational action possible
The Differential Impact of Emotions on Rational Schemes 225
in the rst place and that rationalized social organization may depend on emotions
for its viability and success. Although Weber is generally scant on discussing
emotions in the entire corpus of his work (Elster, 2000), he is not consistent on
this question, as we will see in part one of this chapter where I examine Webers
ideal-typical depiction of bureaucracy.
In the second part of this chapter, I will examine a proposal by James Coleman
for a rational form of social organization that is clearly premised on the exclusion
of emotion and ts very much into the spirit of Webers position quoted above.
My stance will be that Colemans failure to consider the emotional consequences
of his plan is, in fact, a fatal aw.
THE EMOTIONAL FOUNDATIONS OF
SINE IRA AC STUDIO
Although Weber is clearly aware of the presence of emotion in social action (see
Albrow, 1992; Elster, 2000) after all, one of his categories of action is Affektuel
which is emotion, out-and-out, while another, Wertrational, is at least partly so it
is mildly ironic that Weber does not examine one of his hallmarks of rationalization,
namely bureaucracy, for the emotions that sustain and make bureaucracy possible.
For present-day students of social organization, a more inclusive and transparent
view is desirable. In his classic statement on bureaucracy, Max Weber (1958, pp.
215216) wrote:
When fully developed, bureaucracy also stands in a specic sense, under the principle of sine
ira ac studio [without passion or bias]. Its specic nature, which is welcomed by capitalism,
develops the more perfectly the more the bureaucracy is dehumanized, the more it succeeds in
eliminating fromofcial business love, hatred, and all purely personal, irrational, and emotional
elements which escape calculation.
Although, as Albrow (1992) argues, this affectlessness is not necessarily realized
in actual practice, it is one component of the ideal type of bureaucracy or orga-
nization, according to Weber. Indeed, the emotionless conduct of ofcial business
as part of the fully developed bureaucratic mechanism compares [in purely tech-
nical superiority over the practices] of other organizations as does the machine
with the non-mechanical modes of production (Weber, 1946, p. 214). Weber did
not endorse this super-rationalization, but attempted to record a tendency toward
ofcial practice that was peculiarly modern.
I contend here that a signicant paradox underlies the creation of the affect-
less functionary whose ofcial acts are impersonal and dehumanized, namely
that only a powerful emotional nexus to the organization makes such emotionless
226 THEODORE D. KEMPER
action possible. The evidence for this contention is found in Webers own dis-
cussion. In this regard, I do not argue that Weber directly contradicted himself.
Rather, there appears to be a curious failure, with scant exception, to appreciate
that even rationally-conceived and rationally-purposed social organizations gen-
erate emotions in their members, and that these emotions enable the human actors
to participate genuinely in the ongoing social life that is ordered by the social
organizational form.
Although sociologists clearly recognize the emotional underlay in organizations
and institutions substantively oriented toward the arousal or processing of emo-
tions, such as the family or religion, they have until recently (see Fineman, 1993)
blinked this recognition in specically bureaucratic organizations and institutions.
With respect to emotions, the concerns of members, as distinct from the purposes,
of bureaucracies are no different from those in other social groups.
Nor is it that Weber was entirely insensitive to how affect is generated in social
life. Indeed, the pivotal claim in The Protestant Ethic and the Spirit of Capitalism
(Weber, 1930) is that certain religious ideas specically the Calvinist doctrine of
Predestination would be likely to induce a certain emotion, specically, anxiety
concerning whether one were destined for eternal salvation. According to Weber,
that anxiety was crucial for the development of the ethos of capitalism in its early,
entrepreneurial form. Yet, the same type of interpretive understanding of the emo-
tional consequences of bureaucratic organization is notably absent from Webers
discussion.
To extract the emotional elements in Weber, I will employ Webers own method
of Verstehen, using as a major interpretive guide the theory of emotions in Kemper
(1978) and Kemper and Collins (1990). This will permit an empathic apprehension
of the position of the ofcial for whomsine ira ac studio is prescribed, but is likely
to be realized only when emotional considerations are in virtually all other ways
paramount for the ofcial.
Fortunately for present purposes, Weber examined the activities of bureaucratic
ofcials in two respects: in regard to clients (or, as Weber frequently writes, the
governed,) with whom affectless decision-making is enjoined; and in regard to
the chief, or master, and other structural features of the bureaucratic apparatus.
Weber weaves his argument back and forth between these two foci. This allows
us to see that a less emotional orientation toward the clients of bureaucracy is not
necessarily accompanied by a reduction of emotion in respect to the master, or to
other elements of the bureaucratic structure.
According to Weber, the achievement of the modern form of affectless admin-
istration depends on a number of factors. These are supposed to permit the ofcial
to act impersonally, in accordance with existing rules or orders, indifferent to his
or her own desire or preference. In terms that Weber himself employs, action
The Differential Impact of Emotions on Rational Schemes 227
is disciplined rather than principled. Yet, paradoxically, the structural conditions
that presumably make it possible for ofcials to act in a disciplined way, without
emotion, are themselves emotion-laden at the core. Indeed, were they not, it is
difcult to see how they could exercise so potent an effect, namely to excise the
personal and emotional grounds of decision-making. (In examining the emotional
grounds for emotionless action, I do not take a stand on whether emotionless action
is possible, and to what degree. I assume that modern bureaucratic decision-making
differs importantly from Kadi justice.
1
But whether the ideal-typical sine ira ac
studio criterion can ever be fully achieved is not an issue here.) I turn now to the
emotional bases of bureaucratic affectlessness.
Employment Security
Again and again, Weber (1958) stresses the trade-off in modern bureaucracies
between disciplined performance and security of employment. Entrance into an
ofce . . . is considered an acceptance of a specic obligation of faithful manage-
ment in return for a secure existence (p. 199). As a factual rule, tenure for life is
presupposed . . . (p. 202, emphasis in original), hence guaranteeing a secure hold
on a position once obtained. The ofcial [can count on] the old age security pro-
vided by a pension (p. 203). Indeed, the motive for security is so strong, according
to Weber, that there are by no means infrequent efforts to obtain a guild-like
closure of ofcialdom (p. 200). Finally, for present purposes, the bureaucracy
strives everywhere for a right to the ofce by the establishment of a regular dis-
ciplinary procedure and by removal of the completely arbitrary disposition of the
chief over the subordinate ofcial. The bureaucracy seeks to secure the ofcial
position, the orderly advancement, and the provision for old age (p. 242).
What do these structural elements and motives in regard to security tell us about
the emotional foundation of ofcialdom? Let us agree rst that to feel secure
is an important affect. It is not a null state emotionally, but rather has positive
tone autonomically, indicating parasympathetic nervous system dominance. It
feels good, hence allowing for condent action, unperturbed by conscious or
unconscious states of emotional neediness or decit. Kemper (1978) and Kemper
and Collins (1990) have identied security as the emotional outcome of social
interactions in which one feels adequate in respect to ones power vis-a-vis
others.
In numerous interactions daily, human actors strive to maintain adequate levels
of power in relation to others, or to improve their power positions so they will be
felt as adequate. The emotional sequel of successful maintenance or attainment of
adequate power is a sense of autonomy and a feeling of security. Others cannot
228 THEODORE D. KEMPER
arbitrarily or easily compel one to act in a manner contrary to ones will. By
structuring the economic facet of occupational life so as to guarantee this emotional
underlay, modern organizations have taken a giant step forward in attaining a
disciplined conformity as quid pro quo from their administrative personnel.
The security motive is so powerful that it will brook even substantive irrational-
ity, as when the bureaucratic apparatus continues to operate even when its head
is lopped off, under the direction of the enemy (p. 229). Weber attributes this
to objective indispensability [based on its] impersonal character . . . (p. 229).
But this is to fail to read the stable emotional palimpsest that underlies the changed
situation. Indeed, it is precisely the guarantee of the indispensability of the whole
corps of ofcialdom, and the consequent occupational, ergo emotional, security
this affords, that makes possible bureaucratic obedience to whoever will continue
the guarantee, even an enemy. The experience of numerous conquered nations in
World War II conrms this general result.
Status Honor and Social Esteem
As powerful as the security motive is understood to be in this gloss of Weber, it
may not be sufcient to guarantee a fully disciplined conformity frombureaucratic
ofcialdom. Hence, the modern organization has evolved another type of emotional
gratication by which to secure compliance to the will of the master, or, what is
largely the same thing, the rules that are set down as guides to action. Bureaucracy,
says Weber, both in business ofces and in public service, is a carrier of a specic
status development . . . (p. 242).
The modern ofcial is the recipient of a special status honor and social esteem
vis-a-vis clients, or, as Weber puts it, the governed (p. 199). In his essay on
class, status, and party, Weber (1958) denes a status group in terms of a specic,
positive or negative, social estimation of honor (p. 187, emphasis in original). It
is realized in a specic style of life (p. 187, emphasis in original) that carries with
it restrictions on social intercourse (that is, intercourse which is not subservient
to economic or other of businesss functional purposes) (p. 187).
Inpublic bureaucracies, the ofcials social positionis guaranteed . . . byspecial
denitions of the criminal code against insults of ofcials and contempt . . .
(p. 199). Indeed, in Durkheimian terms, the ofcial stands in lieu of the collective
presence and interests of the whole society, participating thereby in its sacredness
(Durkheim, 1915). Weber himself refers to the (not entirely outdated) view of the
policeman as the representative of God on earth (p. 213). More practically,
modern ofcials accrue honor through the social prestige of educational certi-
cates (p. 241), which are prerequisites for entry into the modern organization.
The Differential Impact of Emotions on Rational Schemes 229
The status elements are also enhanced by the tendency toward status-group clo-
sure so that ofces are treated as prebends (p. 203), and by the ofcials readiness
to subordinate himself to the chief without any will of his own . . . (p. 208). (The
last is a bit puzzling, but seems to imply that there is a solidarity that emerges
among those who are peers vis-a-vis a more powerful master.) Finally, even in
putative democratic states, there is the leveling of the governed [vis-a-vis] the
ruling and bureaucratically articulated group . . . (p. 226, emphasis in original). In
sum, a congeries of elements of the social organization of bureaucracy in modern
society coalesce to foster status-group formation among bureaucratic ofcialdom.
What emotional gains do social esteem and status honor bring? Elsewhere
(Kemper, 1978), I have proposed that esteem and status ordinarily indicate
voluntary conferral of benets and rewards by others. The emotional result of
such interaction outcomes in which status is conferred is satisfaction, happiness,
pleasure. One simply feels good when others defer to oneself, acknowledging
thereby ones merit, conrming a good opinion of oneself. In addition, we
may reckon in the pride that stems from the sense of social superiority that the
deference of others engenders.
Even if the governed are not providing deference willingly, as would be
the case in a true status system, where most persons are genuinely admiring and
respectful of the position and accomplishments of others (see Kemper, 1978), the
experience of the bureaucratic stratum should nonetheless be highly satisfying.
Ordinarily, the governed except when the system is on the verge of tumbling
are careful to mask their resentments, so that deference appears to be authentic,
rather than coerced. Indeed, Weber (1958, p. 188) speaks of status honor as
usurped, but this moderates very little the satisfactions of felt superiority. Status
group formation, by whatever means, provides another emotional anchor for
ofcial discipline in the performance of duty.
Finally, the leveling of the governed necessarily entails a sense of superiority,
tinged perhaps with disdain or contempt for the governed. These sentiments ease
the task of enforcing the rational-legal forms in given instances, regardless of how
unpleasant they may be to the clients of the organization. Hence bureaucrats need
suffer few qualms, sparing themselves guilt and/or shame and certainly scanting
compassion in troubling cases.
Loyalty
In addition to security and satisfaction, a third positive emotion enables bureau-
cratic ofcials to play their assigned parts with impersonal imperturbability. This
is the emotional complex that emerges either in the case of loyalty to charismatic
230 THEODORE D. KEMPER
gures or to symbolic representations of societal ideals. According to Weber
(p. 199), in the pure type [of bureaucracy] loyalty is to an ofce, not a person
modern loyalty is devoted to impersonal and functional purposes. The intention
is to free the ofcial from specic devotion and susceptibility to inuence from
a particular occupant of a higher ofce, as in feudal arrangements of vassalage,
hence possibly violating the rules according to which decisions are to be made.
But loyalty is not a cognitive notion. It is, in all instances, an emotional commit-
ment based on high status regard for another, shading also into genuine authority
(see discussion in Kemper, 1978, pp. 377384) that overrides other interests that
may emerge. In terms of Webers own categories, it is a combination of affec-
tive and Wertrational orientation, both of which are antagonistic to the kind of
purely practical read unemotional means-ends action that ts into the category
of Zweckrational. Instead of eliminating emotional considerations in what binds
lower members to those above them in the hierarchy, modern bureaucratic orga-
nization reorients the emotion so that it is directed toward the organization itself,
rather than toward any merely dispensable individual, regardless of position.
Yet this is not always the case. Weber acknowledges the possibility of
Caesarism, in which loyalty is paid to the personal genius (p. 202). Here is
simply charisma in bureaucratic clothes, the dynamic and energetic leader who
summons personal devotion from subordinates. Although this is an unpredictable
phenomenon, its periodic emergence does not act to destroy the bureaucratic
form, but to reafrm it either through revitalization or change. In either case,
the consequence for administration is the subordination of personal interests in
the making of decisions, by afxing emotional energy (Collins, 1981) to the
organization or to an individual who is an organizational icon.
Appointment to Ofce
The three positive emotions of bureaucratic organization discussed thus far
security, satisfaction and loyalty are augmented by at least one negative emotion.
This is fear. It is the counterpart of security on the positive side, hence it constitutes
a negative inducement to conform to discipline. The fundamental source of fear in
bureaucracy is the fact that in the ideal-typical case ofcials are appointed, hence
they depend on the chief for recommendations, promotions and other benets.
Dependency places one in the power of another (Blau, 1964; Emerson, 1962) and
power differences generate fear (see Kemper, 1978; Kemper & Collins, 1990).
Weber speaks of the arbitrary disposition of the chief (p. 242), and lower ranking
ofcials are justly fearful of it, seeking to curb the chiefs power by the institution
of formal disciplinary procedures that allow them some rights to their position.
The Differential Impact of Emotions on Rational Schemes 231
Private bureaucracies tend to be mainly of the appointive type, giving immediate
superiors virtually absolute power over the careers of subordinates. In public
bureaucracies, unionization and civil service procedures such as examinations di-
minishsignicantlythe control of the superior. Insome instances, Weber points out,
ofcials are elected, a systemthat also reduces control of these superiors over their
subordinates (pp. 201203). Electoral systems do not ipso facto reduce the fear
component of bureaucratic careers, since such systems only transfer the source of
control over selection and retention to political parties and their bosses (Weber,
p. 203). But election to ofce would not ordinarily guarantee the affectless
performance of ofcial duties, such as the appointive system does. This is because
the chief has a position of his own within the bureaucracy and hence an interest in
conformity to its rules, while political party functionaries are ordinarily interested
only in how their man inside the bureaucracy can serve them (Weber, p. 203).
According to Weber, a mitigating factor on the development of fear among
bureaucratic ofcialdom is the fact that in the usual case, the chief or master is
at something of a personal disadvantage vis-a-vis lower ranking ofcials. This
is because ofce is based on expertise in matters regarding which the chief is
only a dilettante (p. 232). Hence it would seem that the chief must defer in
some respects to this expertise of his or her subordinates, hence dependence is not
entirely one-sided. Notwithstanding this advantage of ofcials over their chief, the
obverse of the dependency of chiefs on subordinates is that the superiors have a
motive to exercise power on their subordinates so that the latter will produce what
the superior is dependent on them for. Hence, lower echelon expertise is less of
a defense against fear than would appear from Webers analysis of the value of
expertise.
Additional Emotional Considerations
Weber provides us with several additional, incidental insights into the emotional
conditioning of bureaucratic ofcialdom. First, the emergence of bureaucracy in
the modern form also introduced the practice of at least some members in some
types of organizations of wearing distinctive uniforms, e.g. in the military and in
those organizations providing public services such as protection and transportation.
Presumably, the intention here is to emphasize recognizability both for public as
well as for organizational uses, hence enhancingthe efcacyof the bureaucracy. But
uniforms, by their nature, bring all who wear them under a common visual rubric.
This is a major source of solidarity, an emotional bond of status obligation toward
those who are identiable as comrades and peers. Since the uniform does not bind
one to a particular other, but rather to what might be regarded as a generalized
232 THEODORE D. KEMPER
other, it serves as a means to create loyalty and solidarity with the organization
and its purposes, as described in the section on loyalty, above.
The particular features of uniforms that emotionalize those who wear them
include such things as distinctiveness from normal attire in cut and color, thus
shifting the occasions of wearing them, even if these are daily, into the realm
of the sacred in Durkheims (1915) sense. There are also occasions, such as
parades, reviews, award ceremonies, and assignment of tasks, when ofcials in
uniform are massed together. To experience oneself in the serried ranks of the
similarly-garbed is to garner a sense of personal well-being and strength from
the group, and emotional energy and condence from its numbers due to ones
close conjunction with the sacred symbols and objects of the collectivity (Collins,
1981). This can only act to reinforce the ofcials sense of loyalty to the ideals of
bureaucratic performance of ones duty in the affectless way prescribed.
Asecond incidental contribution by Weber to our understanding of the emotional
foundation of affectless ofcial practice in bureaucracy is found in his assertion of
a virtual biological basis for conformity, tantamount to a residue, as described by
Pareto (1935). Weber proposed that the naive idea of Bakuninism of destroying
the basis of acquired rights and domination by destroying public documents
overlooks the settled orientation of man for keeping to the habitual rules and
regulations (p. 229, emphasis in original). This is a quite remarkable idea for
Weber to espouse as a sociologist, but, that apart, it suggests another emotional
basis for bureaucratic formalism, namely an innate requirement for following rules
once these have been set down and practiced. We must assume that failure to follow
the wonted pattern is unsettling in some way, productive of anxiety. Hence the
more comfortable course is to stay with the familiar. Whether this is an innate
attribute of humans, or can be explained better by socialization to norms for which
there are sanctions for violation, is less important than that the ultimate explanation
for conformity as presented by Weber resides in an emotion, not a cognition.
Weber reveals a last emotional nexus in organizational life by virtue of what I
believe is an erroneous example on his part. He writes: Every reorganization of
beaten or dissolved troops . . . is realized by appealing to the trained orientation
of obedient compliance to such orders (p. 229). This would make it appear that
obedience under difcult or contrary circumstances is attained simply by focusing
attention on the requirements of duty. There is a naive S-R quality to this that is a
little astonishing in a sociologist with Webers orientation to a Verstehen method.
Indeed, an appeal to beaten or otherwise disorganized cadres must of necessity be
couched in terms that will arouse strong countervailing emotions to fear, despair,
shame, the emotions that are likely to be prevalent in situations of defeat and
disorganization. More likely then, as successful inducements to reorganization
in the face of peril, are appeals to courage so as to overcome fear, to pride and
The Differential Impact of Emotions on Rational Schemes 233
self-respect to overcome shame, and to love of ideals, referring to home, family,
country, and the like, so as to overcome despair. All else failing, a nal emotional
appeal may succeed. It depends on creating an even greater fear of remaining
disorganized than of confronting the dangers that are currently feared. In military
settings, soldiers are threatened with court-martial or with being summarily
shot.
Only at one point does Weber appear to touch upon the emotional links of duti-
ful performance by bureaucratic ofcials. Referring to the superiority of positive
incentives over direct coercion by the master as a means to assure steadiness
of performance, Weber writes:
Strict discipline and control, which at the same time has consideration for the ofcials sense of
honor, and the development of prestige sentiments of the status group, as well as the possibility
of public criticism, work in the direction of strict mechanization (p. 208, emphasis added).
Here Weber addresses the pride in the occupational prestige of ofcialdom, further
the pride of membership in what might be regarded as an elite corps (of ofcials)
relative to other groups in society, and, nally, the sense of vulnerability and fear of
exposure that might follow from less than precise adherence to bureaucratic rules.
This insight that brushes on emotions is the rare exception in Webers otherwise
strictly unemotional examination of bureaucracy.
Over all, Weber has given us a striking image of modernity the bureaucrat who
performs organizational tasks and makes decisions in accordance with previously
stipulated rules while suppressing his or her own personal bias or preference.
There are reasons to quarrel even with this depiction, since only automatons could
conformfully to Webers ideal-typical portrait. But the purpose here is not to negate
Webers model of affectless ofcialdom, but to argue that the ideal-typical model
of the ofcial in virtually all respects depends on emotional nexi that bind the
bureaucrat to the organization. It is an ironic cage that the modern organization
is capable of dehumanizing its staff only by the expedient of emotionalizing it in
virtually every aspect of its connection with the organization, except at the point
where the staff executes its primary duty, i.e. in the making of decisions that affect
others, hence producing emotions in the lives of clients, while, in the process,
sternly controlling them in its own.
I turn now to a proposal for social organization that also is shot through with
emotions, and also is not recognized as such. Unlike the case of Webers bureau-
cracy, analyzed above, in which emotion is critical to any possible success of that
organizational form, James Colemans proposal for rational social melioration ap-
pears doomed by virtue of the emotions that likely inhere in its implementation.
234 THEODORE D. KEMPER
ON THE SOCIAL INEFFICIENCY OF MONEY:
NEGATIVE EMOTIONAL EFFECTS
OF OVER-RATIONALIZATION
James Coleman has offered sociology a proposition it almost cannot refuse. In a
visionary statement aimed generously and single-mindedly at social melioration,
Coleman (1993) proposed that sociologists must bring what they know to the
crumbling institutions and organizations of our society in order to reconstruct
them on a more rational basis. The goal is to salvage a society where social
indicators of decay and deterioration mark large sectors and population segments,
the costs of which, in crime, welfare budgets, hampered efciency at work, not
to speak of broken lives, are intolerably burdensome. In Colemans view, it was
not always thus. He sees societal decay as a result of the decline of what he calls
primordial institutions.
In primordial institutions, the hallmark of which is the family, informal social
controls accomplished societal objectives without much purposive planning.
Small size, clear authority structure, and relatively undivided labor contributed to
the ability to pursue those objectives with little ambiguity and relative effective-
ness. But the forces of modernity have converted small into large, clear authority
into multiple and attenuated foci of control, and widespread common labor into
enormous differentiation of roles and tasks. A major result of this transformation
is that the original institutions are no longer able to accomplish the tasks of
socialization and control that gave society a predictable, if often unjustly based,
stability. Today, a kind of chaos threatens, as crime, ignorance, and incompetence
emerge as a result, in part, of the impotence of the primordial institutions.
Colemans solution, and one that conforms suitably to the pattern established
by modernity, is to construct (or reconstruct) institutions and organizations on a
rational basis to enable them to accomplish what the old ones cant any longer.
And it is to this task that sociology is called.
So far, so good. The mainly academic employment and career structure of soci-
ology has inhibited a sleeves-rolled-up approach to the solution of social problems.
Scholarly reection and research, rather than purposeful reconstruction, mark
sociology in the main, despite slowly increasing numbers who nd employment
in so-called, applied settings. But Coleman has something much grander in
mind, something in the manner in which economists apply themselves to practical
problems. Economics is also a social science, one that overlaps sociology in some
parts, as both institutional economics, and economic sociology reveal. The
major difference between them is that economists are found at the highest levels
of government, community, and business enterprise, where they conduct analyses,
The Differential Impact of Emotions on Rational Schemes 235
forecast outcomes, and formulate policy to achieve or avoid those outcomes.
Coleman envisions a comparable role for sociologists. And this deserves three
cheers for its unabashed pride and boldness in behalf of a discipline that has much
to offer, but has had few takers over the years.
But Coleman offers more than a cogent rationale for sociologists; he sees
their energetic participation in the reconstruction of failing organizations and
institutions. He presents also a theory to underwrite the work, especially as it
relates to motivating individuals to perform the desirable tasks the organizations
and institutions are reconstructed to accomplish. Here the economic model is
important again, since the theory, rational choice, (RCT) is one that is almost
entirely contained in the economic canon.
Principally, it asserts that people act to maximize their interests, preferences
or utilities. Stated thus, it is unexceptionable, since it permits any goal, taste,
inclination, desire to be an interest, preference or utility. Thus dying for the welfare
of others (i.e. altruism) is as permissible here as seeking to amass the largest fortune
in history (i.e. greed). Furthermore, nothing in the theory prohibits the recognition
of subsidiary interests, preferences and utilities, so that even if greed is colossal,
it does not mean that any means, such as murdering every human obstacle to the
goal, will be employed, in the same way that it does not mean that the altruist will
sacrice his or her life for anyone in peril, as opposed to those who are genetically
or socially close.
Almost needless to say, the RCT approach has stimulated a signicant body
of criticism, much of it from scholars who object that RCT is best conceived as
an effort to make rationality normative, but that it fails badly as description of
what people do (Elster, 1989; Tversky & Kahneman, 1990). Yet others object to
the apparent selshness implicit in models of conduct that prima facie exclude
concern for others (Etzioni, 1988; Held, 1990). Still others reect that the entire
theory is a tautology, since it does not have independent sources of verication of
its premises other than the very behavior it seeks to predict (Kemper, 1993; Plott,
1990). These are valid concerns, but they omit important issues that becloud the
question of whether rational choice, as currently conceived, is possible at all.
This is not to say that it is not possible to choose wisely and well, but rather that
the current formulation of how such choice operates is limited, hence awed as a
predictive device.
My major contention here is that what is considered rational is already
thoroughly imbued with emotion. This view is empirically supported by the
work of Damasio (1994) who found that accidental or surgical excision of the
emotional centers of the brain prevented the cognitive centers from successful
performance. Turner (2000) takes the position further by melding this insight
with an evolutionary model of human social development. And, on the frontier of
236 THEODORE D. KEMPER
current explorations in articial intelligence, emotion is seen more and more as the
necessary complement to the cognitive elements of thinking machines (Trappi
et al., 2002). Collins (1993) argues further that rationally-oriented conduct is at its
heart an effort to maximize or optimize emotional energy, a fundamental factor
in creating social solidarity and for receiving its ensuing benets. Thus, so-called
rational conduct is universally at the service of an emotional motivational
scheme, or, as Hume (1963, p. 227) fervently asserted:
Reason is and ought only to be the slave of the passions, and can never pretend to any other
ofce than to serve and obey them.
With emotions thus in place ab initio, when emotion interferes with rational con-
duct, it does so by substituting a different emotion from the one that is already op-
erating. An example at some length will serve here to capture the essence of what is
a much larger discussion. The most elementary representation of rational choice is
the problem(or game) known as the Prisoners Dilemma. In its paradigmatic form,
it involves two suspects who are being interrogated separately about their recent
criminal activities together. Each has the option to confess, implicating the other
as well, or not to confess. If one suspect confesses and other doesnt, the confessor
receives a light sentence (e.g. one year) while the non-confessor receives a heavy
sentence (e.g. 10 years). If both confess, they each receive an intermediate sentence
(e.g. ve years). But if neither confesses, they each receive a yet lighter sentence
(e.g. two years). Under the circumstances, what would a rational person do?
Clearly, the remarkable thing about this game, which models many real
situations, is that, from the perspective of each player, the most rational strategy
is to confess, thus minimizing ones own sentence. Since each player ordinarily
reasons that this is what the other player will do, in order to avoid the worst
possible outcome for oneself (the 10-year sentence), each player will choose to
be selsh (let the other fellow serve 10 years!), The irony here is that this rational
choice by each suspect leads to an outcome (a ve-year sentence) that is worse
than that which an irrational (i.e. cooperative or unselsh) choice would have
provided, namely a two-year sentence.
Theorists in the rational choice tradition have made a good deal of progress
in dealing with this dilemma of rational conduct. For example, what has come
to be known as the tit-for-tat strategy of play seems to be effective in inducing
cooperation in games where players will interact more than once (Axelrod, 1984).
That is, choose the cooperative alternative on the rst play, and then do exactly as
the other player did in the previous play on all subsequent moves. Another restraint
on the rationally-oriented selsh strategy is when players will be interacting
with each other subsequently outside the game context (Mansbridge, 1990). In
both conditions, the purely rational (and self-defeating) strategy is moderated.
The Differential Impact of Emotions on Rational Schemes 237
What is missing from this analysis is the understanding that the so-called rational
strategy is itself highly emotion-laden, as is any alternative strategy. Without
understanding the emotional foundations of so-called rational choice, theorists in
this tradition miss the vital understanding of how to moderate behavior effectively
in the cooperative direction.
Indeed, what would prompt a prisoner to defect from a fellow prisoner but one
or more of the following emotions: fear, anxiety, loathing, and horror over the
prospect of a long prison term; anticipatory envy and resentment of the fellow
sinner who, by all rational calculation, is going to act selshly and suffer much
reduced punishment; perhaps even hatred of himor her on this account. And should
we omit guilt, shame, or remorse? Indeed, in the circumstance, rational choice is
over-freighted with emotion, and, frankly, it is hard to see it as anything but. If we
now escape from the district attorneys ofce to the general case of the prisoners
dilemma game, where some advantage such as money is the cardinal good to be
gained, we can add obsession with possessions, competitiveness, and greed, which,
though not standard emotions, have little relevance to social conduct unless they
are viscerally, or emotionally, charged (Kemper, 1993).
Where Colemans application of rational choice theory turns dubious is in
its virtually complete subjugation (at least implicitly) to a classical economic
application of the theory, namely, that the most effective rational incentive for
shaping human conduct is money. Weber (1968, p. 86) took the position that from
a purely technical point of view, money is the most perfect means of economic
calculation. Calculation in terms of money . . . is thus the specic means of rational
economic provision. It is precisely this ultra-rationality that Coleman is seeking.
In his most developed example, Coleman offers to reconstruct failing fami-
lies and schools as follows: Since these two old-line institutions arent producing
well-socialized, educated, and, at least, adequately performing individuals see
delinquency and crime rates, single-parenthood rates, ignorance in the workforce,
and so on reconstruct them through offering monetary incentives, or bounties,
for an improved product. The incentives would go to parents, or parent surrogates
(should parents chose to delegate their traditional role and authority), if they were
able to produce human products, i.e. children, who would cost society less than
some statistically estimated norm for their particular social identities, and gain
society some productive advantage that is above a statistically estimated norm
for their group. In effect, undressed, Colemans proposal is strictly rational in the
economic sense.
Given such an incentive scheme as the apparent foundation stone of Colemans
vision for a reconstructed society, one may wonder what part sociologists would
have to play. A rational choice is offered between money and no money. Although
Coleman urges sociologists to join in his reconstruction project, it may be asked,
238 THEODORE D. KEMPER
wheres the sociology here? Regrettably, Colemans inspiring call to action has
ended in the narrowest cul-de-sac of the classic homo economicus. Sociologys role
here can only be to oppose Colemans misleading notion, which can be called the
fallacy of over-rationalization. Although economic models require such simplied,
single-minded, super-rational actors, who must survive in competitive markets,
social life is broader than the market, and the effort to impose a market rationale
on questions that transcend the market must fail. Here that failure will be addressed
from the perspective of emotions, the motivational matrix out of which all social
behavior, including the rational, emerges.
THE NEGATIVE EMOTIONS IN
COLEMANS SCHEME
Were Colemans proposal for a hyper-rationalized socialization and education
institution productive only of positive emotions joy, pride, satisfaction, liking,
love, and such who would complain? Indeed, if Colemans social logic were
correct, these would indeed be the resulting emotions: children would enter
productive and rewarding lives, free from the faults and follies that land them in
long-term welfare-dependence and penury, or jail, or worse; parents would reap
the monetary benets of having done their task especially well, in addition to the
normal pride of parents of successful offspring. And society at large citizens
and community members would be pleased at their reduced tax bill, improved
quality of life (i.e. fear down, satisfaction up), and relief from both guilt and
anxiety over the previously deteriorating state of society.
But, in the tradition of economic theories of all stripes, there are costs to
be counted along with the benets Coleman envisions, and they are not simply
the opportunity costs of the money-bounty that must be paid. Coleman does not
consider that money incentives to achieve some social goals may over-rationalize
the process and thus prove to be socially inefcient. Providing monetary incen-
tives to ensure that parents or their surrogates (hereafter called agents) train
offspring to a certain standard, or exceed it, can have the following emotional and
motivational defects:
(1) Money incentives can reduce agents intrinsic motivation to engage in the
more difcult, sometimes heroic, aspects of the training task, which may
be accomplished only through strong emotional commitment to goals or
values other than money. One wants a healthy dose of Webers Wertrational
to mobilize agents, who must perform tasks that the primordial institutions,
evolved over eons of time and social experience, had accomplished. A
The Differential Impact of Emotions on Rational Schemes 239
relatively sturdy tradition of research in psychology (see Deci et al., 1999)
nds that when extrinsic incentives (such as money) are offered for task
performance, intrinsic motivation is attenuated. Robert Frank (1988),
himself an economist, puts is well:
Satisfaction from doing the right thing must not be premised on the fact that material
gains may later follow, rather it must be intrinsic to the act itself. Otherwise, a person
will lack the necessary motivation to make self-sacricing choices . . . (p. 253).
Childrearing is an arduous, often thankless, task that requires the capacity
to sacrice ones own interests. What we know about money, as a special in-
centive, tells us that while it may induce people to take even criminal risks, it
does not sufce to get most people to do morally difcult things. Frank (1988,
p. 253) concludes his statement as follows: Once others sense that [the moti-
vation to make self-sacricing choices is lacking], material gains will not, in
fact, follow. Furthermore, Kohn (1993) found that when organizations use
rewards to control behavior, there is likely to be an increase in surveillance,
evaluation and competition. These also decrease intrinsic motivation (Deci
& Ryan, 1985). Thus, Colemans money bounty is at risk to achieve its
intended goal precisely through its inability to evoke the kind of emotional
and motivational commitments that the successful agent must have.
2
(2) There is a likelihood that the child who is enrolled in a bounty system will
come to resent the agent for subjecting him/her to arduous efforts not alone
for the childs own good but for the agents gain, especially when there is a
cultural overlay about the meaning of parenthood as springing mainly from
love, not the cash nexus. In adolescence or young adulthood, we court the
possibility of active rebellion against the goals of the bounty system itself.
In an inevitably hyperbolic overreaction, conformity to socially useful goals
may be rejected. For bounty children, that may be the only way to regain an
authentic sense of self as other than an economic commodity in the agents
accounting scheme.
(3) Conversely, the agent will come to resent the child if the child does not
perform up to standard and thus threaten the agent with loss of the monetary
bounty. All kinds of emotional dangers can ensue, including signicant
abuse of the child for being refractory, dilatory, lazy, inattentive, and so on.
Harsh discipline will be a likely resort in many cases, inspiring fear, hatred,
deep anger and resentment not only against the agent, but the agents aims
and goals. Again, highly counter-productive from Colemans perspective.
(4) Status dishonor, inferior repute, and contempt will adhere to those children
who are known to be the products of bounty-formula training. It is inevitable
that childrearing modes be ranked in an order of preference and desirability.
240 THEODORE D. KEMPER
It is also inevitable that the money-bounty method will rank lower than the
strictly voluntary mode. (Some evidence of this is available in the attitudes
expressed toward mothers who are induced to attend a maternal-and-child-
health clinic because they receive $10 each time they keep an appointment
(Kolata, 1994). This kind of judgment will not long be hidden from the
children themselves, as they discover a newformof ism in the playground:
bounty-ism, with its attendant sequel of shame.
(5) With the agent as model, the child can develop a cynical attitude toward
efforts that are not tied to monetary gain, nor feel motivated to be productive
for any reason but money, despite the aws in such a motivational scheme.
Not only is the agent subject to the possibly disabling effects of monetary
incentives (point one, above), but the process can disable the child as
well, narrowing the focus of major commitments solely to those that pay,
literally, with time-and-a-half for overtime, and double-time on Sundays
and holidays. Or, like lawyers, every working minute becomes billable.
Otherwise, theres nothing in it for me.
(6) To ensure success that is, that the bounty be paid the agent may engage
in coercive or procrustean efforts to shape the child according to the stan-
dardized external criteria for success, disregarding individual differences of
temperament, disposition, ability, and the like. Among the things we know
about how work-interests and commitments are shaped, is that parental
ambitions for their childrens career are often frustrated. If a child wishes
to spend time drawing, or playing a musical instrument, or just pondering
the complexity of a leaf or an insect, instead of keeping up with the learning
regimen of the moment, this does not automatically bespeak failure, or lack
of interest in the more formulaic aspects of the learning routine. But there is
a greater likelihood that any divergence from the bounty-inspired learning
plan will be seriously discouraged. A straitjacket approach to learning,
with attendant listlessness, indifference, and a sandpapering down of
individuality is the likely result. Webers (1958, p. 182) terrible prescription
for bureaucratic personalities may thus come strangely to be realized. As he
put it, they will be specialists without spirit . . . Indeed, they may then also
become, as a recourse to emotions denied them in the work they do without
passion, (continuing Webers bleak anathema) sensualists without heart.
In one thing, however, Weber will be wrong: This nullity, as he calls them,
will imagine it has attained a level of civilization never before achieved.
In another metaphor Weber called the outcome mechanized petrifaction.
But it will surely not be embellished with convulsive self-importance.
(7) Non-bounty families (mainly middle class) will resent the bounty families
(mainly lower class) for imposing upon them the externality of having
The Differential Impact of Emotions on Rational Schemes 241
to pay for the program. In other words, there is a strong moral sentiment
that childbearing entails responsibility indeed, the argument has been
made that that is what families are by-and-large created for, namely to
shoulder the responsibilities attendant on childbearing. For some members
of the population to resist the responsibility, absent a cash payment which
those who do take the responsibility seriously must supply, is inevitably
to evoke powerful ritual (following Collins, 1981) abhorrence as well as
resentment. Suppose even that the bounty system were ultimately less costly
than the deterioration it was designed to remedy. This would not mollify
or assuage those who must sacrice monetarily (!) because others will not
sacrice emotionally for their children.
(8) The monetary dependence of the parent on the childs performance gives
the child a degree of control over the parent and an ability to resist adequate
socialization in other domains, thus creating a whole new set of emotional
relations in the family. Mutatis mutandis, Parsons (1954) analysis of
the incest taboo ts nicely here. One sociological reason for the taboo
is that it undermines parental authority. Each parent becomes vulnerable
to sexual blackmail by the child, hence the parents cannot exercise their
necessary parental role. In the bounty situation, Do your homework! may
need to be phrased, Please, wont you do your homework? Although
this may sound like an improvement over many a coercive formula, it
fails, if the response is, Ill do my homework, if I dont have to clean
my room, (or go to the store, or go to grandma with you, or come home
before 10 or 11, etc. oclock, etc.). As in Richard Emersons (1962) classic
formulae (P
ab
= D
ba
and P
ba
= D
ab
),
3
the parents dependence on the
child gives the child power over the parent, and not always for the childs
good.
Finally, I offer a few additional defects of the bounty system which
have negative emotional aspects, but will not elaborate upon them in this
presentation.
(9) Agents efforts must be tied closely to the conventional wisdom of experts,
which may lead to rigidity-inexibility or over-uidity of methods as expert
opinion necessarily changes (e.g. the oscillating preference for breast-vs.
bottle-feeding). In a rapidly changing set of knowledge conditions, agents
may well develop both indifference and contempt for the experts whose
guidance they are supposed to respect.
(10) Fertility may increase as children come to be viewed as a cash crop. Increased
family size may overload parents, leading to emotional short-fuses in
family relations and alienated family members, including the very ones who
are supposed to be learning via the bounty system.
242 THEODORE D. KEMPER
(11) Inevitable disagreements over methods and incentives between agents and
state administrators will focus energy on issues of program control rather
than on issues of effectiveness, leading to varieties of rancor and disdain
between agents and administrators.
(12) Since the bounty-payoff is in a distant future, reliance on monetary incentives
may be ineffective fromthe outset, although this wouldnt be known for some
time. Once results become available in individual cases, if good outcomes
are absent, the potential for blame and guilt and/or shame is abundant.
(13) Agents may be induced to engage in fraudulent methods so as to ensure their
nancial benet regardless of true accomplishment. Brandon (1999) points
out that while the state may allot funds to parents, it cannot easily monitor
how the money is spent.
(14) State interest in both setting standards and in paying agents for achieving
them grossly enlarges the scope of state authority.
Is this too bleak an assessment of Colemans rational choice proposal? Doubt-
less, some bounty families will succeed in overcoming all decits, such as those
indicated above. But a rational evaluation of the families Colemans proposal
aims at suggests that they are among the least likely to exhibit the cognitive
and emotional suppleness and stamina to deal with the negative emotional
consequences of their participation in the bounty scheme.
CONCLUSION
I have assumed throughout, and on evidence, that the rational and emotional may
not be hermetically sealed from each other. Nor should they be. Social processes
are inherently laced with emotion and social theories that ignore this risk their
validity. While it is not always necessary to elaborate the emotional foundations
or concomitants, a complete statement requires this. Here I believe I have shown
that emotion as a part of social process is value neutral, so to speak. In the case of
Webers formulation of bureaucracy, emotion clearly assisted in the success of that
enterprise, while in the case of Colemans bounty scheme, emotion would more
than likely cause it to destruct.
NOTES
1. Weber (1958, p. 216) refers to Kadi justice as a mode of action which knows
no reasoned judgment. That is, the Kadi (a type of Islamic ofcial) was not guided by
a set of rational-legal principles devised to apply to all situation, but could improvise
The Differential Impact of Emotions on Rational Schemes 243
a decision on the basis of whatever emotion or impulse arbitrarily moved him at the
moment.
2. It is only fair to say that Decis intrinsic motivation theory is opposed mainly by
behaviorists. See for example Dickinson (1989) and Carton (1996).
3. Emersons equations are read: The Power of A over B equals the Dependence of B
on A, and the Power of B over A equals the Dependence of A on B.
REFERENCES
Axelrod, R. (1984). The evolution of cooperation. New York: Basic Books.
Becker, G. (1981). A treatise on the family. Chicago: University of Chicago Press.
Blau, P. (1964). Exchange and power in social life. New York: Wiley.
Brams, S. (1985). Superpower games. New Haven: Yale University Press.
Brandon, P. D. (1999). The state, the child and imperfect parenting. Rationality and Society, 11,
399418.
Carton, J. S. (1996). The differential effects of tangible rewards and praise on motivation: Acomparison
of cognitive evaluative theory and operant theory. Behavior Analyst, 19, 237255.
Coleman, J. (1986). Individual interests and collective action. New York: Cambridge University Press.
Coleman, J. (1993). The rational reconstruction of society. American Sociological Review, 58, 115.
Collins, R. (1981). On the microfoundations of macrosociology. American Journal of Sociology, 86,
9841014.
Damasio, A. R. (1994). Decartes error: Emotion, reason and the human brain. New York: Putnam.
Deci, E. L., Koestner, R., & Ryan, R. (1999). A meta-analytic review of experiments examining the
effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 125, 627668.
Deci, E. L., & Ryan, R. (1985). Intrinsic motivationa and self-determination in human behavior.
New York: Plenum.
Dickinson, A. M. (1989). The detrimental effects of extrinsic reinforcement on intrinsic motivation.
Behavior Analyst, 12, 115.
Durkheim, E. (1915). The elementary forms of the religious life. London: Allen & Unwin.
Elster, J. (1989). Solomonic judgments. New York: Cambridge University Press.
Elster, J. (2000). Rationality, economy and society. In: S. Turner (Ed.), The Cambridge Companion to
Weber (pp. 2141). Cambridge: Cambridge University Press.
Emerson, R. (1962). Power-dependence relations. American Sociological Review, 40, 252257.
Etzioni, A. (1988). The moral dimension. New York: Free Press.
Fineman, S. (1993). Emotions in organizations. Thousand Oaks, CA: Sage.
Frank, R. (1988). Passions within reason: The strategic goal of the emotions. New York: Norton.
Hechter, M., & Kanazawa, S. (1997). Sociological rational choice theory. Annual Review of Sociology,
23, 191214.
Held, V. (1990). Mothering vs. contract. In: J. Mansbridge (Ed.), Beyond Self-Interest (pp. 287304).
Chicago: University of Chicago Press.
Hume, D. A. (1963). Treatise of human nature. In: The Philosophy of David Hume (pp. 11312).
New York: Modern Library.
Kemper, T. D. (1978). A social interactional theory of emotions. New York: Wiley.
Kemper, T. D. (1993). Reason in emotions or emotions in reason. Rationality and Society, 5, 275282.
Kemper, T. D., & Collins, R. (1990). Dimensions of microinteraction. American Journal of Sociology,
96, 3268.
244 THEODORE D. KEMPER
Kohn, A. (1993). Punished by rewards. Boston: Houghton-Mifin.
Kolata, G. (1994). Clinic entices patients with $10 a visit. NY Times, May 4, C13.
Mansbridge, J. (1990). On the relation of altruism and self-interest. In: J. Mansbridge (Ed.), Beyond
Self-Interest (pp. 133143). Chicago: University of Chicago Press.
Overbye, C. (1995). Making a case for the rational, self-regarding ethical voter . . . and solving the
paradox of not voting in the process. European Journal of Political Research, 27, 369396.
Pareto, V. (1935). The mind and society. New York: Harcourt Brace.
Parson, T. (1954). The incest tabu in relation to social structure and the socialization of the child. British
Journal of Sociology, 5, 101117.
Plott, C. R. (1990). Rational choice in experimental markets. In: K. S. Cook & M. Levi (Eds), The
Limits of Rationality (pp. 146174). Chicago: University of Chicago Press.
Tittle, C. R. (1980). Sanctions and social deviance. New York: Praeger.
Trappi, R., Petta, P., & Payr, S. (2002). Emotions in humans and artifacts. Cambridge, MA: MIT Press.
Turner, J. (2000). On the origins of human emotions. Palo Alto: Stanford University Press.
Tversky, A., & Kahneman, D. (1990). Rational choice and the framing of decisions. In: K. S. Cook &
M. Levi (Eds), The Limits of Rationality (pp. 6089). Chicago: University of Chicago Press.
Weber, M. (1930). The protestant ethic and the spirit of capitalism. New York: Scribner.
Weber, M. (1958). From max weber: Essays in sociology. In: H. Gerth & C. Wright Mills (Trans. and
Eds). New York: Oxford University Press.
Weber, M. (1968). Economy and society. Berkeley: University of California Press.
CONSCIOUSNESS, EMOTIONS,
AND SCIENCE
Jack Barbalet
ABSTRACT
The centrality of emotions to all signicant social, indeed human activities is
now broadly acknowledged. Nevertheless, discussion of emotions in core ac-
tivities of science, as distinct fromthe motivation of scientists, is undeveloped.
In reviewing the role of emotions in science the paper shows that emotions
provide consciousness of objects of scientic relevance. It is also shown that
emotions necessary to scientic activities are typically experienced noncon-
sciously. These two issues, of emotional consciousness and nonconscious
experience of emotion, raise a number of questions for the study of both
consciousness and emotions.
INTRODUCTION
Consciousness is rather like time, as Augustine described it: You know what it is
until you try to explain it. The ease of intuitive grasp of consciousness, coupled
with the difculty of intelligible exposition, offer clues to its nature. Because it is a
necessary aspect of experience, we correctly feel that we knowwhat consciousness
is. At the same time, we struggle to articulate what is meant by consciousness
because it is not external to our mental processes, but is a part of them. Con-
sciousness, then, is not readily conceived as an object, in that sense, which would
facilitate intellectual apprehension of it. Also, the mental processes in question
are inherently complex and, possibly as a result of this complexity, paradoxically,
Theory and Research on Human Emotions
Advances in Group Processes, Volume 21, 245272
2004 Published by Elsevier Ltd.
ISSN: 0882-6145/doi:10.1016/S0882-6145(04)21010-4
245
246 JACK BARBALET
are not adequately captured by self-awareness. With regard to the complexity
of consciousness, William James, for instance, remarked that a purely cognitive
description of it would leave consciousness dry and without purpose, a situation
corrected by appreciating the importance of emotion to consciousness (James,
1890a, pp. 141142). Even more counter-intuitive, it can be fairly asked whether
persons are necessarilyaware of their consciousness, are theynecessarilyconscious
of being conscious? These are the focal questions the consciousness of things
emotions can provide, and the possibility of lack of awareness of such conscious-
ness when considering the role of emotions in science, which this paper shall
address.
Before discussing emotions in science it is appropriate to change key and
subject, and briey consider the science of consciousness. For most of the 20th
century suggestion of a science of consciousness would have provoked derision.
Psychology studied behaviour not mind, and rejected introspection as a method
of inquiry when it was the most accessible means to consciousness. Neurology
was simply uninterested in such ethereal and nebulous things as consciousness,
which had no meaning for science. Since the 1990s, however, consciousness
has been rehabilitated and is not only now respectable for psychologist and
neurologist to study, but highly topical for them to do so. This sea change is a
result of developments in technologies, especially those that delineate and provide
functional imaging of the electrical correlates of neurological activity. Such
technologies reveal a correspondence between neural processes and aspects of
conscious experience.
The social sciences, with the possible exception of anthropology, have been left
out of the current concentration of scientic interest in consciousness. It is ironic
that this too is reversal of an earlier trend. In a 26-page review of developments
in the scientic study of consciousness, for instance, half a page only discusses
social theories, and of the 6 items cited in that short space, none is written by
a social scientist (Zeman, 2001, pp. 12811282). No doubt the majority of social
scientists today share the image presented here, of a lack of interest in conscious-
ness amongst them. And yet Marxs treatment of class-consciousness, for instance,
dened his account of capitalist society; Durkheims discussion of collective con-
sciousness similarly captured his approach to social solidarity. The importance of
consciousness to the sociology of Weber, Mannheim, Simmel, Cooley and Mead,
to name only the most obvious, cannot go unremarked. While the causes of the
tendency for sociologys disengagement from consciousness require careful con-
sideration, reference to its recent cultural turn indicates a shift from a concern
with social actors apprehension of a world to which they relate, to a concern
with a shaping spirit to which subjects without agency yield. In such realignment,
consciousness is lost.
Consciousness, Emotions, and Science 247
A sociological interest in emotions raises questions concerning consciousness.
While the issues treated in such an inquiry will not be the same as those promoted
by recent neurological and psychological research, drawing on the ndings of
such research will enrich sociological discussion. The broad concern of the
following discussion is the role of emotions in science. This may at rst seem
remote from a concern with consciousness. Nevertheless, as we shall see, this
enquiry unavoidably raises the question of emotional consciousness and aspects
of the nature of consciousness itself. From the point of view of prevailing
conventions emotions have only a disrupting or distorting role in science. Science
practitioners, science educators and philosophers of science have long insisted
that science can only proceed when emotion is expelled. But as emotions are
pervasive in giving direction to and energizing human activity, it is a fair working
hypothesis that some emotions will have a supporting role in science. Indeed,
the signicance of emotions in motivation to science is frequently acknowledged.
Nevertheless, that emotions may have a positive role in core scientic activities is
an idea that continues to attract scepticism. One reason for this, it will be argued,
is that the emotions involved operate below the threshold of awareness, which
is to say that scientists are typically not conscious of the emotions central to
their activities. This idea, that emotions may be nonconsciously experienced, also
frequently meets resistance and will also be discussed below.
PART 1: SCIENCE AND EMOTIONS
Science is arguably the dening social institution of western societies, at least since
the 17th century. The importance of science is not simply in its direct output of
reliable knowledge, or even only in its contribution to technological development
and the consequent expansion of economic production. The signicance of these
aspects of science is enormous, of course. But in addition to the direct yield of
science is another element of its signicance that makes it a pervasively domi-
nating institution, namely that its form has tended to be replicated by all other
social institutions including those associated with cultural production. There are
broad imitative tendencies in areas outside of science to adopt its form. Under the
aegis of science practically all areas of human endeavour tend to favour analysis
rather than synthesis, for example, measurement rather than rule-of-thumb, and
validation rather than enchantment. There is additionally a prevalent suspicion
of emotion, exemplied in science and sometimes regarded as a leading if not
dening feature of it. Any evidence of a recent weakening of the residual cultural
suspicion of emotion may be explained in terms of the declining status of science
in contemporary western cultures, as much as any other factor.
248 JACK BARBALET
While not able to pursue all of the above themes in the present paper, the fol-
lowing discussion will focus on the relationship between science and emotion.
The supposed antipathy between science and emotion is typically associated with
Cartesian origins (Toulmin, 1990, pp. 113115). The 17th century mathematician
and philosopher, Ren e Descartes, famously held that mind and body are radically
different substances as it is possible, he argued, todoubt the existence of all physical
objects, including ones own body, but impossible to doubt the existence of oneself
as a thinking being. Passions, being of the body, are by hypothesis fundamentally
distinct from mind. It follows, then, that science as a mental activity can only be
disrupted by inuence of the emotions. This extrapolation, that emotions disrupt
science, while frequentlyattributedtoDescartes, is not a complete representationof
his position, as we shall see below. It is necessary to be clear, in any event, that con-
cernhere is the impossibilityof emotioninthe intrinsic activities of science. Yet this
was not the original basis on which science and emotion were held to be opposed.
Passion in the Performance of Science but
not the Communication of Science
When the Royal Society was founded in 1662, according to Thomas Sprats History
of the Royal Society (1667), eloquence in language was abolished because it is
contrary to reason and abets passion (quoted in Jones, 1953, p. 85). The underlying
assumption seems to be that passion or emotions undermine science. Yet this is a
curious conclusion to draw because the scientists who constituted the movement
of this period, now known as the Scientic Revolution, frequently referred to the
emotions they believed were necessary to their scientic activities.
William Harvey, for instance, writing in 1628, refers to emotional turmoil
caused by the puzzlement he experienced in distinguishing systole and diastole.
He explains how his mind was therefore greatly unsettled, and how this
emotional discomfort was resolved through experiment and observation, when
I had attained to the truth, that I should extricate myself and escape from this
labyrinth (Harvey, 1628, p. 273). Isaac Newton, writing in 1672, indicates
different emotions but imputes similar importance to them when referring to his
response to light shone through a prism:
It was at rst a pleasing divertissement to view the vivid and intense colours produced thereby;
but after a while applying myself to consider them more circumspectly, I became surprised to
see them in an oblong form; which, according to the received laws of refraction, I expected
should have been circular . . . Comparing the length of this coloured spectrum with its breadth,
I found it about ve times greater, a disproportion so extravagant that it excited me to a more
than ordinary curiosity of examining from whence it might proceed (quoted in Daston & Park,
2001, p. 303).
Consciousness, Emotions, and Science 249
A century after Newton, Joseph Priestly reports his discovery of oxygen, when
burning a candle in air extracted from mercuric oxide, in terms of the same
emotions:
But what surprised me more than I can well express, was, that a candle burned in this air with
a remarkably vigorous ame . . . but as I had got nothing like this remarkable appearance from
any kind of air besides this particular modication of nitrous air [nitric oxide], and I knew no
nitrous acid was used in the preparation of mercurius calcinatus [mercuric oxide], I was utterly
at a loss how to account for it (Priestley, 1776, p. 142).
After more detailed description of activity and reaction, following this passage,
Priestley goes on to say:
I wish my reader be not quite tired with the frequent repetition of the word surprise, and others
of similar import; but I must go on in that style a little longer (Priestley, 1776, p. 146).
The appreciation of the importance of emotion to science during this period
has been noted by Lorraine Daston and Katherine Park when they wrote that
[m]using admiration, startled wonder, then bustling curiosity these were the
successive moments of seventeenth-century clich es describing how the passions
impelled and guided natural philosophical investigations (Daston & Park, 2001,
p. 303). Even Descartes gave a place in science to wonder, which he calls the
rst of all the passions (Descartes, 1649, p. 358). This is because the only
object of wonder is the knowledge of the thing that we wonder at (Descartes,
1649, p. 363) and it therefore disposes us for the acquisition of the sciences
(Descartes, 1649, p. 365). Too much wonder, though, Descartes warns, can lead
to pursuit of triviality and therefore in spite of its positive attributes he continues
to regard it cautiously (Descartes, 1649, p. 365).
What concerned the Royal Society was not passion in science but the dangers
of passion in persuasion, a fear of distortion in scientic communication through
emotion in eloquence. The source of this concern was the 17th century revival of
classical rhetoric. Rhetoric is the study and discipline of persuasion. Aristotles
Rhetoric, which established the eld, included a consideration of means of per-
suasion through the structure of argument and forms of speech. It also included a
lengthy discussion of the emotions, the nature and character of each, its origin,
and the manner in which it is produced (Aristotle, 330BC, p. 19). Aristotle wrote
extensively on emotions in his Rhetoric because he believed that the success of
persuasive efforts depend in part on the emotional dispositions of the audience.
During the 16th century in England, however, works on rhetoric largely focussed
on linguistic form and neglected the emotions. This was corrected during the 17th
century, througha returntothe traditionof Aristotle (Sloan, 1971, pp. xxxiixxxvi).
This is best exemplied in Thomas Wright, The Passions of the Minde in Generall,
a highly inuential work of the time which went through 5 editions from 1601,
250 JACK BARBALET
when it was rst published, to 1630, when the last edition appeared. If persuasive
effort depends on the emotional dispositions of the audience, then the emotional
manipulation of the audience became a further tool of persuasion. It was this latter
prospect that led the Royal Society to caution concerning the role of emotion in
artful persuasion.
Since the 17th century, concern regarding the place of emotions in science has
tended to generalize from distortion in science communication to distortion in the
practice of science itself. We shall see, though, that within this latter concern are
a number of qualications giving rise to quite different specic appraisals of the
relationship between science and emotions. Indeed, the 17th century situation was
more complex than described above, for against the Royal Societys concern with
the possibility of distortion of science communications through emotions, and Har-
veys and Newtons acceptance of emotional guidance in scientic investigation,
there is Francis Bacons position that:
The human understanding is no dry light, but receives an infusion from the will and affections;
whence proceed sciences which may be called sciences as one would . . . Numberless . . . are
the ways, and sometimes imperceptible, in which the affections colour and infect the under-
standing (Bacon, 1620, p. 267).
Bacons statement resonates with the view that became dominant from the late
18th century, namely that for science to precede emotion must be expelled.
The convention from the 18th century, captured by the quotation from Bacon
above, and frequently summarized as Cartesian, holds that as science is to
provide knowledge of the world external to the scientic observer, then science
must adequately represent that world, and thus it must be independent of
the scientic observer. According to this position, then, scientic knowledge is
necessarily impersonal. Representation of the impersonality of science is typically
achieved by characterizing science not in terms of its human attributes, including
the scientists emotions, but in terms of its formal properties, especially those
associated with its methods. The exclusion of emotion from science by virtue of
its impersonality is in fact compatible with emotions having a role in two aspects
of scientic activity, namely the motivation of the scientists investigation and
commitment of the scientist to the social institution of science.
Emotion and Scientic Motivation
The motivation of scientists, like that of human agents in general, can be character-
ized emotionally. The emotions that Newton, Harvey, and Priestley refer to, quoted
above, are the affective base of their motivation for scientic investigation. The
Consciousness, Emotions, and Science 251
joy of discovery, for instance, is frequently mentioned as a continuing emotional
incentive to engage in scientic work: the joy of discovery is a very real incentive
to research, despite the rareness of its realization (Baker, 1942, p. 17; see also
Feuer, 1963, pp. 1, 7, 13). Barker, who was a British zoologist, goes on to showthat
the emotionality of the scientists motivation can coexist with the impersonality of
science:
It is an error to suppose that the scientist is unemotional, or could succeed if he were. The error
has arisen through a misconception. The absolute necessity that a scientists ndings shall not
be changed from objective truth in response to emotional urges of any kind does not result in
his becoming a particularly unemotional person: whether a discoverer or anyone else is pleased
with a discovery has no effect on its validity (Baker, 1942, pp. 1718).
Writing at the same time as Baker, the American sociologist Robert Merton shifted
the focus away from individual motives to institutional control.
The individual or personal motives of scientists are necessarily diverse,
as Merton says, and may include, among a host of possibilities, a passion
for knowledge, idle curiosity, altruistic concern with the benet to humanity
(Merton, 1942, p. 613) and so on. Against these individual or distinctive motives
Merton posits the distinctive pattern of institutional control of a wide range of
motives which characterize the behaviour of scientists (Merton, 1942, p. 613).
As he indicated in an earlier article, institutional control does not operate to
the exclusion of emotions but refers to a sociologically signicant emotional
attachment to institutional norms: The institution of science itself involves
emotional adherence to certain values (Merton, 1938, p. 601). Merton makes his
point in terms similar to Bakers on individual motivation:
although it is customary to think of the scientist as a dispassionate, impersonal individual
and this may not be inaccurate as far as his technical activity is concerned it must be
remembered that the scientist . . . has a large emotional investment in his way of life, dened by
the institutional norms that govern his activity (Merton, 1938, p. 596).
Thus emotions can operate at the level of both the scientists individual motivation
and commitment to institutional norms, without undermining the essential
impersonality of what Merton calls the technical activity of science.
In these accounts, then, emotions are conned to the motivational framework of
science and not its interior. The sustaining institutional controls of science summa-
rizedbyMertonas the ethos of science andrepresentedas four sets of institutional
imperatives, namely universalism, communism, disinterestedness, and organized
scepticism, are infused with emotion. Indeed, Merton refers to the ethos of science
variously as an emotionally toned complex of rules, prescriptions, mores, beliefs,
values and presuppositions (Merton, 1938, p. 595 Note 16), and, with greater
succinctness and increased neutrality, as that affectively toned complex of values
252 JACK BARBALET
and norms (Merton, 1942, p. 605). In each case, though, the emotional element
is pronounced. And yet, the ethos of science is one limited aspect of science
as an institution namely its cultural structure, and must be distinguished from
both the set of characteristic methods by means of which knowledge is certi-
ed . . . [and the] stock of accumulated knowledge (Merton, 1942, p. 605). Thus
in this representation the interior of science remains free of emotions.
Mertons account of the norms of science has been frequently challenged
(Barnes & Dolby, 1970; Kuhn, 1977, pp. 330339). Even more serious, Mertons
norms have been transgressed without undermining the scientic enterprise but
with the effect of enhancing it. In a study based on interviews with 42 eminent
scientists who worked on the Apollo moon rocks, for instance, it was shown that
rather than open-mindedness among scientists it was frequently dogged com-
mitment and even bias that proved to be the strongest sustaining forces both
for discovery of scientic ideas and for their subsequent testing (Mitroff, 1974,
p. 73). Such emotional attachment to ideas and assessments idiosyncratically held
and possibly contrary to existing evidence contravenes in various ways the entirety
of the ethos of science as Merton conceives it. Nevertheless, such commitments
proved necessary for the advancement of science in this case, and were elements
in the practice of scientic rationality (Mitroff, 1974, p. 249). While this brief
account may undermine the Mertonian ethos, it leaves intact Mertons view that
emotions are in the framework of science but not its interior. Commitment and
bias are given to the ndings of investigation and are not interior to investigation
itself.
In a quite different assessment Lorraine Daston (1995) has shown that because
of their implicit metaphysical grounding Mertons norms are unfortunately
impervious to historical variation and cultural modication. In place of this rigid
ethos of science Daston posits a moral economy of science which is a web
of affect saturated values that stand and function in well-dened relationship to
one another . . . a balanced system of emotional forces, with equilibrium points
and constraints (Daston, 1995, p. 4). Through a discussion of the signicance of
moral economies in the generation and operations of quantication, empiricism,
and objectivity, Daston demonstrates the role of the emotions and values of histor-
ically situated scientists in the formation of scientic rationality: By examining
in a new light just those ways of knowing once thought to exempt science from
the realm of emotions and values, a study of moral economies may illuminate the
nature of the rationality that seemed to exclude them (Daston, 1995, p. 24). While
rightly criticising Mertons insensitivity to developments of custom and practice,
Daston fails to go beyond his relegation of emotion to the framework of science
and leaves unexplored the emotional components of the practical activities of
science.
Consciousness, Emotions, and Science 253
Emotions Within Science
Michael Polanyi (1974) presents a very different approach to the role of emotions
in science than Mertons. He says:
The outbreak of such emotions [as elation] in the course of discovery is well known, but they are
not thought to affect the outcome of discovery. Science is regarded as objectively established
in spite of its passionate origins . . . I dissent from that belief . . . and want to deal explicitly with
passions in science . . . [S]cientic passions are no mere psychological by-product, but have a
logical function which contributes an indispensable element to science (Polanyi, 1974, p. 134).
The function of scientic passion, according to Polanyi, is that of distinguishing
between demonstrable facts which are of scientic interest, and those which
are not . . . [and] also as a guide in the assessment of what is of higher and what
of lesser interest (Polanyi, 1974, p. 135). In addition to this selective function
Polanyi identies two further functions of scientic passion, the heuristic function
and the persuasive function. The latter is self-explanatory. The heuristic passion,
which Polanyi says is the mainspring of originality, the force which impels
us to abandon an accepted framework of interpretation and commit ourselves,
by the closing of a logical gap, to the use of a new framework, serves as a
guide to enquiry as it links our appreciation of scientic value to a vision of
reality (Polanyi, 1974, p. 159). Whereas Merton excludes emotions from the
method and substance of science, Polanyi says that they are indispensable within
it.
Polanyi trained as a scientist and was for many years Professor of physical
chemistry at the University of Manchester in England. His statement of the
importance of scientic passions therefore can be regarded as a conclusion of
participant observation. And yet his account deals only with the functional aspects
of scientic passions and lacks a description of the particular emotions involved.
Indeed, the scientic passions that he refers to have the form and character
of postulated entities rather than of empirical phenomena, more like quarks in
quantum chromodynamics than DNA in molecular biology. Another scientist,
whose writings have also attracted a philosophical and sociological readership,
Ludwik Fleck, also advocated the functional importance of emotions for scientic
work. Flecks discussion complements Polanyis and suggests why description of
emotions in science may be so difcult.
Fleck insists that emotionless thinking is meaningless (Fleck, 1935, p. 49)
when arguing against the German sociologist WilhelmJerusalem, writing in 1924,
that the appropriately trained individual acquires the ability to state facts purely
objectivelyandthus learns tothinktheoretically, that is, free fromemotion (quoted
in Fleck, 1935, p. 49). In countering the notion that there is thinking free from
254 JACK BARBALET
emotion Fleck argues that emotion is only either noticed or not noticed, and when
not noticed it may be thought to be absent:
There is only agreement or difference between feelings, and the uniform agreement in the
emotions of a society is, in its context, called freedom from emotions. This permits a type of
thinking that is formal and schematic, and that can be couched in words and sentences and
hence communicated without major deformation (Fleck, 1935, p. 49).
The society Fleck refers to is the thought collective constitutive of a particular
scientic community, and the agreement of feelings he postulates derives from
the shared thought style that characterizes the stock of knowledge and level of
culture of a scientic community (Fleck, 1935, p. 39).
According to Fleck, the emotional basis of science is pervasive (see Barbalet,
2002, pp. 140144). He shows that it is in the structure of the thought collective
(Fleck, 1935, pp. 105112), in the epistemological modelling of a discovery into
a fact (Fleck, 1935, pp. 86, 119, 144), and in the sense of certainty surrounding
factual knowledge (Fleck, 1935, pp. 117, 145) Given that pattern recognition
has such a central role in the observations that are at the heart of scientic research
(Ziman, 1978, pp. 4353), it is appropriate to focus on the affective dimension of
this aspect of Flecks account of scientic discovery. It is here that Polanyis rst
function of scientic passion is to be located, that of distinguishing . . . facts that
are of scientic interest.
Fleck indicates, rstly, that scientic observation can occur only after a period
of scientic training has been undertaken (Fleck, 1935, p. 48). He immediately
goes on to say that the role of training is to generate appropriate experience in the
trainee. Experience is so important because it is the basis of emotional sensibility,
as we shall see. Through possession of appropriate experience the trainee thereby
qualies for membership in the thought collective. This experience and its con-
sequences therefore provide access to observations consonant with the correlative
thought collective: Direct perception of form requires being experienced in the
relevant eld of thought . . . At the same time, of course, we lose the ability to see
something that contradicts the form (Fleck, 1935, p. 92). The epistemological
importance of experience referred to here is in the fact that scientic experience
or training produces particular expectations and therefore provides an appropriate
focus of attention.
Attention is the property of cognition that selects only a portion of the vast
range of sensory involvement for conscious awareness. Fleck gives the example
of the selection of different bacterial colonies from over a hundred different
cultures. The observations are not pure, he says, but anticipate differences
(Fleck, 1935, p. 90). Expectancy or anticipation is an affective element of a
number of particular emotions. It is the fundamental emotional basis of all
Consciousness, Emotions, and Science 255
vision and observation. Because they are not given social representation these
emotions are not culturally labelled, and are therefore without names. Thus,
scientists are seldom conscious of experiencing them, even though they underpin
the conscious awareness of pattern recognition that is central to scientic
observation.
Given the signicance of emotions in Flecks characterization of thought
collective it is important to appreciate the consistency of his account of thought
style. According to Fleck, truth and facts derive their meaning from particular
thought styles. A thought style, Fleck says, consists of a certain mood and
of the performance by which it is realized. He continues by saying that a
mood has two closely connected aspects: readiness both for selective feelings
and for correspondingly directed action (Fleck, 1935, p. 99). As mood
changes, so does meaning (Fleck, 1935, p. 110). The link indicated here
between emotions and meaning has been observed in a number of contexts,
including empirical observation and theory formation (see James, 1890b, pp.
312317; 1902, pp. 128129). Again, while conscious awareness of meaning is
taken for granted, the underlying emotional basis is typically not experienced
consciously.
To summarize discussion to this point: Fleck demonstrates the functions
of emotion in science that Polanyi similarly points to. The emotions involved
remain unnamed by the scientists who experience them, and it is highly
probable that scientists are not consciously aware of the emotions involved.
While the emotions in question are central to selection of the relevant facts
connected with scientic pattern recognition and observation in general, and
consciousness of those facts, scientists are not conscious of the underlying
emotions. Even the more externally apparent Mertonian emotions essential
for motivation and the institutional framework of science are frequently not
consciously experienced by the scientists subject to them, judging by the
continuing denial by numbers of scientists of emotional involvement in scientic
activity.
PART 2: EMOTIONS: CONSCIOUS
AND UNCONSCIOUS
The conclusion of the previous section, that key emotions in science are experi-
enced nonconsciously, is unacceptable to a body of literature. Other relevant lit-
eratures, though, in different ways, acknowledge the possibility of nonconscious
emotional experiences. The present section of the paper will review arguments
concerning this aspect of emotions.
256 JACK BARBALET
Emotions as Conscious States
The question, of whether emotions can be nonconsciously experienced, was asked
in a recent state-of-the-art summary of current ndings of emotions research
(Ekman & Davidson, 1994). A psychologist (Clore, 1994) and a neuroscientist
(LeDoux, 1994) each answered explicitly that it was not possible for emotions
to be nonconscious. A third emotions scientist, a psychologist (Zajonc, 1994),
was less explicit in so far as his answer to the question seemed to move from
nonconscious emotion to nonconscious inuence on emotion, things which Clore
and LeDoux appropriately distinguished. The arguments of Clore and LeDoux are
essentially identical.
Clore holds that while the cognitive processing that causes the emotion is
unconscious, the informational and motivational effects of emotion depend on con-
scious experience in order to capture the attention of the experiencer (Clore, 1994,
p. 285). He goes on to explain that the feeling component of emotion is necessary,
and that it is the fact that the emotion is felt that makes it conscious (Clore, 1994,
pp. 286287). These same elements of the argument feature in LeDouxs account:
while unconscious processes underlie emotions, emotional feeling the fact that
emotions are affectively charged means that emotions are conscious states
(LeDoux, 1994, p. 291). It is almost a matter of denition, then, that emotions are
conscious states, and that they cannot, therefore, be experienced nonconsciously.
Against these conclusions animportant contributiontothe sociologyof emotions
has demonstrated that certain emotions, at least, can be experienced without being
felt. By combining aspects of the work of sociologist Charles Horton Cooley
(1922) and psychologist Helen Block Lewis (1971), Thomas Scheff, for instance,
has been able to demonstrate that adult humans typically experience pride and
shame with such low-visibility that [they] do not notice it (Scheff, 1988, p. 399).
It is possible to conclude from Scheffs argument, therefore, that when the feeling
component of the emotion is by-passed a person may experience shame without
being conscious of the emotion.
Scheff argued that it is possible to indicate an incidence of shame, through
presentation of behavioural evidence, even though subjects may fail to report ex-
perience of shame. He did this by drawing upon Lewis analysis of a large number
of clinical encounters, which showed that while participants were not conscious
of the majority of shame episodes that they experienced, nonverbal or behavioural
shame-markers were nevertheless manifest (Scheff, 1988, p. 401). Scheff went
on to conrm Lewis conclusions through a re-analysis of Solomon Aschs cele-
brated conformity experiments. He shows that Aschs subjects who unreasonably
conformed to a group norm did so under the pressure of unacknowledged and un-
recognized shame (Scheff, 1988, pp. 402405). In this manner Scheff shows that
Consciousness, Emotions, and Science 257
a key mechanism of social organization, namely conformity, functionally depends
on experiences of shame that occur below the threshold of awareness, that is to
say, nonconsciously.
Juxtaposing the conclusions of Clore and LeDoux on the one hand and Scheff
on the other raises a number of issues, for both sets of conclusions cannot be
correct. One question is the signicance and role of feelings in emotions and
in consciousness. Not all approaches to emotion accept the primacy of feeling
in emotional experience that Clore and LeDoux assume. According to one
philosophical theory of emotions, cognitivism, individuals subjectively infer
their emotions not from the feelings associated with them but through the beliefs
inculcated by the emotions. Cognitivism need not be understood to claim that
feeling is not necessary to emotion, only that it is not sufcient. The issue here is
not the truth or otherwise of cognitivism, but the question of the status of feeling
in emotional experience. Scheff, for one, points to conforming behaviour that
the subject may subjectively explain in terms of beliefs, without being aware
of a feeling of shame. Another question: even if feeling is a sufcient element
of emotional experience, does it follow that the feeling must be experienced
consciously? Clore discredits an account that purports to show that fear can
be experienced without the feeling component being consciously experienced
(Clore, 1994, pp. 286287). This question does not arise for Scheff when he
argues that the feeling of an emotion may be bypassed. We shall see that it is
not unintelligible to hold that feeling might be experienced nonconsciously.
Unconscious Emotion
When consciousness becomes an object of enquiry, the possibility of an alternative
state, unconsciousness becomes a matter of interest. To this point of the discussion
nonconsciousness rather than unconsciousness has been mentioned because the
former term carries less cultural baggage than the latter, as we shall see. We shall
also see that under certain interpretations the terms are more or less equivalent,
but not under all interpretations. To appreciate just how confusing the matter of
unconsciousness is we might consider Clores remark, that in agreement with
Freud, I would argue that it is not possible to have an unconscious emotion be-
cause emotion involves an experience, and one cannot have an experience that is
not experienced (Clore, 1994, p. 285). This is to ask: is it possible to have an
experience that the experiencing agent is not aware of?
Much routine experience, of course, involves activities of which the subject is
not necessarily conscious. Indeed, in habitual action, as WilliamJames observes,
mere sensation is sufcient guide, and the upper regions of brain and mind are
258 JACK BARBALET
set comparatively free (James, 1890a, pp. 115116). Typical examples of this
include the experience of driving a car or locking the front door when leaving the
house. These types of experience have signicant unconscious elements in so far
as there is a lack of consciousness awareness of key facets of driving and locking
the house on the part of those engaging in such experiences. This occurs even
though the experience of driving involves tactile and practical awareness of the
clutch and accelerator, of the road, and so on; just as experience of leaving the
house involves experience of the door, the lock and the key. It is simply that those
persons having these experiences need not be consciously aware of being aware of
the things they engage when they drive or lock the door. In philosophical jargon,
there is an absence of second-order awareness even though rst-order awareness
occurs. Could not emotional experiences conform to this type of possibility?
It is implicit in the position of Clore and LeDoux that emotional experiences
are different than nonemotional experiences insofar as emotions are themselves
a special type of awareness by virtue of the feeling component. This position is
rejected here. Emotions provide those experiencing them with awareness of both
internal processes of arousal and also external objects. Emotions link awareness
of internal state and external objects through intention, which is experienced
internally but whose object is most frequently external. Let us accept that emotion
is registered to the person experiencing it through its affective component or
dimension, through the feeling of the emotion. As we shall see below, it is possible
that the emoting subject may not be aware of being aware of their internal arousal
or an external body that is the object of emotion, any more than the driver who is
aware of the road is necessarily aware that they are aware of the road. The absence
of second-order awareness is the condition of unconsciousness in this sense.
For Freud, however, whomClore invokes in making his case, unconsciousness is
not a lackof consciousness but a mental entityinits ownright, parallel toconscious-
ness, and possessing its own separate aspirations and concerns. This distinction,
between unconsciousness as an absence of consciousness, and unconsciousness
as a distinct mental entity, is crucial in addressing the question of consciousness.
It is not always understood (Grifths, 1997, pp. 151155). According to Freud,
personally painful and otherwise emotionally unacceptable memories and thoughts
become integrated into a subjects unconscious through a process of repression. It
is a matter of doctrine for Freud that the associated emotions loose their connec-
tion with the repressed material and are not themselves repressed (Zangwill, 1987,
p. 277). The entity Freud postulates as the unconscious, and the attendant notion
that there are no unconscious emotions, simply offers nothing to an understanding
of whether emoting subjects are aware of the emotions they experience.
WilliamJames was aware of the problemof confusingunconscious as anabsence
and as a substance. Indeed, he developed a critique of pre-Freudian notions of
Consciousness, Emotions, and Science 259
unconscious substance (James, 1890a, pp. 164176). He did not dismiss the
importance of unconsciousness as an absence, however, for he observed the
signicance of experiences of which the subject is not aware. To describe these
latter he explicitly rejected the adjective unconscious, being for many of them
almost certainly a misnomer, replacing it with the vaguer termsubconscious or
subliminal (James, 1902, p. 170). What James calls subliminal consciousness
might be described, in light of preceding discussion, as rst-order conscious, in
which there is experience of awareness of objects by virtue of affective relations
with them, but without attendant second-order consciousness of subject-awareness
of rst-order consciousness.
Terminological proliferation is not usually a source of clarity, but James does
endeavour to avoid confusion when he distinguishes between consciousness of
the ordinary eld on the one hand, and a consciousness existing beyond the eld,
or subliminally on the other (James, 1902, p. 188). Subliminal consciousness,
the discovery of which James says is the most important step forward that
has occurred in psychology during his lifetime (James, 1902, p. 188), refers
to a set of memories, thoughts, and feelings which are extra-marginal and
outside of the primary consciousness altogether, but yet must be classed as
conscious facts of some sort, able to reveal their presence by unmistakable signs
(James, 1902, p. 188). The context of this discussion is religious conversion,
and James is here concerned with the consequences of strongly developed
subliminal consciousness of religious innovators (James, 1902, p. 189). The
general signicance of subliminal consciousness, however, is its universality
(James, 1902, p. 188). Indeed, experience of subliminal consciousness is
held to be a peculiarity in the constitution of human nature (James, 1902,
p. 188).
The role of this peculiarity, as opposed to its incidence, can be readily demon-
strated. To take an instance unconnected with James or religious conversion,
the behavioural zoologist Niko Tinbergen, for example, discusses the study of
posturing behaviour in the Herring Gull:
In a social species such as the Herring Gull, numerous movements of the individual are un-
derstood by its companions, who react to them in special ways. Some of these movements
and postures are not difcult even for the human observer to appreciate, though the detection
of most of them requires careful study. There are a multitude of very slight movements, most,
if not all, of them characteristic of a special state of the bird. The student of behaviour is to a
high degree dependent on his ability to see and interpret such movements. In the beginning, he
will notice them unconsciously. For instance, he will know very well on a particular occasion
that a certain gull is alarmed, without realizing exactly how he knows it. Upon more conscious
analysis of his own perception (an important element in behaviour study), he will notice that
the alarmed gull has a long neck (Tinbergen, 1961, p. 7).
260 JACK BARBALET
Here is an account of the progress of scientic knowledge and understanding
in terms of a movement from subliminal consciousness to second-order con-
sciousness. The rst phase includes perceptual awareness of gull behaviour in
the absence of awareness of that concrete perception. This is knowing that the
gull is alarmed without knowing how it is known. The second phase includes the
observers awareness of her perception, awareness of being aware of the extension
of the gulls neck. This account is continuous with the discussion of pattern
recognition in the treatment above of Flecks contribution to our appreciation of
the role of emotions in scientic observation.
Emotion and Core Consciousness
What James refers to and Tinbergen instances, modern neuroscience explains.
Antonio Damasio, for instance, distinguishes three stages of processing along
a continuum: a state of emotion, which can be triggered and executed non-
consciously; a state of feeling, which can be represented nonconsciously; and
a state of feeling made conscious, i.e. known to the organism having both emotion
and feeling (Damasio, 2000, p. 37; emphasis in original). According to Damasio,
feeling is a private, mental experience of an emotion, whereas emotion is
understood as the collection of responses, many of which are publicly observ-
able (Damasio, 2000, p. 42). The distinction between feeling and emotion drawn
by Damasio is curious insofar as these are not symmetrical elements of a frequently
associated couple. More typically emotion is understood to comprise a number of
elements or components of which feeling is one, the others being motor expression,
motivation and behavioural tendencies, cognitive stimulus processing, and neuro-
physiological processes (see Scherer, 1984). It is possible to say that these compo-
nents are indicators of emotion and that no one of them singly or in combination
comprises the emotion, whichremains a particular type of experience (Leventhal,
1984, pp. 271272). The matter of interest in Damasios account, however, is that
the feeling state need not be registered consciously for it to be experienced.
Damasio demonstrates his claim with a clinical case study. A patient in
Damasios neurological clinic had 20 years previously suffered extensive damage
to both temporal brain lobes that had profound consequences for his capacity to
learn and remember. The patient, David, was as a result of his injuries physically
incapable of recognizing or naming any of the persons with whomhe interacted on
a daily basis, and was incapable of ever remembering whether he had seen any of
them before. And yet David nevertheless seemed to display consistent preferences
and avoidances for certain persons. This apparent inconsistency between Davids
condition and his behaviour was examined through an experiment designed by
Consciousness, Emotions, and Science 261
Damasio and a colleague (Damasio, 2000, pp. 4344). The experiment consisted
of David being subject to 3 distinct types of interaction over a period of time.
The rst type of interaction, with an experimental good guy, was pleasant,
welcoming and rewarded; the second type, with a neutral guy, was emotionally
neutral and involved tasks that were neither pleasant nor unpleasant; the third type,
witha badguy, was extremelytedious andboring. Subsequent tothese encounters
David was given 2 further tasks involving photographs of the 3 persons who had
played the good, neutral and bad guys. It is important to appreciate that David was
able to neither recognize the persons in the photographs nor remember whether he
had ever encountered them before. Nevertheless, when asked to whom he would
go if he needed help David consistently chose the good guy and consistently failed
to choose the bad guy. Although he was unable to say that he knew anything about
the persons in the photographs and could not remember ever seeing them before,
when asked who was his friend he consistently chose the good guy (Damasio,
2000, p. 45).
While nothing in Davids consciousness could be responsible for his correct
characterization of the good guy and the bad guy, his nonconscious preferences
are to be explained in terms of the experimentally induced emotions in the different
types of interactions he hadbeensubjectedto(Damasio, 2000, pp. 4547). Damasio
notes that Davids brain could generate actions commensurate with the emotional
value of the original encounters (Damasio, 2000, p. 46). It is likely that Davids
emotions were accompanied by a feeling of those emotions, but in the absence of
an appropriately related set of images that would explain to him the cause of the
reaction the feelings remained isolated and disconnected from their antecedent
conditions and from Davids subsequent behaviour (Damasio, 2000, p. 46). David
seemed to have no consciousness of his emotional feelings. Yet Davids emotions
provided him with a subliminal consciousness, to revert to James terminology,
of the good and bad guys. Damasio accounts for the differences indicated here in
terms of the distinction between what he calls core consciousness and extended
consciousness: Davids core consciousness was intact, but his neurological damage
prevented extended consciousness.
Core consciousness, according to Damasio, is essentially a selfs consciousness
of immediacy, of the here and now; whereas extended consciousness provides
elaboration of each of the component parts of core consciousness to such a degree
that it is qualitatively distinct, drawing upon quite different neurological structures
and processes (Damasio, 2000, pp. 234276). The sense of self, while rudimentary
in core consciousness, is elaborate in terms of identity and personhood in extended
consciousness. Extended consciousness goes well beyond the here and nowof core
consciousness and provides a rich awareness of both a lived past and anticipated
future. Extendedconsciousness but not core consciousness comprises second-order
262 JACK BARBALET
conscious experiences. This is because extended consciousness treats memories
and other mental representations as objects, giving rise to a sense of self-knowing.
In Damasios words: extended consciousness is the precise consequence of two
enabling contributions: First, the ability to learn and thus retain records of myriad
experiences . . . Second, the ability to reactivate those records in such a way that,
as objects, they, too, can generate a sense of self knowing, and thus be known
(Damasio, 2000, p. 197).
It is nowpossible to state the relationship posited by Damasio, between emotion,
feeling, and the two forms of consciousness, core and extended. The relationship
between emotions on the one hand and core consciousness on the other is nec-
essary, they occur together or are absent together because they require the same
neural substrates (Damasio, 2000, p. 100). There is no such functional relationship
between emotional processing and extended consciousness, continues Damasio,
which is why impairments of extended consciousness are not accompanied by
a breakdown of emotion (Damasio, 2000, p. 101). Indeed, emotion is part of
the mechanism that drives core consciousness. In briey describing this process,
Damasio suggests what is the relationship between feeling and consciousness when
he says that:
emotion has a truly dual status in relation to consciousness: the actual responses whose conse-
quences, as an ensemble, eventually produce an emotion are part of the mechanism that drives
core consciousness; a frame of time later, however, the collections of responses which constitute
a particular emotion can also be treated as an object to be known. When the emotional object
is made conscious, it becomes a feeling of emotion (Damasio, 2000, p. 350).
Emotional feeling, then, gives representation to the emotion in thought, introducing
a possibility of mental regulation as well as enhanced self-knowledge (Damasio,
2000, p. 56). Emotional feeling, then, is an object of extended consciousness, in
which it has instrumental possibilities and through which the emoting subject is
aware of it and the emotion with which it is associated. Emotional feeling is a
necessary product or consequence of core consciousness, but not known to the
emoting subject through it.
Emotional Consciousness and Consciousness of Emotions
Damasios account described above provides a means of overcoming the limita-
tions of those that fail to distinguish between levels of consciousness. It is, however,
in many ways similar to an earlier statement in the history of writing about emotion
that it nonetheless fails to mention, namely Jean-Paul Sartres phenomenological
theory of emotion. Damasios distinction between emotion as apprehension of
the here and now in core consciousness, on the one hand, and the apprehension
Consciousness, Emotions, and Science 263
of emotional feeling as an object in extended consciousness on the other, has its
analogue in Sartres earlier statement of a distinction between emotion as a form
of consciousness on the one hand, and consciousness of emotion on the other.
Sartre begins by describing emotion, in contradistinction to the usual psycholog-
ical and psychoanalytic accounts, as an indispensable structure of consciousness
(Sartre, 1939, p. 15). The consciousness that pertains to emotion, however, is
not consciousness of the emotion, but rather the emotions consciousness of the
world. Sartre says that:
Fear is not originally consciousness of being afraid . . . Emotional consciousness is, at rst,
unreective . . . Emotional consciousness is, at rst, consciousness of the world . . . the person
who is afraid is afraid of something (Sartre, 1939, pp. 5051).
As emotion is necessarily conscious of its object, which is in the world, Sartre
insists that emotion is therefore never unconscious but only unreective (Sartre,
1939, pp. 5657). None of this departs from the position that has been developed
in preceding sections of this discussion, even though there is a continuing absence
of terminological consistency which reects the state of the literature and not
the present writer conceptual agreement is nevertheless achieved. Emotion is
consciousness but unreective in the sense that as an apprehension of the world,
or at least objects in it, emotion entails rst-order consciousness. Sartre insists that
second-order consciousness of emotion is not required for emotional experience.
This latter, for Sartre, is emotional apprehension of the world rather than the
emoting subjects awareness of those emotions.
Here, then, Sartres idea that emotions are never unconscious is merely termino-
logical. An absence of consciousness of emotion is not an instance of nonconscious
emotion because emotion itself is a form of consciousness as relating to an object
in the world. Incidentally, Sartre also explicitly rejects the notion of unconscious
substance, conducted in a slightly earlier work (Sartre, 19361937, pp. 5458). It
is ironic, given his critique of William James theory of emotion (Sartre, 1939, pp.
2240), that Sartres critique of unconscious substance is reminiscent of James
argument against unconscious substance, mentioned above.
In all of this Sartre says in a different way what has already been treated in
discussion above. The real novelty of Sartres account of emotions, however, that
for our purposes is especially relevant for an understanding of the role of emotions
in science, is his account of the transformative capacity of emotion. But what is
of value in this notion has to be surgically extracted from Sartres incomplete and
misleading treatment of the magical nature of emotional transformation.
Sartre says that emotional consciousness, a consciousness of the world through
emotion, is a certain way of apprehending the world in which the affected
subject and the affective object are bound in an indissoluble synthesis (Sartre,
264 JACK BARBALET
1939, p. 52). The apprehension of the world through emotional consciousness,
he goes on to say, is a transformation of the world (Sartre, 1939, p. 58). Sartre
explains it thus:
When the paths traced out become too difcult, or when there is no path, we can no longer live
in so urgent and difcult a world. All the paths are barred. However, we must act . . . Before
anything else [emotional consciousness] is the seizure of new connections and new exigen-
cies . . . Thus, through a change of intention, as in a change of behaviour, we apprehend a new
object, or an old object in a new way . . . The impossibility of nding a solution to the problem
objectively apprehended as a quality of the world serves as motivation for the new unreective
consciousness [emotion] which now perceives the world otherwise and with a new aspect, and
which requires a new behaviour through which this aspect is perceived and which serves as
[material] for the new intention (Sartre, 1939, pp. 5860).
The characterization of emotional consciousness, as transformation of the world
through seizure of new connections and the apprehension of a new object or an
old object in a new way, is entirely equivalent to Adam Smiths (1795) account
of scientic discovery in The History of Astronomy, to which we shall return
shortly. While Sartre is correct to insist that emotion is . . . accompanied by belief
[about the world] (Sartre, 1939, p. 73), his claim that such beliefs are necessarily
false cannot be sustained.
Sartre is committed to untruth in emotion. He regards emotions as essentially
magical not because he associates emotions with imagination but because he
assumes that actors engage emotional consciousness only when they are impotent
in facing a world hostile to their desires (Sartre, 1939, p. 67). Even joy is ratio-
nalized in this way when Sartre describes its expressions, dancing and singing,
as symbolically approximate behaviour, incantations designed to realize the
possession of the desired object as instantaneous totality (Sartre, 1939, p. 69).
The characterization of emotion as magical in transforming, indeed distorting,
reality in order to match frustrated desire, leads Sartre to describe it as degraded
consciousness (Sartre, 1939, pp. 7577, 83). Inherently ineffective, according to
this view, emotion is at best a form of coping behaviour, for Sartre. And here is the
awin Sartres otherwise suggestive and potentially useful account of emotion: not
all emotions are expressive, nor are they meaningfully describable as behaviour, as
Sartre argues (see Solomon, 1981, p. 222). Indeed, the emotions central to science
that manifest transformative powers of reconceptualization of natural relationships
are not behavioural and entirely nonconscious, or, in Sartres terms, unreective
and not the objects themselves of consciousness. Before discussing further the
transformative power of emotions, we shall briey consider their truth value.
In the passage quoted above Sartre refers to the impossibility of nding a
solution to the problem objectively apprehended as a quality of the world [which]
serves as motivation for the new unreective consciousness [emotion] which now
Consciousness, Emotions, and Science 265
perceives the world otherwise and with a newaspect. The objective world is some-
how unacceptable thus promoting a response that is an emotional apprehension
that presents the world differently. There is a double error in Sartres statement.
First, what might be called an empiricist dogma that objectively ascribed qualities
are timelessly true; and second, that emotional consciousness is necessarily false.
Sartres archetypical emotional transformation is performed by Aesops fox,
who in his disappointment and frustration in failing to reach ripe grapes turns
them sour (Sartre, 1939, p. 61). This is one possibility, certainly. But it is not the
only one. Compare Sigmund Freuds characteristically similarly acknowledgment
of the transformative power of emotion, when he refers to love, for example,
as leading to sexual-overestimation, to idealization that falsies judgment
(Freud, 1921, p. 190), with James treatment of love, in which special qualities
of another, otherwise missed without the penetrating cognitive benet of love, are
apprehended through it (James, 1899, pp. 266267). While there is no guarantee
of truth in emotional experience, there is no necessity of untruth either. In the
context of science, emotion is foundation to advancement of knowledge through
its transformative capacity.
In a remarkable but neglected treatment of emotions in science, contemporary
with historically early scientic achievement, Adam Smith (1795), in a posthu-
mously published essay, points to a transformative capacity of emotions to which
Sartre is blinded through magical incantation. Natural philosophers, says Smith,
through training and disposition, are liable to feel an interval betwixt two objects,
which, to more careless observers, seem very strictly conjoined (Smith, 1795,
p. 45). The natural philosophers experience of a things dissimilitude with all
the objects he had hitherto observed has the consequence of producing in him
uncertaintyandanxious curiosity (Smith, 1795, p. 40). These anxious andpainful
emotions that beset the scientist or natural philosopher are cured by scientic
engagement:
Philosophy, by representing the invisible chains which bind together all these disjointed objects,
endeavours to introduce order into this chaos of jarring and discordant appearances, to allay
this tumult of the imagination, and to restore it, when it surveys the great revolutions of the
universe, to that tone of tranquillity and composure, which is most agreeable in itself, and most
suitable to its nature (Smith, 1795, pp. 4546).
Science, then, relies on an emotional process of the anxious discomfort of percep-
tion of disjuncture leading to the construction of explanations representations
of invisible chains that ease the pain, as the repose and tranquillity of the
imagination is the ultimate end of [natural] philosophy, according to Smith (1795,
p. 61). In Smiths account, then, a problem objectively apprehended promotes an
emotional experience that leads, through the transformative capacity of emotion,
266 JACK BARBALET
to a new perception of objectively ascribed qualities of the world, which can
be described as more appropriately, accurately or truthfully representing that
world.
PART 3: SELF-TRANSCENDING EMOTIONS
AND SCIENCE
Self-Asserting and Self-Transcending Emotions
Rebuttal, above, of Sartres supposition that emotional transformation necessarily
leads to false constructions of reality was that such transformations do not
necessarily lead to false constructions. An understanding of the role of emotion
in science requires more than so ambiguous a conclusion. The transformative
capacity of emotions that Sartre focuses upon, and that Adam Smith also points to
in his history of the progressive attainment of improved astronomical knowledge,
is connected with emotional intentionality. There is broad agreement that emotions
are not mere apprehensions of reality but interested apprehensions that lead those
who experience them to particular dispositions.
A dening feature of emotions is that they perceive the world from the perspec-
tive of the emoting subjects needs or preferences. Few would disagree with the
statement that:
one of the major functions of emotion consists of the constant evaluation of external and internal
stimuli in terms of their relevance for the organism and the preparation of behavioural reactions
which may be required as a response to those stimuli (Scherer, 1984, p. 296).
Different evaluations are entailed in different emotions. If the environment is expe-
rienced as unduly inhibiting, for instance, the organism will evaluate it as hostile
and the accompanying emotion will be anger through which there is preparation
to act against what is perceived as the constraining factor. Sartres insight is to see
that a persons inability to cope with changes in their circumstances may transform
a notice fromthe bank, for instance, into the service managers hostility. But this is
not the onlypossibility. It is not the onlypossibilitynot just because different people
see the world differently but also because there are different types of emotions, and
the intentionality of some do not function to assert self-needs in this narrow sense.
When Scherer wrote the passage quoted above it is likely that he had in mind
those emotions that are associated with the possibility of behavioural reaction, that
typically have high expressivity and clear if not strong physiological correlates,
such as anger, fear, love, disgust, shyness and so on. There are a number of ways in
which emotions can be categorized; the most frequently cited being the distinction
Consciousness, Emotions, and Science 267
between basic or primary emotions and secondary emotions. The categorization
that is most useful in understanding the role of emotions in science, however,
is that which distinguishes self-asserting and self-transcending emotions. This
categorization can be traced back to the work of WilliamJames, and no doubt many
other writers, andis explicitlystatedbyArthur Koestler (1964). The emotions listed
above, and that best t the statement of function in the quotation from Scherer in
the previous paragraph, are self-asserting emotions. Self-transcending emotions
have a very different prole and serve quite different purposes.
Koestler noted that self-transcending emotions tend to be ignored by emotions
researchers, whotypicallyfocus insteadonself-assertingemotions (Koestler, 1964,
p. 285). The neglect of self-transcending emotions, which continues still, is a mea-
sure of their lack of salience but cannot be construed as an index of their lack of
importance or infrequent incidence. Examples of self-transcending emotions iden-
tied by Koestler include grief, longing, worship, raptness, and aesthetic pleasure
(Koestler, 1964, p. 285). While all emotions are in principle delinked from action,
self-asserting emotions entail characteristic dispositional programs in which the
possibility of action consequent upon them is part of their meaning. Experience of
fear, for instance, does not necessarily lead to ight or ght, but the possibility of
these associated behaviours is integral to what is meant by fear. A characteristic
feature of self-transcending emotions, on the other hand, is the relative absence
of behavioural concomitant. As Koestler says, grief, longing, worship, raptness,
aesthetic pleasure are emotions consummated not in overt but in internalized be-
haviour . . . [that] induce passive contemplation, silent enjoyment (Koestler, 1964,
p. 285). Here is the lack of salience or visibility in self-transcending emotions. But
here also is a property of a type of emotion that would perform service to and is
likely to be associated with science.
By its nature, science is a cognitively expansive set of practices, ever seeking
new ndings and new means for attaining new ndings. Indeed, one concern of
those fearful of the disruptive role of emotions is the way in which emotions may
inappropriately narrow scientic focus through the introduction of self-interested
goals. This was the issue mentioned above in Bakers expression of the concern
that a scientists ndings shall not be changed from objective truth in response
to emotional urges (Baker, 1942, pp. 1718). This possibility is a justiable
concern relating to self-asserting emotions because of their propensity to narrow
the consciousness of those who experience them. Self-transcending emotions, on
the other hand, partly because they share a participatory dimension, Koestler says,
involve an expansion of consciousness by identicatory processes of various
kinds (Koestler, 1964, p. 286). This point can be demonstrated by considering
one particular self-transcending emotion, aesthetic pleasure, and its role in
science.
268 JACK BARBALET
Aesthetic Emotion in Science
It is not infrequent that sound reasons exist for a number of different options when
a choice has to be made regarding which particular option would move scientic
research forward (Kuhn, 1977, p. 328). Under these circumstances, appeal to
the facts cannot offer closure, and Kuhns alternative, a half-hearted appeal to
values they inuence but do not determine choice (Kuhn, 1977, p. 331)
requires support and correction. Scientists themselves frequently appeal not to
values but to aesthetic pleasure in such circumstances, but seldom if ever report
these emotions in scientic publications. Nevertheless, implicit in the notion of
a scientic community is the possibility that aesthetic pleasure is not a private
idiosyncratic experience but an outcome of consistency of social milieu. Indeed,
the importance of training and experience to scientic cognition, discussed above,
further suggests a convergence of aesthetic judgement in scientic communities
and its possible application to scientic decision making.
Aesthetic experience has a number of elements, of which the emotional compo-
nent is the most obvious. A second element, that reinforces the self-transcending
nature of aesthetic pleasure, is the intellectual component of aesthetic experience.
This is because aesthetic experiences are disinterested in the sense that focus on an
objects appearance or structure is at the same time to put that objects instrumental
or practical attributes, that relate more directly to the needs of self-assertion, out of
focus. In this way the intrinsic properties of objects are emotionally apprehended
(see Maslowquoted in Kemper, 1979, p. 305). The emotional-intellectual pleasure
that derives from aesthetic experience can only be fully understood when the
phenomenal object of such an experience is properly specied.
The aesthetic object is never a thing in itself, but an arrangement of things
or a perspective that permits a view of arrangements. It is the perception of
organization in an apparent chaos of parts that is at the heart of aesthetic
experience and the source of joy that accompanies it. Aesthetic experience, then,
arises through perception of a particular type of organization of elements that
realizes certain values. To this extent Kuhn is correct. The values in question have
been identied by Maslow, for instance, as wholeness, uniqueness, and aliveness
(quoted in Kemper, 1979, p. 303). Within the context of a scientic community,
the appreciation of these values will be inherent in the thought collective, to use
Flecks (1935) term, referred to above. Indeed, the experiential basis of scientic
cognition, discussed above, will necessarily include a practical specication of
precisely those values central to such aesthetic experience.
Aesthetic experience, then, is a response to the realization of the values
characterized here as wholeness, uniqueness, and aliveness, in the circumstances
of scientic research and decision-making. In other words, aesthetic experience
Consciousness, Emotions, and Science 269
is a response to a correspondence between these values of the scientic thought
collective and the particular conditions encountered in research. The attainment of
correspondence between values and condition evokes the emotions of joy, delight,
and pleasure. Kuhns reference to values is therefore insufcient because the
efcacy of values in scientic choice is through the role of emotions in signalling
correspondence between values and conditions in aesthetic experiences.
Kuhn is nevertheless correct to suggest that values can be regarded as a guide
through ambiguous situations. That a scientist realizes her values is a necessarily
desirable event, in which satisfaction, indeed joy, is inherent. This is the point: the
experience of joy, which is the affective force of all aesthetic experience, is the
emotional expression of the realization of values, of attaining a correspondence
between the values which guide a scientists choice and the conditions that
scientist perceives. In signalling a perception of correspondence between values
and conditions, joy is precipitated as a feeling of self-actualization and of the
meaningfulness of ones activities, indeed being (de Rivera, 1977, p. 64). In this
regard, joy and wonder are parallel emotions (de Rivera, 1977, p. 66), a point of
some importance given the role of wonder in scientic discovery.
The aesthetic experience, then, in the context of scientic choice, is an emo-
tional apprehension of one possibility over others on the basis of a felt realization
of pertinent values in one option and their absence in alternative options. That
this is an emotional apprehension of choice reects the nature of scientic choice.
It must be emphasized that these latter arise when there is always at least some
good reasons for each possible choice (Kuhn, 1977, p. 328). The question of
choice arises not because there is an absence of evidence for one nding or
theory as against others, but because there is evidence for all, and therefore an
absence of determining evidence. The necessity of aesthetic emotion or joy is
therefore unavoidable in the development and progress of science. Because the
emotions involved are self-transcending they direct rst-order consciousness
of the objects they apprehend that then becomes the object of second-order
consciousness.
Through the emotions within scientic activity the scientist is aware of the
objects the emotions bring to consciousness, but the emotions do not draw
attention to themselves. As archetypical self-transcending emotions they are
experienced through the aesthetic objects they draw to the scientists attention
and are themselves experienced nonconsciously. An extremely astute observer,
the plant morphologist Agnes Arber, has captured something of this:
It is not possible to offer strict scientic evidence for the idea that not only reason but emotion
has a function in biological discovery . . . we can only point to slight indications which are at
least compatible with its truth. It is recognized, for instance, that the moment at which a fruitful
combination of ideas enters the awareness, is often charged with a particular feeling of joy,
270 JACK BARBALET
which precedes and seems independent of, the rational satisfaction of goal-attainment (Arber,
1954, pp. 2021).
The joy that Arber refers to is not the motivational joy exterior to science that comes
with the satisfaction of any job well done. It is an emotional basis of activity inte-
rior to science. Neither is this joy register of the conclusion of a piece of scientic
work, but the affective mechanism that allows it to continue toward its conclusion.
CONCLUSION
It has been shown that emotions are not only central to the sustaining framework
of science but also necessary in its core activities. Emotions are able to discrim-
inate between relevant and irrelevant objects of scientic concern, and underlie
the scientists consciousness of those objects. That such emotions are typically
experienced nonconsciously is a dimension of emotional experience deserving of
further investigation. Indeed, it is likely that the incidence of an absence of full
conscious awareness of emotional experience is not conned to science, and in the
broad span of human activity is statistically very high. Study of science, then, in
addition to its intrinsic interest is also in this regard a model for research of other
social institutions where nonconscious emotional experience is to be found.
REFERENCES
Arber, A. (1954). The mind and the eye: A study of the biologists standpoint. Cambridge: Cambridge
University Press.
Aristotle (330BC[1926]). Art of rhetoric. Loeb Classical Library Cambridge, MA: Harvard University
Press.
Bacon, F. (1620 [1905]). Novum organum. In: J. M. Robertson (Ed.), The Philosophical Works of
Francis Bacon (pp. 256387). London: Routledge.
Baker, J. R. (1942). The scientic life. London: Allen & Unwin.
Barbalet, J. (2002). Science and emotions. In: J. Barbalet (Ed.), Emotions and Sociology (pp. 132150).
Oxford: Blackwell.
Barnes, S. B., & Dolby, R. G. A. (1970). The scientic ethos: A deviant viewpoint. European Journal
of Sociology, 11, 325.
Clore, G. L. (1994). Why emotions are never unconscious. In: P. Ekman & R. J. Davidson (Eds),
The Nature of Emotion: Fundamental Questions (pp. 285290). New York: Oxford University
Press.
Cooley, C. H. (1922 [1964]). Human nature and the social order. New York: Schocken Books.
Damasio, A. (2000). The feeling of what happens: Body and emotion in the making of consciousness.
London: William Heinemann.
Daston, L. (1995). The moral economy of science. OSIRIS, 10, 324.
Consciousness, Emotions, and Science 271
Daston, L., & Park, K. (2001). Wonders and the order of nature, 11501750. New York: Zone Books.
Descartes, R. (1649 [1955]). The passions of the soul. In: E. S. Haldane & G. R. T. Ross (Eds),
Philosophical Works of Descartes (pp. 329427). New York: Dover Publications.
Ekman, P., & Davidson, R. J. (Eds) (1994). The nature of emotion: Fundamental questions. New York:
Oxford University Press.
Feuer, L. S. (1963). The scientic intellectual: The psychological and sociological origins of modern
science. New York: Basic Books.
Fleck, L. (1935 [1979]). Genesis and development of a scientic fact. Chicago: University of Chicago
Press.
Freud, S. (1921 [1957]). Group psychology and the analysis of the ego. In: J. Rickman (Ed.), A General
Selectionfromthe Works of SigmundFreud (pp. 169209). GardenCity, NY: DoubledayAnchor.
Grifths, P. E. (1997). What emotions really are: The problem of psychological categories. Chicago:
University of Chicago Press.
Harvey, W. (1628 [1952]). An anatomical disquisition on the motion of the heart and blood in animals.
In: R. M. Hutchins (Ed.), Great Books of the Western World: Gilbert, Galileo, Harvey (Vol. 28,
pp. 267304). Chicago: Encyclopaedia Britannica.
James, W. (1890a). The principles of psychology (Vol. 1). New York: Henry Holt and Company.
James, W. (1890b). The principles of psychology (Vol. 2). New York: Henry Holt and Company.
James, W. (1899). Talks to teachers on psychology: And to students on some of lifes ideals. London:
Longmans, Green and Co.
James, W. (1902 [1958]). The varieties of religious experience: A study in human nature, forward by
J. Barzun. New York: Mentor Book.
Jones, R. F. (1953). The triumphof the englishlanguage: Asurvey of opinions concerningthe vernacular
from the introduction of printing to the restoration. Stanford, CA: Stanford University Press.
Kemper, T. D. (1979). A social interactional theory of emotions. New York: Wiley.
Koestler, A. (1964). The act of creation. London: Hutcheson.
Kuhn, T. S. (1977). Objectivity, value judgment and theory choice. In: T. S. Kuhn (Ed.), The Essential
Tension: Selected Studies in Scientic Tradition and Change (pp. 320339). Chicago: University
of Chicago Press.
LeDoux, J. (1994). Emotional processing, but not emotions, can occur unconsciously. In: P. Ekman
& R. J. Davidson (Eds), The Nature of Emotion: Fundamental Questions (pp. 291292).
New York: Oxford University Press.
Leventhal, H. (1984). A perceptual motor theory of emotion. In: K. R. Scherer & P. Ekman (Eds),
Approaches to Emotion (pp. 271291). Hillsdale, NJ: Lawrence Erlbaum.
Lewis, H. B. (1971). Shame and guilt in neurosis. New York: International Universities Press.
Merton, R. K. (1938 [1968]). Science and social order. In: R. K. Merton (Ed.), Social Theory and Social
Structure, Enlarged Edition (pp. 591603). New York: Free Press.
Merton, R. K. (1942 [1968]). Science and democratic social structure. In: R. K. Merton (Ed.), Social
Theory and Social Structure, Enlarged Edition (pp. 604615). New York: Free Press.
Mitroff, I. I. (1974). The subjective side of science: A philosophical inquiry into the psychology of the
Apollo moon scientists. Amsterdam: Elsevier.
Polanyi, M. (1974). Personal knowledge: Towards a post-critical philosophy. Chicago: The University
of Chicago Press.
Priestley, J. (1776 [1965]). The discovery of oxygen abstracted from: Experiments and observations
on different kinds of air. In: J. A. Passmore (Ed.), Priestleys Writings on Philosophy, Science
and Politics (pp. 139150). New York: Collier Books.
de Rivera, J. (1977). A structural theory of the emotions. New York: International Universities Press.
272 JACK BARBALET
Sartre, J.-P. (19361937 [1957]). The transcendence of the ego: An existentialist theory of conscious-
ness. New York: The Noonday Press.
Sartre, J.-P. (1939 [1948]). The emotions: Outline of a theory. New York: Philosophical Library.
Scheff, T. J. (1988). Shame and conformity: The deference-emotion system. American Sociological
Review, 53, 395406.
Scherer, K. R. (1984). On the nature and function of emotion: A component process approach. In:
K. R. Scherer &P. Ekman (Eds), Approaches to Emotion (pp. 295317). Hillsdale, NJ: Lawrence
Erlbaum.
Sloan, T. O. (1971). Introduction. In: Facsimile reprint of T. Wright, The Passions of the Minde in
Generall (1630) (pp. xilxiii). Urbana: University of Illinois Press.
Smith, A. (1795 [1980]). The history of astronomy. In: A. Smith (Ed.), Essays on Philosophical Subjects
(pp. 33105). Oxford: Oxford University Press.
Solomon, R. C. (1981). Sartre on emotions. In: P. A. Schilpp (Ed.), The Philosophy of Jean-Paul Sartre
(pp. 211228). La Salle, IL: Open Court.
Tinbergen, N. (1961). The herring gulls world: A study in the social behaviour of birds. New York:
Basic Books.
Toulmin, S. (1990). Cosmopolis: The hidden agenda of modernity. New York: Free Press.
Zajonc, R. B. (1994). Evidence for nonconscious emotions. In: P. Ekman & R. J. Davidson (Eds),
The Nature of Emotion: Fundamental Questions (pp. 293297). New York: Oxford University
Press.
Zangwill, O. (1987). Freud on mental structure. In: R. L. Gregory (Ed.), The Oxford Companion to the
Mind (pp. 276278). Oxford: Oxford University Press.
Zeman, A. (2001). Consciousness. Brain, 124, 12631289.
Ziman, J. (1978). Reliable knowledge: An exploration of the grounds of belief in science. Cambridge:
Cambridge University Press.
A THEORY OF THE SELF,
EMOTION, AND CULTURE
Erika Summers-Efer
ABSTRACT
This paper uses systems theory to clarify the crucial point that there is a
basic, inborn, bodily motivation, and that a social theory of the self cannot
simply be a theory of process. By bridging across current neuroscience,
cognitive science, and systems theory, I propose a self that is fundamentally
emotional energy seeking. There are other bodily needs (food, drink, etc),
but these satiate quickly, and although they can override everything else at
moments when they are low, they are not the central switching mechanism,
the top of the hierarchy in the subsumption architecture of the self. Basing
the formation and ongoing processes of the self in the motive to maximize
emotional energy can explain the seeming conict between tendencies
towards self-consistency and the potential for creativity and change. It also
allows us to detail the mechanisms that underlie the process of individuals
drawing on culture as a resource and in turn diffusing new symbols and
meanings into the larger culture.
Theory and Research on Human Emotions
Advances in Group Processes, Volume 21, 273308
Copyright 2004 by Elsevier Ltd.
All rights of reproduction in any form reserved
ISSN: 0882-6145/doi:10.1016/S0882-6145(04)21011-6
273
274 ERIKA SUMMERS-EFFLER
SENSING AND CONTEXTUALIZING: A THEORY
OF CULTURE, EMOTION, AND THE SELF
Introduction
Dening culture as the tools, resources, or repertoires for strategic action
(Bourdieu, 1990; Hays, 1994; Sewell, 1992; Swidler, 1989; Tilly, 1993) implies
an active self geared toward problem solving, but such theories of culture do not
specify what problems we might be trying to solve. What ends are we strategizing
towards, what are the dynamics of the selves that do the strategizing, and how do
these selves use culture? In order to answer these questions, we need to understand
our basic motives for behavior. On the most supercial and immediate level, we
may say that we cannot know this information about people. We might say that
the uidity and diversity of both culture and actors prevents us from being able to
begin to understand individuals motivations. In contrast, I suggest that there is a
level belowthe specic substantive content of culture and cognition, the level of the
process of using culture. This is the level where we will nd common motivations.
The goal of this paper is to establish a process of the self that is built on the inter-
action between our basic motives and our environment. As selves, we use cultural
tools and help to create new ones through strategizing to reach our fundamental
goals and the socially constructed ones that are scaffolded on top of these.
The Self as Process
As Mead has noted (1934), the self is pattern or process rather than substance.
The key to a useful understanding of the self is to realize that we are trying to
understand a process that can take countless substantive forms, and will over any
given lifetime. The self is a perpetually shifting dynamic system that is more like
the ow of a river in which patterns emerge and disappear than a static backdrop
(Kelso, 1995, p. 1). This exibility of form is possible because the self is more
of an associater than a knower. By focusing on the organizational processes of
the self, we will begin to understand the strategizing and emergent self that we
experience.
Can a Single Force Organize the Self?
Systems that are organized by a single force are the most straightforward. They
are simple input/output arrangements that allow for high levels of predictability.
A Theory of the Self, Emotion, and Culture 275
Could the self be such a system? In order for the self to be determined by a single
force, it would have to be entirely determined biologically or socially, with no
mix between the two. We know that the self cannot be entirely determined by our
genes because DNA cannot contain this much information (Kelso, 1995, p. 14;
LeDoux, 1996, p. 72; Tredway et al., 1999, p. 114). We require input from outside
of our genes to shape the process of the self.
Similarly, neither can our environment solely determine the organization of
the self. The environment plays an important role in the ongoing process of the
self, but clearly the innate capacities and limits of perception play a role as well.
A social constructivist position tends to focus more on the content of what is
known rather than the underlying purpose of gathering information. Symbolic
interactionism suggests that individuals attach symbolic meaning to objects,
behavior, themselves, and other people, and they develop and communicate these
meanings through interaction (Howard, 2000, p. 371). The process of meaning
making is active, and we can assume purposeful, but symbolic interactionism
does not specify what the purpose might be. One could suggest that motives are
also completely constructed from ones environment, but this is impossible as we
require some orienting value of good and bad to be able to learn (Cisek, 1999;
Clark, 1997; Damasio, 1994). Even the most rudimentary classical conditioning
models require this. Where would this basic good or bad association come from,
at least initially, if not from our body (Franks, 1999, p. 158)?
A Self Organized Through Inborn Motivation
and Environmental Conditions
Some social theorists are wary of any discussion of genes or inborn capacities in
relation to behavior because of a misunderstanding that this would imply a lack
of freedom or uniqueness on the level of the individual, or that it would reduce the
signicance of the environment in determining the shape and meaning of action.
On the contrary, we will see that genetic information provides the foundation for
the role of life experience in creating our individuality. Our inborn capacity is
the source of the partial program that results in the development of a self; it helps
set a general rather than precise arrangement of systems and circuits (Damasio,
1994, p. 110). Clark (1997, p. 9) uses the example of a wall following robot
that is programmed to veer to its right until it senses contact, then it veers ever
so slightly to the left and follows the process again. The result is that the robot
follows the wall. To our eyes it looks as if the robot had been programmed to
follow the wall, but following the wall is only the nal result; it is a gloss over the
simplicity of the program. The solution is elegant because it relies on information
276 ERIKA SUMMERS-EFFLER
in the environment, not just information available internally. The soft assembly
of a partial program provides exibility by requiring extensive environmental
inuence. The solutions that emerge are tailored to the idiosyncrasies of context,
yet they satisfy some general goals (Clark, 1997, p. 44). So it is with the self.
Our genes provide general directives that are lled out only through interaction
with our environment (Brothers, 1997; Clark, 1997; Franks, 1999; Tredway
et al., 1999). Gopnik and Meltzoff point out that from an evolutionary point
of view three of the most distinctive features of human beings are their long
protected immaturity, the plasticity of their behavior, and their ability to adapt
to an extremely wide variety of environments (1997, p. 19). The self is neither
fully genetically programmed, nor completely socially constructed. It is this
arrangement that allows for the emergent quality of the self that we experience.
We know that all humans are motivated to satiate thirst, hunger, etc, and
similarly there is a social motive. Despite wide cultural differences all humans
experience socialization, indicating a basic propensity for positive interaction
as a motive for action. That is to say, people are attracted to social interaction
and this provides a motive for socialization. We can also see evidence of a social
motive in the very process of our socialization. Socialization does not produce the
same results that classical conditioning based in simple tactile pain or discomfort
would, rather it produces a highly reexive self. We care what others think and
feel about us. We try to see ourselves from their perspective and train our behavior
to this opinion more so than any physical or material reward or punishment. It is
this fundamental social motive that allows us to be able to develop self-reexivity
and consciousness as we try out different positions in our mind to evaluate our
own behavior and strategize for positive social positioning. Our self-awareness is
based not only in our awareness of the other, but also in the understanding that the
other has implications for controlling our environment. Anticipating the other in
reference to ourselves creates self-awareness (Schore, 1994). While not directly
addressing innate motivations, this is similar to what Mead said about the role of
viewing ones self from the position of the other in developing the capacity for
self-reexivity (Mead, 1934). The social motivation ties the self to the body, at
least initially, as this drive for positive social position is inborn, genetic, an evolved
trait for the survival of a species (Barbalet, 1998, p. 25; Damasio, 1994, pp. 85,
191; Geen, 1991, p. 394; Shilling, 1999, p. 552; Stevens & Fiske, 1995, p. 190).
However, this social motive also highlights the centrality of the environment, as
the goal of the drive is rmly situated out in the interactional world.
If we do not grant this inborn motivation, and leave motivation and goals to be
entirely socially constructed by the environment, not only do we have a hard time
explaining self-reexivity, we also have difculty explaining howpeople are actors
instead of direct reections of their environment (Wrong, 1961). A completely
A Theory of the Self, Emotion, and Culture 277
social constructivist model leaves little potential for emergent ideas and behavior.
If, however, we consider that we are not born blank slates, that we have certain
predispositions that are the product of evolution, the long terminteraction between
the genetics of a species and the changing environment, we will see how individu-
als actually do have an important role as a strategists and decision makers. Because
our fundamental motive is not derivative of our environment (although in real life
it can never be manifest separate from our environment) we have the capacity for
deep socialization as well as emergent and creative behavior. We do not merely
regurgitate culture; we use it to work toward the fundamental goal of positive
social positioning, and by doing so we continuously modify both the culture
we use and ourselves.
The Self as a Self-Organizing System
If we have some innate predispositions that serve as basic motivations, but the
environment is crucial for lling out the specics for how we think and operate
as well as the socially constructed motives that are scaffolded on top of our basic
drives, what sort of self could we have? The self might be random, but this could
not be the case because there would be no stability of personality or culture.
Alternately the self could be directly organized in a cause and effect fashion, but
we have ruled out both the environment and innate dispositions as completely
programming or ineffectual, so the self could not be a systemcontrolled by a single
inuence. Finally, the self could be a self-organizing system. In self-organizing
systems, there is no specic code for the formation of the system, rather pressure
from multiple forces results in ongoing pattern formation.
Self-organizing systems arise from the interaction of a large number of factors.
These systems are open, which means that they exchange energy, matter, or
information with their environment. They are also non-equilibrium; without being
pumped or energized from the outside (or inside as well for metabolic machinery)
they cannot maintain their structure (Kelso, 1995, p. 4). non-equilibrium systems
have no feedback-regulated set points or reference values that they are designed
to maintain (Kelso, 1995, p. 9). This is quite different from equilibrium systems
such as hydration and nourishment that are designed to bring the system back to
a point of stability and balance. In a self-organizing non-equilibrium system the
control parameter leads the system through patterns and is not dependent on the
pattern itself (Kelso, 1995, p. 16). The order parameter is the pattern that results
from the interaction between the control parameter and the individual parts of the
system. It is set by the micro conditions in the environment; once it is set it has
an organizing force of its own.
278 ERIKA SUMMERS-EFFLER
Looking at the self as a self-organizing system avoids many of the difculties
that sociology has previously encountered when looking at complex social
systems. We no longer have to worry about any part of the system (environment,
brain, body, genetic predispositions, etc.) being reduced to any other part; the
force of all of the different parts sustains the system. We also need not content
ourselves with merely describing the dynamics of complex systems, which has
been an impediment for theorists concerned with complex systems exhibiting
circular causality (Giddens, 1991; Sewell, 1992). Instead, we can look to uncover
control variables that stimulate the system of the self to create the patterns that
we see. By considering how shifts in the individual aspects of the system of
the self result in varying patterns of action we can explain the exibility of the
system without sacricing the potential to predict outcomes. Since we are not
relying on the linear causality that is usually associated with predictive capacity,
we are prevented from downplaying the importance of the multiple aspects of
the system.
In order to understand a self-organizing system we must determine the control
parameters, order parameters, and the order parameter dynamics (Kelso, 1995,
p. 10). We can understand the control parameter as the basic motivation behind the
development and organization of the self. Through the specication of the control
parameter we can delineate the relevant parts of the system, the aspects of the body
and environment that interact to continually create the self. Finally, we will look
to the patterns of the self, the processes that result from the interacting parts of the
system. When the picture is complete, we will begin to be able to see the potential
to anticipate how changes in parts of the system of the self would put pressure
on other parts of the system to create new patterns. When we say that the self as a
system is not linear and that multiple forces result in circular causality, this does
not mean that we need to stop with a description of the process; rather we can go
about predicting change in the system by considering counterfactuals, if . . . then
statements, about pressures on parts of the system. We will be closer to an
understanding of the self that offers predictive potential without reducing the self
to linear operations.
ISOLATING THE CONTROL PARAMETER FOR
THE SELF THE MOTIVATION BEHIND THE
ONGOING FORMATION OF THE SELF
The potential for developing a self has an evolutionary history that, like everything
in evolution, is based in survival. People are a crucial source of life support,
and the need to survive requires learning in order to deal with the social world.
A Theory of the Self, Emotion, and Culture 279
Humans are vested with self-interest, but this self-interest is tempered by a reliance
on and identication with groups (Stevens & Fiske, 1995, p. 191). Over time
characteristics associated with human social interaction have become genetically
specied (Barbalet, 1998, p. 25; Stevens & Fiske, 1995, p. 191; Turner, 1999,
p. 87). The evolution of a social orientation has happened not through individual
selection or group level selection, but rather a socially oriented selection in which
the group acts as the selection environment for evolution at the individual level
(Stevens & Fiske, 1995, p. 190). This suggests that the ability to adapt to the
group, any group rather than a particular one, has been genetically selected as
advantageous for humans (Gopnik & Meltzoff, 1997).
We are born with the neural machinery to learn in response to our social
environment through our primary emotions (Turner, 1999, p. 87; Turner, 2000).
The existence of such innate social perception is supported by evidence that certain
social stimuli contain enough information to provide beginning levels of meaning
without elaborate cognitive processes (Brothers, 1997, p. 61; Schneider, 1991,
p. 545). How can perception without cognition be possible? Representations of
the outside world cause modications in the body whenever interaction between
the organism and environment takes place, and this modication of the body
registers not in the form of cognition but rather emotions (Damasio, 1994, p. 230).
Emotions are the experience of sensing, and feelings are the awareness of emotion
(Damasio, 1994, p. 159). Our emotional machinery is geared to process informa-
tion relevant to our personal and social behavior; it does this through pairing social
situations with adaptive bodily responses (Damasio, 1994, p. 177; Schore, 1994,
p. 301).
The development of innate social drives to assist in organism survival created
the potential for self-awareness, as knowing and understanding how others felt
about us, beyond how they immediately responded to us, became relevant for
survival (Brothers, 1997). As we were concerned with the shape of others minds,
we took shape in our own. This still happens on an individual basis when each
person is born and socialized. Symbols become attached to this process through
parallel processing of emotion and symbols from the environment; through this
process we are able to develop placeholders for abstract concepts and language.
Language does not allow for self-reexivity, rather self-reexivity allows for the
development of language (Turner, 1999, p. 87).
The potential for learning though bodily responses and the resulting feelings
is not, however, the point of the system. It is an adaptive feature of a system built
to accomplish certain tasks; learning helps the system adapt to the environment
so that its goals are more likely to be achieved (LeDoux, 2002, pp. 89). As
Clark (1997, p. 1) notes, minds evolved to make things happen. The process of
learning tells us that we have such a purpose, but we must dig deeper to understand
280 ERIKA SUMMERS-EFFLER
the goal of the system, the control parameter for the self-organizing system of
the self.
Specifying the Emotional Social Motive
There are a number of suggestions about what this fundamental motivation or
social impulse might be. Some suggest a self-consistency theory. These theorists
argue that people act to be assured that the world is as they perceive it; they act
to minimize ontological anxiety or a fear of meaninglessness (Garnkel, 1967;
Giddens, 1991; Robinson & Smith-Lovin, 1992). This assumption runs into
problems when it comes to explaining the potential for social change, especially
how individuals come to envision and take risks for change. It does, however,
point to a trend that is evident in the case of consistently abused people, such
as abused children and victims of domestic violence, where individuals choose
courses of action that reafrm their position in their world, even if this position
is abusive (Hardcastle, 1999, p. 240). There is also experimental evidence that
suggests that people are motivated at least to appear consistent and integrated over
time (Abelson, 1981, p. 722). The fundamental social motive should be able to
explain the tendency towards self-consistency as well as how people can work for
imagined alternatives.
Some theorists assume that people are motivated to experience solidarity
(Durkheim, 1995 [1912]; Goffman, 1967), but if we assume this motive, we are
at a loss when attempting to fully explain power, change, and conict (Wrong,
1961). Often the creation or maintenance of a group can only be secured through
conict with another group. As Simmel asserts, conict can be central to the
foundation and preservation of group unity (Simmel, 1950). While a motive of
solidarity can explain much of human behavior, it cannot be the fundamental
motive because it cannot account for conict, which is the counterpart to group
identity formation and solidarity (Howard, 2000, p. 370).
Other theorists suggest that people are utilitarian; that individuals weigh the
relative costs and benets of different outcomes, and act to achieve their greatest
personal gain. The central motivate in classical utilitarian theory is the desire
to maximize gratication, or utilities, and to avoid deprivations or punishments
in social transactions (Turner, 1987, p. 16). Contrary to utilitarian assumptions,
there is experimental evidence that suggests that people distribute money or other
utilities in ways that favor persons in their own randomly assigned category, and
provide the least benet to persons in other randomly assigned categories. People
continue to behave in this way even when other distributions would benet them
(Swanson, 1992, p. 100). This supports the assumption that there is a social motive
A Theory of the Self, Emotion, and Culture 281
that works above individual utilitarian motives. A less specic motive towards
value preferences has been added to the utilitarian model to account for prosocial
behavior (Olson, 1965), but this solution is too vague without specifying how
value preferences are formed. As I stated above, learning, including the formation
of value preferences, suggests a response to an underlying motive, not the
motive itself.
While we all experience the specic goals of solidarity, self-consistency, and
maximizing utilities, these goals must be more surface manifestations of a deeper
social orientation because they conict with each other. None of them could be
the fundamental social motive and allow for the expression of the others. Two
theories of motivation suggest a social impulse, but allow for conict, change,
self-consistency, solidarity, and power: Collinss theory of a drive to maximize
emotional energy, and Aron and Arons theory of a drive towards self-expansion.
The two theories can be integrated, with Collins providing the goal and Aron and
Aron providing the process for reaching the goal. Aron and Arons detailing of
the process of self-expansion claries and species Collinss assertion that both
power and solidarity experiences boost ones level of emotional energy.
Collins argues that individuals are motivated to gain emotional energy, which
at a high level feels like enthusiasm, personal strength, and/or willingness to
initiate interaction, and when at a low level it feels like emotional atness,
lack of condence, and/or depression (1988). There are two types of emotions,
transient emotions that can change abruptly and often disrupt the ow of everyday
activity (for example, feelings of joy, disgust, anger, etc), and emotional energy,
which is an enduring emotional tone within the individual that reects a history
of interactions (1990). Emotional energy is drawn from two possible sources:
rst, solidarity experiences based in ritual interaction, which reect ones status
group position, and second, being on the more authoritative end of a hierarchical
interaction. Alternately, if one is excluded or subordinated in an interaction one
will lose emotional energy (Collins, 1988, 1990). We can extend this argument
to say that when individuals are low on emotional energy they may use defensive
strategies designed to minimize loss rather than maximization strategies.
Emotional energy is experienced physically in ones body and is indicative of
the interactional social reality that one is experiencing. A persons biography,
or chain of interactions over time, teaches an individual that certain types of
interactions are likely to either corrode or build their level of emotional energy
(Collins, 1988). People prefer the types of interaction rituals that have been the
most successful in the past. Ones history of experiences in particular types of
interactions determines whether one will be inclined towards solidarity, the exer-
tion of power or authority, or a combination for ones source of emotional energy.
When all is well, and ones level of emotional energy is not threatened, this process
282 ERIKA SUMMERS-EFFLER
of seeking some interactions over others happens at a level below consciousness.
This assertion is supported by research on emotion that suggests that the workings
of emotion are generally below the level of awareness (LeDoux, 1996).
The self-expansion model proposes that a central human motivation is self-
expansion, which literally means that the self takes up more social space (Aron &
Aron, 2000, p. 110). Aron and Aron suggest that people can expand themselves
through both physical and social inuence, and social and bodily identity (Aron &
Aron, 2000). We can integrate the two to say that self-expansion would produce
emotional energy.
1
Being on the more powerful end of a hierarchical interaction
produces emotional energy because it results in greater physical and social inu-
ence. Likewise, when people enlarge their sense of self in solidarity interactions
by incorporating people within a close relationship, a shared ritual, or social
identity, they enlarge their social and bodily identity and experience an increase
in emotional energy. Seeking emotional energy through self-expansion is the
non-equilibrium control parameter that energizes and leads the self-organizing
system of the self.
The self-expansion model offers an important psychological piece to the
emotional energy maximization model. The metaphor of self-expansion captures
the dynamic nature of the motivation the drive forward implied in maximization,
the reason why social movements or religions need to grow or die, and why we do
not just continue to do the same things that brought us emotional energy but use
this information as a guide for strategizing for more. It also adds some complexity
to the process of emotional energy attainment by highlighting that there is a self
that must be dealt with during the acquisition of emotional energy. There are
moments of solidarity, such as protests, religious services, or sporting events that
can carry us away in a rush of emotional energy. However, there are also people
in those same moments who stay outside, and who feel all the more alienated
when the solidarity becomes the most intense for those around them. Through
ethnographic observations of protests, I have seen how members of the press will
mingle with protesters at the beginning of a protest, before any signicant unied
chanting, singing, or marching starts taking place. However, as the chanting etc
begins, the casual connection between the press and the protestors breaks down,
and the members of the press physically move to the fringes and take up a posture
that publicly marks them as members of the press, not protestors.
The self is also important for the experience of power. Some people may enjoy
being on the powerful end of interactions while it makes others uncomfortable.
Deference is shown in these interactions, which should produce emotional energy,
but the dynamics of the interaction are more complicated than that. Helping can be
a particular type of hierarchical interaction that results in an imbalance of power.
Again, through ethnographic observation of charity organizations and interactions
A Theory of the Self, Emotion, and Culture 283
I have seen some people attracted to the same helping interactions from which
others shrink. For example, all involved in a particular organization are happy to
make sandwiches for the hungry who came to their door, but some enjoy actually
handing the sandwiches over, while it makes others uncomfortable. Although there
is the opportunity for emotional energy and self-expansion in these moments, there
is also a self that has a history of experiences that mediate the structural situation
that affects ones level of emotional energy. There is more going on than the
immediate social positioning in determining the impact on ones level of emotional
energy; through the self, we bring our history of interactions with us into each
new interaction.
Both Collinss theory of emotional energy as well as Aron and Arons theory of
self-expansion can explain how people turn to defensive rather than maximizing
strategies, such as the above mentioned cases of child abuse and domestic
violence. Aron and Aron address this directly by saying that in cases of continued
harsh rejection people cease to attempt self-expansion (2000, p. 111). Similarly
Collinss theory can be used to explain how in cases where there are consistently
limited options for gaining emotional energy, people will try to minimize the
loss of emotional energy (Summers-Efer, 2002). Trying to acclimate to a new
environment, even an apparently safer one, requires an expenditure of energy it
is an emotional energy risk, one that someone who has little available emotional
energy can afford to take. This is especially true if the new environment cuts
subordinated individuals off from their old environment, where they have the few
limited sources they can depend on, and life has taught them that there are few
better options. This can happen on the level of the individual, but also on a social
level, explaining the phenomenon of the complicity of the oppressed. We could
also anticipate defensive strategies might be more frequent in environments that
are socially chaotic (e.g. city life, intense or numerous workplace interactions).
In these environments people may be more likely to minimize interaction as a
way to recover from social anxiety or unknown potential for emotional energy
gain or loss. In such circumstances one might choose to stay home to listen to
music or read a book rather than risk social interaction that has consequences they
cannot predict.
Emotion, Body and the Social Self
We rely on the environment to indicate interactions necessary for our survival. To
satiate our equilibriumdrives (nding food and shelter, for example) and maximize
our drive for emotional energy, we need our bodies to sense our environment
to smell, taste, touch, hear, and see what is around us (Damasio, 1994, p. 125).
284 ERIKA SUMMERS-EFFLER
I would add that we also use our bodies to sense social position. We cannot perceive
separately from the body. This is obvious in the case of stubbing ones toe, but it
is also true in the interactions that would appear to be purely emotional and social.
Consider the phrases about being jilted by a lover, it was like a knife in the heart
or a punch to the gut, or losing a loved one is like having your heart ripped
out of you. People can feel like doubling over or nauseated from the sensation of
shame. The list could go on. These social interactions have emotional affect that
is felt in the body.
The experience of feelings offers us conscious access to our visceral and
musculoskeletal states as they are affected by the innate sensing mechanisms re-
sponding to our expectations and our immediate environment. The brain is directly
connected to the body and only indirectly to the arbitrary culturally constructed
symbols that the brain uses for recognition and more elaborate cognition. What
comes rst constitutes a frame of reference for what comes after, thus feelings
and emotions are a primary guide for the brain and cognition (Damasio, 1994,
p. 159). Emotions are the aspects of the self that are the closest to the body, oldest
in evolution, and most central to any process of reasoning. This places the body
and its emotional signals at the foundation of the highest reaches of reason and
creativity (Damasio, 1994, p. 200).
The James/Lange theory of emotion suggests a model for a bodily emotional self
(see Barbalet, 1999; Hochschild, 1983; James, 1890). James argues that emotion
accompanies the response to physical stimulus, either novel or physically recalled
through memory; the cognition that follows the emotion results from further
processing.
2
The James/Lange theory focuses primarily on solitary activities; for
example, James proposes that the act of running makes one afraid. My argument
suggests that social interaction is the more important application of the theory.
For example, getting into rhythm with someone laughing makes us happy.
Recent work in emotions supports Jamess assertions about the bodily founda-
tion of emotions. Ekman has demonstrated that information about our emotions
is available through minute facial expressions and body gestures, even when we
are motivated to cover our emotional experience (1992). Turner has noted the
centrality of body language for communication (2000). Hateld et al. (1994)
have found that when we mimic anothers micro emotional gestures (such as facial
expressions), we call up the same emotion in ourselves that we are witnessing
in the other person. They also found that people are inclined to mimic the micro
gestures of other people, especially people who are energy sinks or charismatic.
3
Bodily emotional information, which usually remains below the level of aware-
ness, is the basis for emotional contagion as well as the production of reciprocal
emotions in the case of conict (Hateld et al., 1994). As I noted above, there
are ndings to suggest that we perceive social information with little cognitive
A Theory of the Self, Emotion, and Culture 285
processing (Brothers, 1997, p. 38; Schneider, 1991, p. 545). Such empirical work
suggests that emotional information about our social position and emotional
state can be indicated and passed physically through micro gestures. Emotional
information, which is connected to the body and perceived directly, is primarily
concerned with assessing social position in terms of power and inclusion.
Reconciling Multiple Drives: The
Subsumption Architecture of the Self
Mead (1934) denes the self as primarily cognitive and symbolic, but I suggest
that this is only one aspect of the self that sits on top of older evolutionary drives
and innate predispositions that also have direct bearing on our behavior. While
maintaining social ties and maneuvering the social landscape are important for
survival, they came later in our history than the basic necessity to bring needs for
water, food, etc., into equilibrium. These are the older less conscious aspects of
our survival map, but they are still relevant to collecting, storing, and processing
information. If we restrict our consideration of the self to the symbolic realm we
ignore other learning capacities that we use as a guide for future behavior.
If we have a drive for food and a drive for social positioning that results in
emotional energy, sometimes they are not in alignment; how are they reconciled?
The image of the self as a central computer solving these conicts is too clumsy
and slow to account for the speed with which we process information and react.
Articial intelligence work has offered a more elegant solution of subsumption
architecture. Subsumption architecture is comprised of a number of distinct
activity directed subsystems. The interaction between distinct layers is restricted
to simple signal passing one layer may support, interrupt, or override the others
(Clark, 1997, p. 13). The systems are fairly independent until there is a conict
where they must be reconciled, and then a hierarchy is revealed that is only there
in a time of conict. Subsumption architecture does not imply any overall central
plan, but rather is extremely adaptive to the environment because it is based on a
simple organization of control parameters (Clark, 1997, p. 14).
Because drives such as hunger and thirst are only seeking satiation, baseline
equilibrium rather than maximization, the non-equilibrium drive to maximize
emotional energy sits at the top of the hierarchy when drives conict. We need
to drink to live, but we live for social success because it is a non-equilibrium
drive that is always pulling us further along, or at least protecting us from further
loss. A need based in maximization is at best temporarily satised; this means
that often other drives will take a back seat to the social drive if they are at least
moderately satised. However, if one of the older drives becomes life threatening,
286 ERIKA SUMMERS-EFFLER
it will become dominant as long as it does not represent a major emotional energy
risk. For example, if we are thirsty we may give up a positive social interaction
as long as it does not incur long-term social damage. On the other hand, we may
give up the interaction if we are only a little thirsty if the social interaction is not
anticipated to be particularly successful or important. In order to anticipate which
choice would be made, we would have to know how thirsty one is as well as how
important an individual anticipates a given social interaction will be.
Occasionally a satiation function can be attached to the social motivation
through a symbolic connection to social success. This can create problems as
behavior that normally falls into the background becomes more dominant and
potentially damaging to the body. For example, this is likely what happens with
eating disorders as the size of a persons body plays a central role in social success
in particular cultures. It may also help to explain addictive behaviors as well,
as there is evidence that addictions are ways of dealing with social emotions
and anxiety (Endler & Kocovski, 2000; Golwyn & Sevlie, 1992; Mitra &
Mukhopadhyay, 2000).
The hierarchy of drives that emerges in conict can also begin to explain
altruistic behavior and even self-sacrice. In ethnographic studies of altruistic
social movement and charity organizations, I have seen that self-sacrice can be
tremendously rewarding, in terms of both status and solidarity. Sometimes altruis-
tic behavior is rewarding because of the opportunity for self-expansion in the form
of prominence, social power, and social status. For example, being arrested for
civil disobedience and doing time in prison as a prisoner of conscious affords one a
sense of moral integrity as well as status within an activist community. In the time
between conviction and the beginning of his or her sentence, the future prisoner
of conscience does a tour of their sympathetic community, sometimes ofcial
speaking engagements, and other times informal gatherings. During these times
the individual is openly praised, encouraged, and held up as a model of morality
and courage.
Rick holds court in the kitchen, standing up eating blueberry pancakes that have been made
especially for him. No one else is eating, and we all stand around facing him, focusing on him,
listening to him, and talking to him. I can only describe himas glowing with pride as we prepare
himfor the three months in federal prison he will start in two days. We are his faithful attendants,
and we feel proud of our association with him. He looks around with forked pancake poised in
air, a huge grin, and says, Ill have to go to prison more often. Fresh blueberry pancakes and
everyone is so nice to you (Field notes October 12, 2002).
During jail time, those who visit the prisoner gain status through the interaction.
It marks them as in among the highest ranks of the movement. Letters written
from prison are published in the movements publications and are often read aloud
during organizational gatherings. After serving time, the prisoner of conscience
A Theory of the Self, Emotion, and Culture 287
does a tour similar to the one before serving time, and is again heralded as a hero.
The status gained from going to prison endures, and those who do time remain
within the movements elite.
I have also seen the emotional energy rewards that can come fromself-expansion
in the form of solidarity, of expanding oneself to include another.
Kara begins to tell a story that Ive heard her tell many times. Ram Dass writes about minister-
ing to those who are dieing. He says that you leave all of your roles at the door: doctor, nurse,
priest, friend, and you go in there as one human being to another. You are both human beings,
looking out the eyes of different bodies and sharing the same human condition. If you are
really lucky, you both realize that you are really one, that there is just one person in the room.
As always, her eyes are shining with tears at the end of the story, not because the story is about
death and sadness, but because it is about a kind of human connection. It is this connectivity
that shes always looking for when she interacts with the people she serves (eld notes
April 18, 2001).
Understanding that in these moments others actually become included within
ones self suggests why these interactions tend to be the most generous and least
demeaning of charity interactions. In these moments, the helpers feel the need of
the other as their own, and get a rush of emotional energy that comes from this
feeling of self-expansion.
EMERGING PATTERNS OF THE SELF
The Purpose of the Self: Solving the Emotional Energy Problem
The body, brain, and environment interact to create the order parameters, or
the self-organizing patterns, of the self. This happens through the inuence of
multiple innate control parameters (thirst, hunger, etc), the most important of
which is the drive to maximize emotional energy through self expansion. The
control parameters set up problems to be solved: what are the best ways to go
about meeting our basic needs? How should one interact in an ever-changing
environment to maximize ones emotional energy through self expansion? The
properties of the self arise for the purpose of solving these problems.
The self is a strategizing system (Higgins, 2000, p. 5; Stevens & Fiske, 1995,
p. 189), which develops out of our ability to anticipate our environment. If
our environment became completely chaotic we could predict the collapse or
disintegration of the self. The mechanism of continued problem solving towards
future emotional energy and self expansion organizes the self, thus the self is future
oriented. The present is merely an informational blip in the ongoing process of the
self. This is similar to Meads point that we do not experience the present; rather
288 ERIKA SUMMERS-EFFLER
Fig. 1.
we are primarily focused on the past (i.e. by referring to the me) (1934). Peirce
likewise suggests that we are not focused on the present; rather we are oriented
towards the future (i.e. by referring to the you) (Wiley, 1994). As Wiley suggests
(1994), we are both: we are focused on the future, but our possible alternatives and
tools for interacting are drawn fromour personal history, our past. All strategizing,
choices, and actions are embedded in the owof time and an ever-changing not en-
tirely predictable environment (Emirbayer &Mische, 1998, p. 966). Strategizing is
based in what we experiences as options. These options come from our immediate
environment as well as our personal history of past environments. The system of
the self responds to the environment, and in doing so shapes the future of the self
by limiting what the self can call on for future strategizing. How do we develop a
personal history? Howdo we make sense of our environment and store information
relevant to our experience? These questions are getting at the part of the self
that becomes uniquely ours, the part that is transportable and transposable from
situation to situation. This is the part of the self that constitutes our own personal
cultural tool kit.
In order to understand the dynamics that comprise the self as a self-organizing
system we need to understand how the self lters information, how it exibly uses
information, and the conditions under which the self is stable and under which
the self changes.
Figure 1 depicts the entire systemof the self. Because the self is a self-organizing
system, the entire sub processes (labeled: A, B, and C) are ultimately related to
each other. However, for the sake of clarity I will discuss parts in isolation.
Sensing The Stimulating Control Parameter
I begin with the process that is closest to the motive to maximize emotional
energy (EE) through self-expansion, which is the control parameter of the system.
What I refer to as sense, is the innate bodily response to an inborn motivation
A Theory of the Self, Emotion, and Culture 289
Fig. 2.
to improve our social position. This is the part of the self that is genetically
predisposed and unreexive. Sense is the bodily/emotional response to ones
environment that indicates social position. It is the impact of our environment on
our level of emotional energy.
Figure 2 illustrates the directness of the relationship between our face-to-face
interactions and their bodily affect on us. As Brothers points out, humans
are extremely responsive to the immediate social environment (1997, p. 28).
Researchers have found cells in parts of the brain that are socially responsive;
these cells are particularly responsive to faces (Brothers, 1997, p. 38). These
parts of the brain that are socially responsive are also associated with emotion
(Brothers, 1997, p. 27). Schore goes so far as to say that,
switching of psychobiological states in response to specic socio-environmental conditions that
are signicant to the individual represents the essential adaptive function of emotional behavior
(1994, p. 301).
There are connections in the orbitofrontal region of the brain between the
processing of social interaction and the brains opioid self-reward system (Schore,
1994, p. 91). Though the release of opioids, positive social response motivates
fusing with the source of satisfaction (Schore, 1994, p. 241). Alternately, shame
results in a rapid de-energizing state that produces deation, loss of energy,
withdrawal, recoiling, and attempts to hide (Schore, 1994, p. 241; Scheff, 1990).
These ndings suggest that social information is processed directly; it has meaning
related to social position that is inherent, as well as the more developed meaning
provided by further cognitive processing. Such conclusions are supported by
ndings that have found correlations between damage to brain areas associated
with emotion and diminished social skills (Brothers, 1997, p. 53).
Although based in emotion, the process of sensing has cognitive-like properties
in that it is a way of assessing or perceiving, but this process is not symbolic.
The indicators for social position and emotional energy are bodily and are stored
bodily. This is similar to other senses, like touch or smell, which are stored bodily
rather than symbolically. Our overriding motivation is to maximize emotional
290 ERIKA SUMMERS-EFFLER
energy, so the emotional energy sensing part of the self is our primary orientation
towards our environment. The sensing aspect of the self is much like what
Mead referred to as the I in that it gives the impression of being spontaneous,
impulsive, and emergent because it does not rely on a history of cognitions.
Instead it relies on the level of stored emotional energy and the experience of
bodily social interaction, such as the patterns of gaze direction, voice tones, bodily
gestures, and rhythms that physically convey our social position (Corsaro, 1982;
Ekman, 1992; Hateld et al., 1994; Katz, 1999; Kendon, 1990).
We have an inborn motivation to maximize emotional energy through
self-expansion, and this motivation acts as the control parameter energizing
the self-organizing system of the self. We begin with what Damasio (1994,
pp. 103104) calls innate dispositional representations based in the subcortical
areas of the brain: the hypothalamus, brain stem, and limbic system. Dispositional
representations are the inborn potential to activate other ensembles of neurons in
order to create changes in the body and the brain. We have dispositional repre-
sentations that are sensitive to social position (Brothers, 1997; Schore, 1994); we
activate this potential through interactions with our environment. For example, if
we are children and excluded on the playground, our inborn potential responds to
our social environment and activates other ensembles of neurons that produce the
changes in the body and brain that are associated with the shame resulting fromthe
loss of social position. The acquisition of newknowledge is built upon these initial
predispositions through somatic markers, where information from our environ-
ment is associated with and triggers dispositional representations (Damasio, 1994,
p. 180). Through this process other more complex cognitive aspects of the self are
built indirectly on varying levels of emotional energy detected by the sensing part
of the self.
Contextualizing Developing the
Order Parameters of the Self
Context is the symbolic information from our environment that is associated
with dispositional representations bodily sensations; we use this information to
strategize and negotiate future social interaction. Emotional information is useful
to us because it marks out symbols for negotiating social interaction. Alternately,
symbols are useful to us because they help to guide future interaction by providing
working theories of how our world works. The process of connecting sense to
context is based in the plasticity of the brain. Neuroscience increasingly empha-
sizes the plasticity of neural structure and its close relation to experience (Gopnik
& Meltzoff, 1997, p. 219). We are born with an innate capacity for synapses to
A Theory of the Self, Emotion, and Culture 291
record and store information (LeDoux, 2002, p. 8), and the contextualizing aspect
of the self is born of our innate ability to learn and adapt to our environment.
Somatic markers are created when information from our environment is paired
with our innate dispositional representations. Social actions elicit amygdala
activation, which in turn writes the social experience of the moment into neural
assemblies that encode social features and links them to action dispositions
(Brothers, 1997, p. 61).
Brothers provides an example of how somatic markers are formed and used.
. . . Suppose I meet someone and have a very unpleasant interaction with that person. During our
encounter, the persons facial features are being registered in my visual cortices, in a widespread
pattern of neural activity. At the same time, because of the nature of the interaction between us,
linking regions in my brain such as my amygdala, in addition to sending signals that affect
my bodily state, is also sending signals to the nucleus basalis. The nucleus basalis showers
my sensory brain with acetylcholine (including the parts of my brain that register voices and
motions), encouraging the patterns of activity now active to become stably linked. As a result,
the particular neural pattern associated with this particular face and voice will become easy to
reactivate as an overall pattern (Brothers, 1997, p. 60).
Through parallel processing of somatic and symbolic information, emotions in
the body are connected with symbols; these symbols are then used to guide future
action. In the example above, the symbols would be used to avoid the particular
person or the type of interaction that resulted in unfavorable social positioning.
The amygdala, orbital frontal cortex, and the cortex of the temporal lobe
are essential for attaching appropriate emotional signicance to environmental
stimuli (Brothers, 1997, p. 43). Because of their connections with the rest of
the brain, they act as a kind of interface between impressions coming in from
outside and bodily states. They produce rapid, nely tuned physical responses to
external events, especially social external events (Brothers, 1997, p. 46). Socially
responsive cells in the cortex that wrap around the temporal lobe are connected to
parts of the brain that are responsible for laying down memories (hippocampus)
and parts of the brain that dictate bodily responses such as hormone production
and the cardiac and circulatory changes that go along with ght or ight responses.
Social interaction is doubly privileged because it provides immediate access
to storage in memory on the one hand, and immediate access to behavioral
dispositions on the other (Brothers, 1997, p. 41).
Somatic markers activate neural connections that make us feel as if we were
experiencing an emotional state due to an interaction between our body and the
environment (Damasio, 1994, p. 155). This is the neurological foundation for as
if loops the memory/bodily response connection enables us to anticipate the
impact of social interaction through calling up an internal sensation as if the
external event had taken place. The trigger is in the form of internal representation
292 ERIKA SUMMERS-EFFLER
Fig. 3.
of meaningful symbols though memory traces laid down in response to external
social stimuli. As if loops generate the feeling associated with the external
stimuli by using symbolic triggers associated with previous social encounters
to call up milder forms of bodily response than those that would result from
immediate external stimuli. Neurons work collectively to encode aspects of the
world. When the memory of a particular episode is invoked, subsets of neurons
that were active together during the episode are reactivated, thereby recreating a
representation of the event internally (Brothers, 1997, p. 58).
Figure 3 represents how bodily response to social positioning and acquired
dispositional representations in the form of somatic markers are the foundation
for an as if (Damasio, 1994) system that produces meaningful symbols that help
us to strategize for emotional energy and self expansion in future interactions.
Exploratory brain surgery provides evidence for the importance of as if
loops that connect environmental information with bodily states. When brains
are probed, patients reveal that social/emotional memories are not stored in story
forms but rather sensory impression, bodily changes, and feeling states (Brothers,
1997, p. 50). Those undergoing the probing of social memories report that it
brings about feeling appropriate to certain social situations (Brothers, 1997,
p. 51, emphasis in the original). The feelings that these patients experience during
the brain probe are response dispositions specically tuned to particular social
situations (Brothers, 1997, p. 51). Sandra Bartky (1990) provides an example of
such a body/social/emotion connection by pointing out that shame is physically
remembered like a punch to the gut, and it can make one blush at remembering.
4
Through the process of sensing and contextualizing, emotions are arbitrarily
connected to actions and symbols. For example, there is nothing inherent about
the word apple that it should be attached to what we know as an apple. This
cognitive exibility in the relationship between symbols, action, and meaning is
adaptive because it helps us meet the needs of our particular environment. Pretend
play, learning to associate arbitrary rules and characteristics with particular
symbols and then transposing them (for example role playing teacher or parent)
A Theory of the Self, Emotion, and Culture 293
prepares a child for social life (Brothers, 1997, p. 29). Childrens pretend play
is the external and obvious development of the as if loops that we rely on
smoothly and implicitly in adult social life.
By pairing symbolic information from our environment with inborn so-
cial/emotional responses, we create experience-driven, non-inherited, arbitrary
connections between symbols and emotions; we create somatic markers that that
help us to strategize by acting as new dispositional representations responses
that signal the activation of patterns of neural ring that bring about changes
in the body. Symbols associated with negative emotions have the power to
reenact painful body states, so they become automated reminders of likely
negative consequences (Damasio, 1994, p. 180). When positive feelings are
associated with symbols they act as beacons for future strategizing. Knowledge
and understanding are based primarily in emotion and secondarily in cognition
(Cisek, 1999). The creation of somatic markers through the association of
symbols with innate emotions gives symbols their meaning. However, despite
the fundamental importance of emotion, meaning is still in large part socially
constructed as the symbolic information that is paired with the emotion to create
meaning is completely dependent on ones environment.
Selecting Symbols for As If Loops
For every social interaction there is almost limitless information available in our
environment that could be associated with it. How will some be processed as
relevant to a uctuation in emotional energy to serve as triggers for as if loops
while others will be left unattended in the background? This is the question that
ultimately ties the sensing and contextualizing parts of the self together as one
self-organizing system that constitutes what we understand to be the self. As
perception is a process that is part of the system of the self, it too is stimulated
more or less directly by the control parameter of maximizing emotional energy
through self-expansion.
Figure 4 depicts this connection between our ever-evolving theories of the
world, perception, and our engagement in face-to-face interaction. Perception
is tied to this goal, so it is not objective and abstract, but part and parcel of the
continually developing strategies to maximize emotional energy.
One answer to this question of what information will be processed as relevant
can be explained by the salving principle, which states that the greatest source
of instability in a self-organizing system will dominate (Kelso, 1995, p. 8). The
control parameter is the source of stability or instability, and in the case of the
self this refers to the motivation to maximize emotional energy through self
294 ERIKA SUMMERS-EFFLER
Fig. 4.
expansion, meaning that the greatest source of instability in emotional energy
(+/) will drawattention and focus. For example, there may be a number of social
interactions occurring at a given moment, but the one that is associated with the
greatest shift, either + or in EE, will be associated in the continually modied
as if loops.
The other answer to this question of what information will be made relevant
is that our past experiences frame our present social environment. Experience
brings about patterns of anticipation and thus focus, and this focus lters future
experience. Our developing theories of the world produce conceptual structures
that limit the information that we nd relevant for negotiating social interaction.
Sensory appraisal is necessary for emotional response, in return emotional
response is necessary for an event in the environment to achieve signicance
(Schore, 1994, p. 110). Attention is critical for bringing about changes in the
characteristics of the neural assemblies encoding experience (Brothers, 1997,
p. 59). Organization of sensory experience changes as a result of learning. The
fundamental key then in both perception and meaning is the focusing of attention
produced by the emotional outcomes as past events. Options for awareness
are based on ones history.
As if loops manifest as habit; they are working theories that we use to
strategize social interaction. The micro forces in the environment set the direction
of the pattern, but once the as if loops are formed and patterns develop they have
a force of their own. The focusing of attention is a central aspect of this force.
Internal models of ones social environment are representations that enable an
individual to have expectations and evaluate interactions (Schore, 1994, p. 317).
Our acquired as if loops act as an editor that lters information from senses
that are socially motivated (Brothers, 1997, p. 12). The foundation of our self is
this ability to store representations of encounters, environmental conditions, and
social outcomes, in order to frame our environment and anticipate consequences
(Schore, 1994, p. 247).
A Theory of the Self, Emotion, and Culture 295
Using and Updating Our As If Loops
Like the pragmatists pointed out long ago, patterns of behaviors represent what
works (Dewey, 1991 [1909]). We are theory builders, continually developing the-
ories of how the world works in response to a continually changing environment.
Gopnik and Meltzoff (1997) detail this process for infants and children, but there
is no reason to believe that the fundamental process of gathering information
about our environment in order to better understand how our world works stops
at any point in development. As if loops are working models of our social
environment; all of our knowledge is ultimately based in a practical understanding
of how to improve our social position or avoid diminished social positions. We
develop our as if loops in response to evidence and use them to understand new
problems. They inuence what new evidence we notice and how we perceive it,
and they allow us to make predictions about a wide variety of evidence (Gopnik
& Meltzoff, 1997). These theories we build necessarily falsely narrow the world
and limit the sorts of inductions we can make, but at the same time by making
our environment comprehensible the process allows us to make inductions. These
theories are again revisable in the face of the induced conclusions constrained by
the theory (Gopnik & Meltzoff, 1997). In this way we forever move towards the
future, but our perception and strategizing skills keep us a step behind and tied
to our past.
The continual updating and modifying of the as if loops that operate as
theories of our environment constitutes the primary activity of the self. The
process of the self may produce relatively stable patterns, but it is never at rest.
This dynamic process tied to the past gives the self both exibility and emergent
properties. History, not spontaneity or agency, accounts for both the feeling of
emergence and new solutions to interactional challenges. Kelso point out that
rather than computing, our brain dwells in metastable states, or what we might
think of as working theories of the world. Because the theories are never right,
and our basic motive is never satised, the developing theories are poised on the
brink of instability where they can switch exibly and quickly. The instability of
our theories orients us towards the future so that we are not simply reacting to the
present (Kelso, 1995, p. 26). Thus we are able to stay exible and oriented towards
the future because of our motive to maximize. Only the patterns that emerge
through their usefulness support stability; the underlying motive is not immedi-
ately connected to stability. The search for better theories is its own internally
driven motivation.
As if loops provide efciency in strategizing for emotional energy because
they allow us to bypass relying on immediate body/environment interactions by
producing the feeling within the brain alone (Damasio, 1994, p. 155). Emotion
296 ERIKA SUMMERS-EFFLER
associated with a symbol produces anticipation, aids in prediction of social
response, and focuses attention (Schore, 1994, p. 111). However, an as if loop
is not as strong as the connection between the real-time interaction of our bodies
and our environment, which means they can be modied.
If things do not go as predicted and social positioning results in a ux in
emotion created by a real-time body/environment relationship rather than the
as if process (indicating that the as if was not an accurate guide), new
information will be turned into a somatic marker and the as if system will be
updated. Inaccurate prediction of the outcome of a social interaction, or failed
interaction strategies, can result in updating the as if loops that constitute our
working theories of the world. Conict will bring about immediate modication
of the loop, but lack of reinforcement will slowly wear down as if connections.
Similar to Deweys position, the process of establishing and reestablishing as
if loops suggests that we continually respond to our environment, and do not
act in a xed means-ends progression (Dewey, 1991 [1909]). Damasio suggests
that as if loops can only be established through personal experience, but we
can go back to the discussion about how the self extends into the environment for
problem solving and emotional energy seeking to conclude, like Higgins (2000,
p. 5), that we can learn from observing those with whom we identify.
Personal History and Style: The Substance and
Methods of the Contextualizing Aspect of the Self
The contextualizing aspect of the self has layers; some are closer to the fundamen-
tal directives of our basic motivations and others are built up many layers on top of
our inborn equipment for negotiating our environment. Our methods and general
strategies for maximizing emotional energy in interaction are more tightly tied to
the inborn equipment, and are more emotional, less symbolic, and less likely to
become conscious or change than the more surface symbolic information. Swidler
points out that people do not proceed with an action piece by piece, but that
action is necessarily integrated into larger strategies for efciency and exibility
(Swidler, 1989). I refer to the level of methods for seeking emotional energy
as personal style.
The contextualizing aspect of the self also contains the history of the symbolic
content that activate specic as if loops though the impact of past outcomes
of interaction on ones level of emotional energy; I refer to this store of symbols
as personal history. For example, we may operate in the world of education
where specic symbolic information is stored to enable us to negotiate this
environment (this is our personal history), but based on our history of interactions
A Theory of the Self, Emotion, and Culture 297
in education we may tend towards power, solidarity, or defensive strategies as our
pathway for maximizing emotional energy or minimizing the loss of it (this is our
personal style).
Despite the exibility and uidity of the self, our personal style of interaction
gives us a sense of consistent perspective and stability. The consistency in process
and slowly evolving as if loops provide a sense of ownership and identity
associated with the self. Although personal style is more constant than personal
history, it still responds to changes in the environment, but more reluctantly
than the substantive knowledge of our personal history. We have an incentive to
continue with our generally successful methods. Interaction style is a skill, so a
change is a costly retooling. It requires dramatic events to affect the very style of
the organization and operation of the self. Religious conversion would have this
effect, and so might a signicant direct attack on ones way of interacting rather
than on what one said or did.
Our store of symbols that comprise our personal history are continually put
to the test and afrmed or modied. We begin life with a sort of Jamesian
preorganized mechanism of emotional response, but then context specic
information is scaffolded on top of this. We continue to experience these initial
emotional responses to basic environmental conditions, but as our personal
history develops, we will have a more individualized store of information to use
in strategizing to maximize emotional energy. The content of our personal history
is relatively situational and uid compared to our personal style, but compared to
the sensing aspect of the self (which is experienced as a level, only tied to history
in the level of emotional energy that one brings into the interaction, and easily
modied by ones present interaction) our personal history is modied slowly.
New symbolic information has to be reconciled with the old, thus this part of the
self lags behind the more spontaneous emotional part.
Within Self Communication and Conscious Thinking
The self is our process of controlling and organizing our behavior in anticipation
of our environment. Our capacity for predicting our environment and controlling
our behavior develops long before we develop a grasp of complex symbols and
language both in phylogeny (evolutionary history of a species) and ontogeny
(developmental history of an individual) (Cisek, 1999, p. 135; LeDoux, 2002,
p. 11; Tredway et al., 1999, p. 111). Symbols and language are the products of
our efforts to strategize and control our environment, not the mechanisms that
produce our ability to control. We need to move language and symbols from the
center of our understanding of the development of the self, and replace them
298 ERIKA SUMMERS-EFFLER
with motivations, bodies, and emotions. The self is centered in the body, and
understanding this prevents us from overly emphasizing symbolic and linguistic
aspects of the self. Because we function primarily through as if loops, the basic
form of within-self communication is associative rather than linguistic. There is
no future, past, or symbolic content in the sensing aspect of the self, so there is
no dialogue between the sensing and the contextualizing aspects of the self. The
internal dialogue that Mead suggests constitutes the self all takes place entirely
within the contextualizing aspect of the self, as symbolic content from ones
immediate environment and personal history are reconciled. This process is more
deliberate and slower than the usual process of the self.
Highly reexive linguistic thinking dominates the self only if we dene the self
as Mead does, as the conscious self-reexive awareness of ones self as on object
(1934). It is a mistake to dene the self so narrowly because the internal dialogue
of conscious thinking that Mead suggests comprises the self is too slow a process
for us to rely on, especially when we are responding to danger (Turner, 2002),
or wrapped up in an activity that seems to be pouring out of us or happening
spontaneously. Should we believe that when there is no internal dialogue we have
no capacity to organize our behavior, or that we are somehow operating on some
level outside of socialization, on some authentic presocialization impulse? On the
contrary, the process of socialization operates in a much more efcient way than
a self reliant on internal dialogue would suggest. Our impulse to act is already
tied up with our history and expectations in the form of as if loops that pair
symbols with emotions. This process of socialized perception happens before we
have an impulse to respond.
Consciousness and Self-Reexivity
As the contextualizing aspect of the self is a store of future action strategies for
our use as actors, all of the information within it is processed and stored relative
to the self. This information is not only self-referencing, but also potentially self-
reexive. As discussed above, our social drive creates the potential for heightened
self-awareness and reexivity, but we do not spend most of our time in a state of
self-consciousness. If events go according to our predictions, no new information
in our environment is singled out as important and the process of contextualizing
is more of an act of keeping track on a level below reexive consciousness.
Figure 5 illustrates how we proceed with little internal dialogue and self-
reexivity when we are able to successfully rely on our existing as if loops
for negotiating interaction. For example, consider driving a familiar route. While
driving is a complex activity, when we are familiar with a drive it takes little
A Theory of the Self, Emotion, and Culture 299
Fig. 5.
conscious awareness to get from point A to point B. When we are absorbed,
we experience a general buzz of bodily feelings (Damasio, 1994, p. 150) and
images (Turner, 1999, p. 102; Turner, 2000, p. 109). Even when we engage in less
deliberate self-talk or casual talk with others, thinking takes the form of a habitual
or supercial consciousness where we call on symbols, but entirely within the
ow of context; it does not require extensive reexive thinking.
Conict between our expectations and actual occurrences can bring about con-
sciousness, reexivity, and deliberate dialogical thought. Subtle unpredictability
brings about slight modications of existing as if loops. More substantial
substantive/symbolic conict requires the dismissal of old or creation of new as
if loops.
Figure 6 depicts howconict between what we expected and the actual outcomes
of interactions can produce conscious thinking and the modication or updating of
our as if loops. This is the sort of situation that would result in internal dialogue,
Fig. 6.
300 ERIKA SUMMERS-EFFLER
as incompatible symbolic information is reconciled. The voices or positions in
the dialogue do not represent emergent impulses and internalized expectations,
but rather two or more sets of internalized expectations or perspectives. It is in the
reconciling of two or more conicting expectations, two or more as if loops,
or in the failure of one as if loop that we have the potential for making choices
and creating change. A conict in style of emotional energy seeking brings
about particularly extended self-reexivity. This sort of conict between the
expectations for a method of interaction and actual results is more pronounced and
brings about more self-reexivity than a substantive/symbolic conict because it
has the potential to have an impact beyond the scope of the immediate context for
ones ability to maximize emotional energy. This argument that conict brings
about consciousness and reexivity is similar to Mead and Deweys assertion
that frustration of impulses or an action brings about conscious thinking (Dewey,
1991; Mead, 1938), and Meads argument in Philosophy of the Present (1932,
p. 71), that conicting responses from the environment create higher levels
of reexivity. Garnkels (1967) breaching experiments also demonstrate how
unexpected interactions create heightened consciousness and reexivity.
Because consciousness is situated in the contextualizing aspect of the self, it
is inherently self-referencing and self-reexive. Despite the centrality of inborn
emotional mechanisms for setting and keeping the organizational process of the
self in motion, our capacity for self-reexive awareness cannot exist outside of
our history of past interactions (Tredway et al., 1999, p. 114). Our experience of
our emotions is ltered through our environment and our expectations. Conict
between what we personally expect and what actually happens brings about
consciousness and self-reexivity. The increased awareness caused by conict
offers us exibility to respond to our environment based on our particular history
of interactions in the environment (Damasio, 1994, p. 133). This means that our
feelings, or our awareness of our emotions, are socially constructed. Understand-
ing this highlights the importance of our own personal history, as what brings
about consciousness has everything to do with our past experiences. We bring part
of our past with us in the form of strategies that help us negotiate our environment,
and in doing so we have a potential for consciousness and feeling that is
uniquely ours.
The Need for Self-Consistency and the Conservative Bias of the Self
Above I discussed the conicting arguments about the nature of our motivation (see
Turner, 1987); some suggest self-consistency (Garnkel, 1967; Giddens, 1991;
Robinson & Smith-Lovin, 1992), while others advocate a motive to maximize
A Theory of the Self, Emotion, and Culture 301
utilities (Homans, 1961). I suggest that the conict between these two positions
can be reconciled by considering the organizational processes of the contextu-
alizing aspect of the self separately from organizational process of the self as a
whole. As a whole, the self is primarily motivated to maximize emotional energy,
but the contextualizing aspect of the self has dynamics of its own born of attempts
to predict environmental conditions and control behavior. We can only predict
if there is a feeling of stability and reliable relationships between symbols and
meaning (Cisek, 1999, p. 137). Our need to predict is a conservative inuence that
keeps us tied to the past instead of totally uid and dependent on current ows of
emotional energy. A need for self-consistency is at the heart of the usefulness of
the contextualizing aspect of the self. Through our need to anticipate, relationships
become routinized and form what feels like a rm structure. We can see this
in Garnkels breaching experiments as people rushed to bring things back to
usual after the disruption of normality (Garnkel, 1967), and in experiments that
demonstrate that people can interact to maintain consistency with their self-image,
even when it is negative (see Robinson & Smith-Lovin, 1992).
The potential for instabilities is omni-present because there is a push and pull
of cooperation and competition different forces with dynamic and open systems
(Kelso, 1995, p. 11). In the case of the self, there is cooperation among the parts of
the self to provide a sense of self-consistency so that the self can accurately predict
its environment. However, this is in competition with the fundamental control
parameter to maximize emotional energy because the position that one has held
consistent in may not reward the greatest amount of EE. The ability to strategize
makes us conservative, adaptive, and cognitively complex. We look to our tools
to act, which are necessarily conservative because they are based in the past.
However, this motive towards self-consistency is only on the level of the symbolic
self, which rests atop the fundamental drive to maximize emotional energy. This
means that self-consistency is not the bottom line. There is instability in that this
pull towards predictability is weighed against the potential for more emotional
energy from an alternative situation. Subordinate positioning undermines the
stability of the system by giving individuals a reason to risk the unknown for a
better social position.
5
When the force is in competition with the stable pattern
overrides the one creating stability, it leads to rich, irregular, and emergent dynamic
behavior (Kelso, 1995, p. 12).
Culture and the Sensing and Contextualizing Self
The cognitive capacity of the contextualizing aspect of the self is compatible
with an understanding of culture as symbolic resources with which to strategize
302 ERIKA SUMMERS-EFFLER
(Bourdieu, 1990; Hays, 1994; Sewell, 1992; Swidler, 1989; Tilly, 1993). While
culture is usually thought of as supraindividual, culture and cognition cannot
easily be separated (DiMaggio, 1997). The symbols that comprise culture only
exist through individuals creating, referencing, and modifying them as part of
their action strategies for continued self-expansion and emotional energy seeking
(Schaller, 2001, p. 89). Our shared drive towards interaction requires that symbols
be communicated so that those who routinely interact can negotiate the social
landscape. Beliefs, behaviors, and symbols are communicable to the extent that
they satisfy our interactional goals. No symbols or patterns of meaning are put into
a person directly; culture only gets inside a person as a tool for solving a problem.
For example, children do not use the words they hear most often rst; rather they
use the ones most relevant to the problems that they are trying to solve (Gopnik &
Meltzoff, 1997).
We think and talk with others about the content of our personal history, and in
doing so create not only our individual world-view but also the aggregate level of
culture. Culture is comprised of shared meanings that are signicant because of
their centrality to interaction patterns within a group.
Figure 7 illustrates how emotional-energy-seeking individuals who are
interacting in particular environments call on their personal as if loops for
strategizing social interaction. This process is the force behind cultural diffusion.
The formation of cultural norms on the aggregate level is a mostly unintended
consequence of immediate interactional goals (Schaller, 2001, p. 82).
The contextualizing aspect of the self draws from the symbols available in
our environment, so our options for action are directly impacted by our cultural
surroundings. For example, giving 18-month olds particular objects to play with
will highlight particular information about the childs world. A child who plays
with mixing bowls will gather different evidence about how the world works
than a child who plays with clay, or spears and arrows (Gopnik & Meltzoff,
1997, p. 71). Our methods and our specic goals that represent emotional energy
through existing as if loops are drawn from our experiences within particular
cultural environments. Our motivation to participate in social interaction causes
us to adopt the speech and practices typical of our society; they become part
of our personal as if loops. Thus, the narratives and meanings we use in
order to belong also become our own as we use them as tools for negotiating
our interactional environment (Brothers, 1997, p. 83). By acting to reach our
specic goals, we recycle symbols back into the environment. The symbols are
modied by our use, but as our use of them is in response to what they represented
when we drew on them, the modication will be tied to the larger history of
the symbol.
A Theory of the Self, Emotion, and Culture 303
Fig. 7. The Self and Cultural Diffusion.
While culture can certainly be examined apart from our individual selves and
face-to-face interaction, the ability to understand and predict the diffusion of
symbols into the larger culture would be lost without considering the role of
interactional structure and the internal dynamics of the self. Similarly, we cannot
make sense of the self or face-to-face interactions without accounting for the role
of culture in forming the contextualizing aspect of the self, our personal store
of as if loops. The contextualizing aspect of the self continuously calls on
the larger culture, but its use of culture is contingent on the structural/emotional
consequences of face-to-face interactions. This is how the larger culture is built
upon interactions, but also how the larger culture is a resource for contextualizing.
Through the process of building and modifying our understanding of how the
world works culture out there gets inside the self as it becomes part of our guide
for maximizing emotional energy.
304 ERIKA SUMMERS-EFFLER
CONCLUSION: WHEN AND
WHY THE SELF IS IMPORTANT
When the Self is Important
As long as our brain is healthy, we always have the ability to anticipate the
future and adjust our behavior to reect our predictions; therefore, the self always
exists and has since the evolution of our adaptive capacity and a social motive
(Franks, 1999, p. 167). This does not mean, however, that the self has always
been of central importance or that it will be. The prominence of the self is socially
constructed through the shape of our networks. The interactional structure of our
environment, which is related to our understanding of the role of individuals,
inuences how actors in different times and places understand the their worlds as
more or less responsive to individual imagination, purpose, and effort (Emirbayer
& Mische, 1998, p. 973).
For example, in a society where networks are particularly stable and dense and
individuals closely identify with the group as a whole, the motive to maximize
emotional energy would be focused almost entirely to the collective level rather
than the individual level. As the strategies are on a group level, the group
rather than the individual would be the focus of higher levels of reexivity (for
comparative evidence, see Carrithers et al., 1985; Mauss, 1985). The individual is
of particular importance in the loose and overlapping networks that characterize
late capitalist societies. Rather than approaching the world primarily as a member
of a family or group, individuals are the hubs of their interactional world, and
the individual takes on a prominent role as she or he strategizes to maximize
emotional energy. In these conditions reexivity will be focused on the level
of the individual, and we will have a strong sense of our selves as individuals.
As Goffman points out, the self is a sacred symbol in this society (1959). The
autonomous self is the symbol associated with our primary method of meeting our
emotional energy goals.
Why the Self is Important
Reconciling the self with macro social theory is at the heart of the structure/agency
problem, which is the impasse in being able to explain the power of social structure
as well as the sense of emergence and choice that actors feel. When we see the self
as being propelled forward by a goal that can only be met through favorable social
positioning, we can see how both the structure of face-to-face interactions and the
symbols available limit the choices of the actor. However, we also see that selves
A Theory of the Self, Emotion, and Culture 305
are active strategizers, and as such are far from passive. The vast, but not entirely
unpredictable, possibilities resulting from the interaction between our history,
our current environment, and our innate goals provides us with the potential
for creative problem solving and the sense of emergence and individuality that
we experience.
By referring to the sensing and contextualizing self, we can also resolve some
of the apparent gaps between cultural and structural approaches to understanding
social life. The self is a site for establishing the inuence and limits of both social
structure and culture. Previous efforts have explored the relationship between
culture and structure on a more macro level (Bourdieu, 1990; Giddens, 1984;
Hays, 1994; Sewell, 1992). In moving the level of consideration to the self, culture
is no longer represented in shared rules and strategies external to the individual,
but in the emotional signicance of symbols that ground more macro and shared
manifestations of culture (DiMaggio, 1997). Similarly, structural limits on the
level of the self cannot be approached through organizational dynamics and macro
patterns of access to material resources, but rather the constraints and impact
of face-to-face interaction and its emotional consequences (Collins, 2000). By
considering the process of sensing and contextualizing, we can see how both
structure and culture are integral forces in the self-organizing system of the self,
as well as how they affect each other in the continued effort to establish and
modify strategies and actions for maximizing emotional energy.
Sociology is splintered into many sub areas that pay limited attention to each
other. It is often difcult to see the relevance of macro level work for micro
investigations, as well as the other way around. By connecting social structure
and culture through motivation, strategy, and action, this model suggests hope
for creating a more unied eld where cultural investigations, structural analysis,
and different levels of inquiry could inform each other. In order to achieve viable
micro/macro and structural/cultural connections we need to develop meso level
formulations that detail: how macro structural constraints are realized on the
level networks and face-to-face interaction, the movement of symbols across
face-to-face interaction, and nally the connection between these two processes.
NOTES
1. This is a testable assumption as there are empirical indicators for whether the self (or
in other terms, one might say ones identity) is taking up more social space (size of networks,
position within networks) as well as emotional energy (enthusiasm, condence, pride).
2. This theory has been controversial since rst presented at (James, 1950), in part
because of Jamess critics overly simplistic representation of his ideas (Redding, 1999).
Often accused of dismissing cognition as a relevant factor in emotional life (Cannon,
306 ERIKA SUMMERS-EFFLER
1927; Schachter & Singer, 1962), James was primarily concerned with bridging the
somatic/cognitive divide, and in doing so saw the necessity of including cognition
(Redding, 1999). Although he argues that emotion is the result of physical stimulation, cog-
nition not only follows, but is also used to identify a particular object as some type of object;
thus cognition primes future physical response to ones environment (Redding, 1999).
3. Hateld et al. (1994) do not provide evidence or theories to suggest why the
inclination to mimic works most strongly at the high and low emotional extremes.
4. Scheffs polarity of pride and shame (1991) is similar to Collinss high and low
emotional energy. Being shamed is an experience of having ones emotional energy
abruptly lowered.
5. See Summers-Efer (2002) for a discussion of when this is likely.
REFERENCES
Abelson, R. P. (1981). Psychological status of the script concept. American Psychologist, 36,
715729.
Aron, A., & Aron, E. N. (2000). Self-expansion motivation and including other in the self. In: W. Ickes
& S. Duck (Eds), The Social Psychology of Personal Relationships. New York: Wiley.
Barbalet, J. M. (1998). Emotion, social theory, and social structure: A macrosociological approach.
Cambridge: Cambridge University Press.
Barbalet, J. M. (1999). William James theory of emotions: Filling in the picture. Journal for the
Theory of Social Behaviour, 29, 251266.
Bartky, S. L. (1990). Femininity and domination: Studies in the phenomenology of oppression. New
York: Routledge.
Bourdieu, P. (1990). The logic of practice. Cambridge: Polity Press.
Brothers, L. (1997). Fridays footprint: How society shapes the human mind. New York: Oxford
University Press.
Cannon, W. (1927). The James-Lange theory of emotions. American Journal of Psychology, 39,
115124.
Carrithers, M., Collins, S., & Lukes, S. (1985). The category of the person: Anthropology, philosophy,
history. New York: Cambridge University Press.
Cisek, P. (1999). Beyond the computer metaphor. Journal of Consciousness Studies, 6, 125142.
Clark, A. (1997). Being there: Putting brain, body, and world together again. Cambridge: MIT Press.
Collins, R. (1988). Theoretical sociology. New York: Harcourt Brace Jovanovich.
Collins, R. (1990). Stratication, emotional energy, and the transient emotions. In: T. Kemper (Ed.),
Research Agendas in the Sociology of Emotions (pp. 2757). New York: State University of
New York Press.
Collins, R. (2000). Situational stratication: A micro-macro theory of inequality. Sociological Theory,
18, 1743.
Damasio, A. R. (1994). Descartes error. New York: HarperCollins.
Dewey, J. (1991). How we think. New York: Prometheus Books.
DiMaggio, P. J. (1997). Culture and cognition. Annual Review of Sociology, 23, 263287.
Durkheim, E. (1995 [1912]). The elementary forms of religious life. New York: Free Press.
Ekman, P. (1992). Telling lies: Clues to deceit in the marketplace, politics, and marriage. New York:
Norton.
A Theory of the Self, Emotion, and Culture 307
Emirbayer, M., & Mische, A. (1998). What is agency? American Journal of Sociology, 103,
9621023.
Endler, N. S., & Kocovski, N. L. (2000). Self-regulation and distress in clinical psychology. In:
M. Boekaerts & P. R. Pintrich (Eds), Handbook of Self-regulation (pp. 569599).
Franks, D. D. (1999). Some convergences and divergences between neuroscience and symbolic
interaction. In: D. D. Franks & T. S. Smith (Eds), Mind, Brain, and Society: Toward a Neu-
rosociology of Emotion, Social Perspectives on Emotion (Vol. 5, pp. 157182). Stamford, CT:
JAI Press.
Garnkel, H. (1967). Studies in ethnomethodology. Englewood Cliffs, NJ: Prentice-Hall.
Geen, R. (1991). Social motivation. Annual Review of Psychology, 42, 377399.
Giddens, A. (1984). The constitution of society: Outline of the theory of structuration. Cambridge:
Polity Press.
Giddens, A. (1991). Modernity and self-identity: Self and society in the late modern age. Stanford,
CA: Stanford University Press.
Goffman, E. (1959). The presentation of self in everyday life. New York: Doubleday.
Goffman, E. (1967). Interaction ritual: Essays on face-to-face behavior. Chicago: Aldine.
Golwyn, D. H., & Sevlie, C. P. (1992). Paraphilias, nonparaphilic sexual addictions, and social phobia.
Journal of Clinical Psychiatry, 53, 330.
Gopnik, A., & Meltzoff, A. N. (1997). Word, thoughts, and theories. Boston: MIT Press.
Hardcastle, V. G. (1999). Its O.K. to be complicated: The case of emotion. Journal of Consciousness
Studies, 6, 237249.
Hateld, E., Cacioppo, J. T., & Rapson, R. (1994). Emotional contagion. Cambridge: Cambridge
University Press.
Hays, S. (1994). Structure and agency and the sticky problem of culture. Sociological Theory, 12,
5772.
Higgins, E. T. (2000). Social cognition: Learning about what matters in the social world. European
Journal of Social Psychology, 30, 339.
Hochschild, A. R. (1983). The managed heart: Commercialization of human feeling. Berkeley:
University of California Press.
Homans, G. C. (1961). Social behavior: Its elementary forms. New York: Harcourt, Brace.
Howard, J. A. (2000). Social psychology of identities. Annual Review of Sociology, 26, 367393.
James, W. (1950). The principles of psychology. New York: Dover.
Kelso, S. J. A. (1995). Dynamic patterns: The self-organization of brain and behavior. Boston: MIT
Press.
Kendon, A. (1990). Conducting interaction: Patterns of behavior in focused encounters. Cambridge:
Cambridge University Press.
LeDoux, J. E. (1996). The emotional brain: The mysterious underpinnings of emotional life. New
York: Touchstone.
LeDoux, J. E. (2002). The synaptic self. New York: Viking.
Mead, G. H. (1932). Philosophy of the present. Chicago: Open Court.
Mead, G. H. (1934). Mind, self, and society. Chicago: University of Chicago Press.
Mitra, G., & Mukhopadhyay, A. (2000). Psychological factors in drug addicts and normals: A
comparative study. Journal of Projective Psychology & Mental Health, 7, 5378.
Olson, M. (1965). The logic of collective action. Cambridge: Harvard University Press.
Redding, P. (1999). The logic of affect. Ithaca: Cornell University Press.
Robinson, D. T., &Smith-Lovin, L. (1992). Selective interaction as a strategy for identity maintenance:
An affect control model. Social Psychology Quarterly, 55, 1228.
308 ERIKA SUMMERS-EFFLER
Schachter, S., & Singer, J. E. (1962). Cognitive, social, and physiological determinants of emotional
state. Psychological Review, 69, 379399.
Schaller, M. (2001). Unintended inuence: Social-evolutionary processes in the construction and
change of culturally-shared beliefs. In: J. P. Forgas & K. D. Williams (Eds), Social Inuence:
Direct and Indirect Processes (pp. 7793). Philadelphia: Psychology Press.
Scheff, T. J. (1990). Microsociology: Discourse, emotion, and social structure. Chicago: University
of Chicago Press.
Schneider, D. J. (1991). Social cognition. American Review of Psychology, 42, 527561.
Schore, A. N. (1994). Affect regulation and the origins of the self: The neurobiology of emotional
development. Hillsdale, NJ: Lawrence Erlbaum.
Sewell, W. H., Jr. (1992). Atheory of structure: Duality, agency, and transformation. American Journal
of Sociology, 98, 129.
Simmel, G. (1950). The sociology of Georg Simmel. K. Wolff (Trans). New York: Free Press.
Stevens, L. E., & Fiske, S. T. (1995). Motivation and cognition in social life: A social survival
perspective. Social Cognition, 13, 189214.
Summers-Efer, E. (2002). The micro potential for social change: Emotion, consciousness, and social
movement formation. Sociological Theory, 20, 4160.
Swanson, G. E. (1992). Doing things together: Some basic forms of agency and structure in collective
action and some explanations. Social Psychology Quarterly, 55, 94117.
Swidler, A. (1989). Culture in action: Symbols and strategies. American Sociological Review, 51,
273289.
Tilly, C. (1993). Contentious repertoires in Great Britain, 17581834. Social Science History, 17,
253280.
Tredway, J. V., Knapp, S. J., Tredway, L. C., & Thomas, D. L. (1999). The neurosociological role of
emotions in early socialization, reason, ethics, and morality. In: D. D. Franks & T. S. Smith
(Eds), Mind, Brain, and Society: Toward a Neurosociology of Emotion, Social Perspectives on
Emotion (Vol. 5, pp. 109156). Stamford, CT: JAI Press.
Turner, J. H. (1987). Toward a sociological theory of motivation. American Sociological Review, 52,
1527.
Turner, J. H. (1999). The neurology of emotion: Implications for sociological theories of interper-
sonal behavior. In: D. D. Franks & T. S. Smith (Eds), Mind, Brain, and Society: Toward a
Neurosociology of Emotion (Vol. 5, pp. 81108). Stamford, CT: JAI Press.
Turner, J. H. (2000). On the origins of human emotions: A sociological inquiry in the evolution of
human affect. Stanford: Stanford University Press.
Wiley, N. (1994). The semiotic self. Chicago: University of Chicago Press.
Wrong, D. H. (1961). The oversocialized conception of man in modern sociology. American
Sociological Review, 26, 183193.
DEFENSIVE STRATEGIES:
THE FORMATION AND SOCIAL
IMPLICATIONS OF PATTERNED
SELF-DESTRUCTIVE BEHAVIOR
Erika Summers-Efer
ABSTRACT
When individuals routinely lack access to interactions that build emotional
energy (EE), they use indirect routes to maximize EE. They build strategies
around attempting to minimize the loss of EE. I refer to these indirect routes as
defensive strategies. Defensive strategies reect what psychologists refer to as
an internal locus of control placing control over ones circumstances within
ones self rather than outside in ones environment. While an internal locus
of control may help an individual to adapt to their current situation, it also
helps to preserve the status quo. I focus on the case of staying with an abusive
domestic partner as an illustration of the social dynamics that underlie
apparently self-destructive behavior and the preservation of abusive interac-
tion patterns, including: the formation of defensive strategies, the emotional
and cognitive implications of relying on defensive strategies, the situations
that are likely to lead to the cessation of defensive strategies in favor of
proactive strategies, and the social implications of defensive strategies.
Theory and Research on Human Emotions
Advances in Group Processes, Volume 21, 309325
Copyright 2004 by Elsevier Ltd.
All rights of reproduction in any form reserved
ISSN: 0882-6145/doi:10.1016/S0882-6145(04)21012-8
309
310 ERIKA SUMMERS-EFFLER
INTRODUCTION
If the self is constructed as we develop, and the structure of the self is the
internalization of interaction between people (Mead, 1934; Wiley, 1994), we can
presume that there is a social motivation that underlies the formation, structure,
and ongoing development of the self. Collins suggests that this motivation is
emotional (as it would have to be if it predates the self that is required for cognitive
processing) and that this motivation is to maximize emotional energy (EE) (1990).
EE is similar to the emotion of solidarity that Durkheim (1912), Goffman (1967),
Wiley (1994), and Scheff (1990) have suggested is the fundamental motivation
behind social organization and individual behavior, but by broadening the
motivation beyond solidarity to maximizing EE, Collins allows for the experience
of power in increasing EE and the experience of subordination in draining EE
(1990). We need power as well as solidarity to explain the formation of defensive
strategies and apparently self-destructive behavior.
We gain EE through interaction rituals where we experience solidarity or
power, and lose EE when we are excluded or subordinately positioned. EE feels
like condence and enthusiasm, and a loss of EE feels like depression and lack
of willingness to initiate interaction. In a successful interaction, ritual participants
build up attunement and a common mood with each other through a shared focus
of attention. This process transforms a shared emotion into collective efferves-
cence and EE on the level of the individual (Collins, 1990). In a power dominated
interaction ritual, there is no opportunity for the build up of collective efferves-
cence and EE for all participants because there is no focus on a shared emotion,
rather the focus is on reciprocal emotions, such as anger and fear. Emotions are
built up, but they are not shared and they do not result in collective effervescence;
only the powerful individual experiences an increase in EE. The subordinately
positioned or dominated individual will lose EE. In addition, the power dynamic
inhibits the expression of the emotion that is generated by being positioned
subordinately, fear for example. The power dynamics in interaction that inhibit
the expression of subordinate emotion require those subordinately positioned to
manage emotions internally (Hochschild, 1983). Over time experience teaches
individuals what sorts of interactions hold the most promise and what sorts of
interactions are likely to be draining. Individuals move towards solidarity or power
based on what they have learned will produce the greatest amount of EE (1990).
Collins depicts individuals as active strategizers, using the interaction ritual
options available to them to move from one situation to the next, looking
to maximize the EE payoffs of every interaction. But what happens when our
opportunities for interaction rituals are limited and we have fewalternatives? What
if we experience more failure than success, more EE draining interactions than EE
Defensive Strategies 311
building interactions? What if we have little control over the interactions in which
we participate? I suggest that in these situations, rather than a straightforward
strategy of seeking solidarity or power, we will use an indirect route to maximize
emotional energy. We will build strategies around attempting to minimize the
loss of EE. I refer to these indirect routes as defensive strategies. Situations that
generate defensive strategies can lead to self-destructive behavior.
When individuals learn that their environment is hostile and not amenable to
efforts to bring about change, they turn to the last part of their environment over
which they have control themselves. Rather than targeting their environment,
defensive strategies are based in controlling the individuals own behavior in an
attempt to insure the best possible outcome within a constraining environment.
The result is what psychologists refer to as an internal locus of control, a sense
that ones circumstances are the product of their own efforts rather than external
environmental forces. Psychologists have argued that an internal locus of control is
adaptive (Myers, 1993). A typical example is that students who take the blame for
their bad grade tend to do better than students who blame the test or the instructor.
However, if we are considering domestic violence or other forms of abuse, we can
see that taking the blame for a bad grade is entirely different from taking blame
for ones own routine subordinate positioning. The rst may be adaptive, but the
second surely is not. I suggest that while an internal locus of control may help
an individual to adapt to their current situation, it also helps to preserve the status
quo. In order to assess the psychological and social impact of defensive strategies
we need to consider not only where individuals focus their efforts at controlling
their environment, but the over all effect of the strategy on the indivi