ArticlePDF Available

Emotion, Social Function, and Psychopathology



The studies of emotion function and emotional disorders complement one another. In this article, the authors outline relations between the social functions of emotion and four psychological disorders. The authors first present a social-functional account of emotion and argue that emotions help coordinate social interactions through their informative, evocative, and incentive functions. They then review evidence concerning the emotional and social problems related to depression, schizophrenia, social anxiety, and borderline personality disorder and consider how the emotional disturbances related to these disorders disrupt interactions and relationships, thus contributing further to the maintenance of the disorder. They conclude by discussing research strategies relevant to the study of emotion, social interaction, and psychopathology. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Review of General Psychology
Vol. 2, No. 3, 320-342Copyright 1998 by the Educational Publishing Foundation
Emotion, Social Function, and Psychopathology
Dacher Keltner
University of California, BerkeleyAnn M. Kring
Vanderbilt University
The studies of emotion function and emotional disorders complement one another. In
this article, the authors outline relations between the social functions of emotion and
four psychological disorders. The authors first present a social-functional account of
emotion and argue that emotions help coordinate social interactions through their
informative, evocative, and incentive functions. They then review evidence concerning
the emotional and social problems related to depression, schizophrenia, social anxiety,
and borderline personality disorder and consider how the emotional disturbances
related to these disorders disrupt interactions and relationships, thus contributing
further to the maintenance of the disorder. They conclude by discussing research
strategies relevant to the study of emotion, social interaction, and psychopathology.
We can be afraid... or get angry, or feel pity, in general
have pleasure or pain, both too much and too little, and
in both ways not well; but [having these feelings] at the
right times, about the right things, towards the right
people, for the right end, and in the right way, is the
intermediate and best condition, and proper to vir-
.. . Virtue, then, is a mean.
—Aristotle, Nicomachaen Ethics
The notion that many forms of psychopathol-
ogy include emotional disturbances dates back
to the classical philosophers, as is evident in
Aristotle's definition of virtue as a mean in
emotional response. Recent empirical research
lends credence to this general proposition.
Different psychological disorders have been
linked to "excesses" in emotion (e.g., depres-
sion, anxiety), "deficits" in emotion (e.g.,
depression, psychopathy), or the lack of coher-
ence among emotional response systems (e.g.,
schizophrenia; Barlow, 1988, 1991; Clark &
Watson, 1991; Kring & Bachorowski, in press;
Lazarus, 1991; Plutchik, 1993; Thoits, 1985,
Dacher Keltner, Department of Psychology, University of
California, Berkeley; Ann M. Kring, Department of
Psychology, Vanderbilt University.
During the preparation of this article, Ann M. Kring was
supported in part by a grant from the National Alliance for
Research on Schizophrenia and Depression.
We thank Jo-Anne Bachorowski for her helpful comments
on a draft of this article.
Correspondence concerning this article should be ad-
dressed to either Dacher Keltner, Department of Psychology,
3210 Tolman Hall, University of California, Berkeley,
California 94720, or Ann M. Kring, Department of
Psychology, 301 Wilson Hall, Vanderbilt University, Nash-
Tennessee 37240. Electronic mail may be sent to or
These connections between psychopathology
and emotion provide impetus for at least two
related lines of
A first approach is to rely
on what is known about the linkages between
emotion and autonomic and central nervous
system structures (e.g., Davidson,
1996) as a guide in the discovery of physiologi-
cal mechanisms that contribute to psychopathol-
ogy. A second approach is to look outward at the
individual's social environment and, guided by
what is known about the social functions of
emotion, to document how emotional features of
psychological disorders relate to specific styles
of interaction and relationships, thus perpetuat-
ing the disorder. Our aim in this article is to
provide a conceptualization and research agenda
for this second approach to the study of emotion
and psychopathology.
The benefits of studying the relations between
the social functions of emotion and psychopa-
thology are twofold. First, given the prevalent
association between emotion disturbance and
psychopathology (e.g., Thoits, 1985), basic
research on emotion and social interaction
provides a conceptual framework for consider-
ing possible causes and consequences of emo-
tional disturbances as well as potential interven-
For example, many emotion researchers
assume that emotions involve multiple compo-
including behavior, communication, expe-
rience, and physiology, and find that these
components are often coordinated (e.g., Leven-
son, 1992). Moreover, the coordinated engage-
ment of these components is important for a
number of emotion-based social interactions.
Interpreted within this framework, evidence that
the emotional responses of schizophrenic pa-
tients are not coordinated (e.g., Kring & Neale,
1996) allows us to make more specific predic-
tions about a number of social and interpersonal
consequences. Similarly, other basic emotion
research has found that the experience of anger
is linked to perceptions of unfairness (Keltner,
Ellsworth, & Edwards, 1993) and that the
expression of anger evokes fear in others
(Dimberg & Ohman, 1996) and communicates
dominance and hostility (Knutson, 1996). These
findings illuminate how the anger that figures so
prominently in borderline personality disorder
might be linked to troubled social interactions
and relationships.
Just as important, research on the emotional
features in psychopathology informs the study
of the functions of emotion (Keltner & Gross, in
much as the study of brain dysfunction
informs studies on the functions of brain regions
and systems. Indeed, studying the social conse-
quences of emotional disturbances can elucidate
how emotions contribute to adaptive social
interactions and relationships. For example,
children who display little embarrassment in the
context of making mistakes tend more generally
to behave in antisocial ways (e.g., Keltner,
Moffitt, & Stouthamer-Loeber, 1995). This
finding is consistent with the notion that one
function of embarrassment is to motivate the
adherence to social norms and morals (Miller &
Leary, 1992).
To make connections between emotion, social
functioning, and psychopathology, we first
review theoretical and empirical evidence that
shows how emotions help coordinate interac-
tions related to the formation and maintenance
of social relationships. We then integrate evi-
dence relating four disorders to disturbances in
emotional processing and social interaction and
then discuss how these disturbances preclude
and harm important social relationships. We
conclude by discussing possible research strate-
gies relevant to the study of emotion, social
interaction, and psychopathology.
A Social Functional Account of Emotion
Initial theoretical and empirical interests in
emotion tended to center on the intrapersonal
characteristics of emotion, addressing such
questions as the determinants of emotional
experience, the patterns of emotion-specific
physiology, and the coordination of emotional
responses. Advances in the understanding of the
intrapersonal characteristics of emotions have
facilitated the complementary study of the
interpersonal functions of emotions. Relevant
research and theory has begun to address the
consequences of emotion beyond the individual
and focus on the ways that emotions are
embedded within ongoing social interactions
(e.g., Averill, 1980, 1982; Campos, Campos, &
Barrett, 1989; Ekman, 1992; Lazarus, 1991;
Abu-Lughod, 1990).
This new emphasis on the interpersonal
characteristics of emotion can be summarized in
a social-functional approach to emotion. This
approach conceptualizes emotions as multichan-
nel responses that enable the individual to
respond adaptively to social problems and take
advantage of social opportunities in the context
of ongoing interactions (e.g., Campos et al.,
Ekman, 1992; Frijda & Mesquita, 1994;
Tooby & Cosmides, 1990). Four assumptions
are central to a social-functional approach to
emotion (Keltner & Haidt, 1997). First, it is
assumed that humans are social by nature and
meet many of the problems of survival in social
relationships (Baumeister & Leary, 1995; Fiske,
Humans respond to threats, generate and
distribute resources, and raise offspring in the
context of social relationships.
Second, it is assumed that emotions are
adaptations or solutions to specific problems
related to the formation and maintenance of
social relationships (Averill, 1992; Barrett &
Campos, 1987; Lutz & White, 1986; Tooby &
Cosmides, 1990). As Campos et al. (1989)
proposed, "Emotions are not mere feelings, but
rather are processes of establishing, maintain-
ing, or disrupting relations between the person
and the internal or external environment, when
such relations are significant to the individual"
395, 1989). Certain emotions (e.g., anxiety,
desire, and gratitude) and emotional
dispositions (e.g., positive affectivity) motivate
individual and interactive behaviors that enable
individuals to form social bonds (e.g., Bowlby,
Buss, 1992; Hazan & Shaver, 1987;
Trivers, 1971; Watson, 1988; Watson, Clark,
Mclntyre, & Hamaker, 1992). Other emotions,
such as sympathy, anger, jealousy, amusement,
and embarrassment, are believed to enable
individuals to maintain, protect, and restore
social bonds in the face of immediate threats to
the individual or relationship (Averill, 1982;
Eisenbergetal., 1989;Keltner&Buswell, 1997;
Solomon, 1990).
Third, it is assumed that emotions are
dynamic, relational processes that coordinate
the actions of individuals in ways that guide
their interactions toward more preferred condi-
tions (Campos et al., 1989; Lazarus, 1991).
not only do emotions organize physiologi-
behavioral, experiential, and cognitive
responses within the individual (e.g., Levenson,
they also organize the actions of
individuals in face-to-face interactions (e.g.,
Klinnert, Campos, Sorce, Emde, & Svejda,
1983;Ohman, 1986).
Fourth, a social-functional account presup-
poses that the experience and expression of
emotions bring about beneficial social conse-
quences for individuals and their relationships
(e.g., Barrett & Campos, 1987; Baumeister,
Stillwell, & Heatherton, 1994; Frijda, 1986;
Keltner & Gross, in press). For instance,
embarrassment evokes forgiveness in others and
produces reconciliation following social trans-
gressions (Keltner & Buswell, 1997). Sadness
and distress elicit sympathy, helping, and
increased proximity (Campos et al., 1989;
Eisenberg et al., 1989). Laughter and smiling
evoke affiliative tendencies (Keltner
The experience and expression of emo-
tions are also associated with more cumulative,
long-term social benefits. For example, amuse-
ment has been linked to more satisfying
personal relations (for review, see Keltner &
Bonanno, 1997), and jealousy correlates with
the increased likelihood of maintaining long-
term intimate relations (Buss, 1992).
Emotion and the Coordination
of Social Interaction
A social-functional approach proposes that
emotions coordinate social interactions in ways
that help humans form and maintain beneficial
relationships. Several theorists have offered
different arguments about the role of emotions
in social relationships. For example, certain
theorists have characterized how emotions
follow from systematic changes in social
relationships (de Rivera & Grinkis, 1986). Other
theorists have portrayed emotions as social roles
in which individuals carry out scripted behavior
(Averill, 1980; Clark, 1990). Finally, a number
of theorists have argued that emotions are
elements in extensive social interactions, such
as courtship, flirtation, grieving, or play, that
help humans meet important social goals
(Eibl-Eiblsfeldt, 1989; Ekman, 1984; Lutz,
White, 1990).
Notwithstanding such theorizing, there has
been little integration of the research that has
examined the specific processes by which
emotions coordinate social interactions. In the
ensuing section we summarize research that
reveals three general processes by which
emotions shape social interactions. First, the
expression and experience of emotion signal
socially relevant information to individuals in
interactions about their own and their interaction
partners' emotions, intentions, and orientations
to the relationship. Second, emotions evoke
emotional responses in others that are associated
with beneficial responses to the emotional event.
Third, emotions serve as incentives for others'
actions, thus helping to structure interpersonal
interactions. Disturbances in these emotion-
based processes will contribute to the break-
down of social interactions, which will then
directly impact social and personal adjustment.
Informative Functions of Emotion
At least since Darwin's (1872) analysis of
facial expression, researchers have focused on
the informative functions of emotional expres-
sion, initially studying how facial displays of
emotion communicate information about the
sender's emotion to receivers in a fairly reliable
fashion across cultures (e.g., Ekman, 1993;
Izard, 1977; but see Russell, 1994). More
recently, researchers have documented that
facial, vocal, and verbal expressions of emotion
communicate other kinds of social information
in addition to the sender's current emotion,
including information about the status of
ongoing relations (e.g., Knutson, 1996), individu-
social intentions and relational orientations
(e.g., Fridlund, 1991, 1992), and significant
objects or events in the environment (e.g.,
Mandler, 1975).
Emotional expression conveys information
about senders. Social interactions depend in
part on the knowledge of others' intentions and
emotions. The expression of emotion provides
at least four kinds of information about others'
internal states and dispositions, which in turn
helps coordinate social interactions. First, facial
and vocal expressions of emotion more or less
reliably signal the sender's emotional state to
receivers (Ekman, 1993; Scherer, 1986). Sec-
ond, theorists have speculated that emotional
displays communicate the sender's social inten-
such as, for example, whether to strike or
offer comfort or play (e.g., Fridlund, 1992).
Third, the verbal and nonverbal expression of
emotion signals characteristics of the sender and
receiver's relationship, including the extent to
which it is defined by dominance and affiliation.
For example, displays of anger communicate the
sender's relative dominance and hostility to-
wards the receiver (Knutson, 1996), whereas
displays of embarrassment communicate the
sender's relative submissiveness and inclination
to affiliate (Keltner, 1995). Fourth, in certain
contexts the expression of emotion can signal
information about the sender's mental and
physical health. For example, in one study,
bereaved individuals' displays of anger and
disgust led observers to infer that the individuals
were suffering from poor psychological adjust-
ment and were in need of psychological
assistance (Keltner & Bonanno, 1997). In
addition, infants born in distress cry in a
characteristic way defined by a long latency,
high pitch, and unusually long duration, which
signals the infants' physical problems to observ-
ers (Zeskind & Lester, 1978).
Emotional expression conveys information
about objects and events in the social environ-
ment. Many social interactions revolve around
individuals' coordinated responses to events in
the environment. Humans respond to threats,
distribute resources, and negotiate conflicts in
interpersonal settings. Two lines of research
elegantly illustrate how the expression of
emotion conveys important information about
objects and events in the environment.
One line of research has documented that
fearful behavior facilitates observational learn-
ing of fearful responses in other individuals
(e.g., Mineka, Davidson, Cook, & Keir, 1984;
Mineka & Cook, 1993). In this research,
observer monkeys viewed model monkeys'
fearful, avoidant behavior, including tfieir facial
displays, in response to snakes or toy snakes. In
subsequent test sessions with the snakes, ob-
server monkeys demonstrated that they rapidly
acquired the model monkeys' fear of the real and
toy snakes. The correlations between models*
and observers' fearful behavior typically reached
the high .80s, even after just one observation of
the model monkey.
Studies of social referencing, the process by
which individuals use
emotional displays
to interpret ambiguous stimuli, have also
documented that emotional displays provide
important information about the environment
(e.g., Klinnert et al., 1983; Walden & Ogan,
These now-classic studies have demon-
strated that parents' facial and vocal displays of
positive emotion or fear will determine whether
their infants will walk across a visual cliff
(Sorce & Emde, 1981), play in a novel context,
or respond to a stranger with positive emotion
(reviewed in Klinnert et al., 1983).
Emotional experience provides an assessment
of social relationships. Recently, theorists
have emphasized the relational nature of the
experience of emotion (e.g., Campos et al.,
According to this perspec-
emotional experience provides information
about intraindividual events such as the activity
in the facial musculature or the autonomic
nervous system (e.g., Buck, 1984; Levenson,
Ekman, & Friesen, 1990), and the conditions of
current social relations.
Empirical evidence indicating that individu-
als rely on their emotional experience to assess
relationships along important dimensions is
consistent with this perspective on emotional
experience. For example, the experience of
anger and guilt relates to perceptions of the
fairness of personal relationships (, Walster,
Walster, & Berscheid, 1978; Solomon, 1990).
The experience of embarrassment and shame
relate to perceptions of social status vis-a-vis
others (Gilbert & Trower, 1990; Tangney,
Miller, Flicker, & Barlow, 1996). Theorists have
also speculated that the experience of emotion
provides an assessment of the level of commit-
ment to a relationship, in the case of love and
sympathy (Frank, 1988), and the extent to which
ongoing social relations are propitious for the
individual's reproductive success, in the case of
happy and sad mood states (Nesse, 1990).
These observations indicate that individuals
rely on their experience of emotion to assess
their relationships with individuals who are both
the cause and target of the emotion. Emotional
experience also provides information about the
condition of social relations in general. Specifi-
cally, individuals experiencing moods and emo-
tions elicited in one context will use those
feelings to evaluate their relationships, even
though the original cause of the emotion is
unrelated to the relationship (e.g., Clore, 1994;
Schwarz, 1990). For example, empirical studies
have documented evidence that current mood or
emotion influences individuals' evaluations of
their general relationship satisfaction (e.g.,
Keltner, Locke, & Audrain, 1993), anger influ-
ences fairness judgments of ongoing interac-
tions (Keltner et al., 1993), and fear influences
the perception of the possible risks and losses
associated with potential social interactions
(Lerner& Keltner, 1997).
To summarize, theory and research indicate
that emotional expression and experience pro-
vide important information about the sender's
emotions, intentions, orientation to the relation-
and well-being; events or objects in the
environment; and the conditions of social
relations. Because emotions provide such valu-
able social information, theorists have argued
that the communication of emotion is an
important component of more complex dis-
course processes in which individuals negotiate
interpersonal conflicts (e.g., Dunn & Munn,
1985) and reach a shared understanding about
concepts of right and wrong (e.g., Bretherton,
Fritz, Zahn-Waxler, & Ridgeway, 1986; White,
Disturbances in emotional experience
and expression, therefore, are likely to disrupt
relationships in important
Disturbances in
the intensity, type, and timing of emotional
expression, for example, would deprive interac-
tion partners of valuable information about
ongoing interactions. Disturbances in emotional
experience would likewise compromise the
nature of information about his or her current
Evocative Functions of Emotions
Emotion theorists have long suggested that
humans evolved adaptive responses to the
emotional responses of others (e.g., Darwin,
Ohman & Dimberg, 1978). This claim is
consistent with the general assumption that the
communicative behavior of sender and receiver
co-evolved in reciprocal fashion (Eibl-Eibes-
feldt, 1989; Hauser, 1996; Owren & Rendall, in
From this perspective, one individual's
emotional expression serves as a social affor-
dance that evokes "prepared" responses in
others (e.g., Ohman & Dimberg, 1978). Studies
of the evocative properties of emotional expres-
sions suggest that emotions can evoke both
complementary emotions and similar emotions
in others.
Evocation of complementary
pirical studies have documented that emotional
displays evoke complementary emotions in
others, which we define as emotional responses
that differ from that of the sender but that
respond selectively to the sender's emotional
display. These complementary emotions moti-
vate important social behaviors, including help-
ing, soothing, and forgiveness. At least three
examples of the evocation of complementary
emotional responses can be found in the
empirical literature.
First, in a series of innovative conditioning
studies, Ohman and Dimberg have documented
how displays of anger evoke complementary
fear in observers (reviewed in Dimberg &
Ohman, 1996). For example, photographs of
angry facial displays paired with an aversive,
unconditioned stimulus were more resistant to
extinction than photos of facial displays of
happiness (Ohman & Dimberg, 1978). Strik-
ingly, one study found that this conditioning
effect was only observed when angry faces were
directed towards the subject (Dimberg &
Ohman, 1983). In subsequent research, anger
faces that were "masked" by a neutral face
presented immediately following the presenta-
tion of the anger faces, and presumably not
consciously represented by the observer, still
produced the conditioning effects described
above and evoked elevated electrodermal re-
sponses associated with fear (Esteves, Dimberg,
& Ohman, 1994).
Research on how distress displays elicit
sympathy in others illustrates a second comple-
mentary emotion
Developmental stud-
ies find that children respond with signs of
distress, concern, and overt attempts to help in
response to others' distress beginning at as early
as 8 months of age (Zahn-Waxler & Radke-
Yarrow, 1982). Studies of adults find that
individuals feel sympathy and concern in
proportion to others' observable signs of emo-
tional distress (Batson & Shaw, 1991). Further-
more, there appears to be a pattern of sympathy-
related expressive and physiological responses
that includes concerned gaze and oblique
eyebrows and reduced heart rate that predicts
helping behavior (Eisenberg et al., 1989).
A third complementary emotional response is
found in studies of the evocative effects of
embarrassment. These studies show that observ-
ers report high levels of affiliative emotions,
such as amusement and sympathy, in response to
others' displays of embarrassment and shame
following social transgressions (Keltner, Young,
& Buswell, 1997). In addition, a number of
studies have found that when an individual
displays embarrassment-related behavior follow-
ing a mistake, observers report greater liking of
the individual and, when relevant, more forgive-
ness (reviewed in Keltner & Buswell, 1997).
Evocation of similar emotions. Theorists
have long been interested in the tendency for
humans to respond to others' emotions with
similar emotions (for historical review, see
Hatfield, Rapson, & Cacioppo, 1994). Indeed,
there is a good deal of research on the related
phenomena of emotional mimicry (e.g., Davis,
empathy (e.g., Hoffman, 1984), and
vicarious emotional response (e.g., Miller,
Although researchers need to more
clearly specify the conditions under which
individuals reciprocate each other's emotions,
as well as the emotions that tend to be
reciprocated, empirical evidence indicates that
one individual's experience and display of
emotion evokes comparable emotional re-
sponses in others. Specifically, studies have
documented such reciprocal responses in the
case of embarrassment (Miller, 1987), laughter
(Provine, 1992), and distress or sadness (Batson
& Shaw,
Eisenberg et al., 1989).
Theorists have offered various speculations
about the social benefits of the reciprocation of
emotional response. First, the elicitation of
similar emotion in others increases the likeli-
hood that individuals will know one another's
emotional states (e.g., Hoffman, 1984). The
knowledge of others' emotional states in combi-
nation with empathic emotion, some theorists
allege, is the foundation of moral emotions
such as guilt (Hoffman, 1984) and moral
behavior such as altruistic helping (Batson &
Shaw, 1991). Second, just as alarm calls in
nonhuman species coordinate the responses
of conspecifics to a shared threat (e.g., Seyfarth
& Cheney, 1990), reciprocated emotions are
likely to coordinate the actions of several
individuals facing a similar object or event in
the environment.
In sum, a substantial body of evidence
indicates that emotional expression evokes
complementary and similar emotional responses
in others. Individuals predisposed to express
certain emotions, therefore, are likely to evoke
specific emotions in others with whom they
interact. In certain instances, these evocative
processes are beneficial for social relationships.
In other instances, however, the evocation of
emotion in others may prove to be problematic.
Incentive Functions of Emotions
Emotions coordinate social interactions in a
third way: an individual's expression and
experience of emotion may provide incentives
for or reinforce another individual's social
behavior within ongoing interactions (e.g.,
Klinnert et al., 1983). There have been few
empirical studies that have directly examined
the incentive functions of emotional expression.
Relevant studies and related theoretical observa-
however, suggest that individuals fre-
quently engage in social behaviors contingent
on or in anticipation of others' emotional
experience and expression.
Developmental researchers have speculated
that the display of positive emotion by both
parents and children rewards desired behaviors,
thus increasing the frequency of that behavior
(e.g., Tronick, 1989). For instance, as infants
carry out intentional behaviors with the assis-
tance of their parents, such as grabbing an
object, they will smile when their parents
engage in behavior that facilitates their own
goal-directed behavior and show signs of
distress when the parents do not act in such
fashion (Tronick, 1989). Other studies have
shown that parents use positive emotional
displays to direct the attention of their infants
(Cohn & Tronick, 1987). More generally, it has
been claimed that parental laughter facilitates
learning by rewarding appropriate behavior in
infants and children (Rothbart, 1973).
Studies of adult laughter reach similar conclu-
sions about the incentive function of emotional
behavior. For example, researchers have found
that the temporal location of laughs within the
ongoing stream of conversations is almost
exclusively at the end of the utterance (Provine,
Although there are several interpretations
of this finding (see Bachorowski, Smoski, &
Owren, 1998), one plausible interpretation is
that laughter serves as a reward for social
behavior preceding it, offering praise for desir-
able utterances in the context of conversations.
More generally, theorists have argued that
laughter rewards many forms of desirable social
behavior in the context of ongoing interactions
(Bachorowski etal., 1998).
Several emotion-centered social behaviors
seem to be motivated to elicit emotion in others.
For example, one motive of teasing interactions
is to embarrass the target of teasing (e.g.,
Keltner, Young, Oemig, Heerey, & Monarch,
Although the embarrassment and humili-
ation produced by teasing can have many
negative consequences (for review, see Keltner
et al., 1997), teasing also can have positive
outcomes. Specifically, teasing related to embar-
rassment often increases affiliation and can
provide important information about
for example, whether individuals are
romantically interested in one another.
Finally, to the extent that emotional displays
provide incentives that guide social behavior,
the reduction or absence of emotional displays
should reduce the likelihood of contingent social
behavior, A few studies provide indirect evi-
dence in support of this claim. For example,
experiments in which parents are instructed to
mute their expressive behavior find that infants
quickly become disturbed and disengaged (Tron-
Adamson, Wise, & Brazelton, 1978). In
conversations with individuals who are prone to
low levels of positive emotion, participants
engage in less responsive social behavior and
experience the conversations as unrewarding
(Thome, 1987).
Summary of the Social-Functional
We have argued that emotions coordinate
social interactions by serving at least three
functions. Emotions provide information about
interacting individuals' emotions, intentions,
and relational orientations. Emotions evoke
complementary and similar emotions in others
that motivate behaviors that benefit social
relationships. The perception of emotion and
anticipated elicitation of emotions in others
serve as incentives for certain social behaviors.
In these three ways, emotions provide structure
to social interactions, guiding, evoking, and
motivating the actions of individuals in interac-
tions in ways that enable individuals to meet
their respective goals. Disturbances in emo-
tional response, by implication, will have
important consequences for the quality of social
interactions and relationships.
Emotion and Social Interaction
Disturbances in Psychopathology
Although different kinds of emotional distur-
bances figure prominently in the descriptions
and manifestations of various psychological
disorders, there have been relatively few empiri-
cal studies on the nature of these disturbances.
By contrast, difficulties in social interactions,
variously referred to as social competence,
social skills, social support, and social adjust-
ment, have been studied more extensively across
a wide range of disorders. Although theorists
have recently recognized the importance of
linking the literatures on emotional and social
disturbances in psychopathology (Blanchard &
Panzarella, in press; Buck, 1991; Feldman,
Philippot, & Custrini, 1991), these literatures
remain largely unintegrated. By emphasizing
the social functions of emotions, including the
ways in which emotions coordinate social
interactions, we believe that researchers will be
able to more clearly make both conceptual and
empirical connections between the nature of emo-
tional and social disturbances in psychopathology.
In the following section, we first review the
evidence for emotional disturbances and social
interaction problems in four psychological
disorders. We then provide a framework for
integrating these two previously disparate litera-
tures by considering the emotional disturbances
in the context of a social-functional approach.
That is, we link the emotion disturbances to
informative, evocative, and incentive functions
of emotion and then point to specific expecta-
tions and hypotheses and, when possible,
relevant findings about the ways in which these
emotion disturbances interfere with social func-
tioning. Although many of the studies reviewed
in the first half of this article considered how
discrete emotions coordinate social interactions,
much of
research on emotion disturbances in
psychopathology has focused on broad dimen-
sions of emotion, primarily due to the pervasive
nature of the emotion disturbances. Nonethe-
links between emotion dimensions and
social interactions can also be made. For
example, positive affect is associated with
ratings of the frequency and duration of a
number of social interactions and activities
(Watson, 1988; Watson etal., 1992).
For several reasons, we chose to focus on four
disorders: unipolar depression, schizophrenia,
borderline personality disorder, and social pho-
bia. First, emotional symptoms and interper-
sonal difficulties, including the formation and
maintenance of relationships, figure promi-
nently in each of these disorders. Second, we
chose to focus on diagnosable disorders rather
than broader psychological dimensions (e.g.,
distress, well-being) in hopes of providing an
impetus for research on emotion and social
deficits in disorders. Third, these disorders serve
as exemplars of how disturbances in emotion
influence maladaptive social interaction pat-
Finally, space constraints limit our ability
to discuss other disorders, although we recog-
nize that many other disorders also are character-
ized by emotional and social disturbances.
Unipolar Depression
Emotional disturbances. The prominent
emotional features of unipolar depression in-
clude the phasic and enduring experience of
sadness and the inability to experience pleasure
(anhedonia). More broadly, depression is charac-
terized by low levels of positive affect and
heightened levels of negative affect (e.g.,
Watson, Clark, & Carey, 1988). People with low
levels of positive affect are likely to experience
emotions such as sadness and to be interperson-
ally disengaged. In contrast, people with high
levels of negative affect are likely to frequently
experience emotions such as anxiety, guilt, and
One line of theorizing relevant to our interests
holds that depression reflects deficits in a
reward-oriented approach motivation system
(Depue & Iacono, 1988; Depue, Krauss, &
Spoont, 1987; Tomarken & Keener, in press).
There is empirical support for this reasoning in
studies linking relative hypoactivation in the left
frontal hemisphere, which is thought to relate to
approach-related emotion and motivation sys-
to depression (e.g., Allen, Iacono, Depue,
& Arbisi, 1993; Davidson, SchafTer, & Saron,
Henriques & Davidson, 1991), risk for
depression (Tomarken, Garber, & Simien, 1997),
and low levels of positive affect (Tomarken,
Davidson, Wheeler, & Doss, 1992). Tomarken
and Keener (in press) have proposed that this
pattern of brain activity may be a marker of risk
for depression that is reflected by a number of
deficits, including the relative incapacity to
respond to positive emotional stimuli and
self-regulatory deficits in the capacity to use
positive events to shift into positive emotional
Either of these deficits will likely
interfere with social interactions insofar as
social interaction requires goal-directed or
reward-oriented approach behavior. Thus de-
pressed people may not derive pleasure or
reward from interpersonal relationships or
interactions while currently depressed, perhaps
due primarily to a diminution in positive
affectivity. Moreover, depressed people may not
seek out social interactions if they fail to help
shift their mood from a predominantly negative
state into a more positive one.
Although most studies of depression and
emotional disturbance have concentrated on
emotional experience, some evidence indicates
that depressed patients may also exhibit flat,
dull, and slowed speech (Buck, 1984; Har-
greaves, Starkweather, & Blacker, 1965; Levin,
Hall, Knight, & Alpert, 1985; Murray & Amott,
Scherer, 1986) and limited facial expres-
particularly expressions of positive emo-
as well as a decrease in overall body
movement (Berenbaum & Oltmanns, 1992;
Ekman & Friesen, 1974; Gotlib & Robinson,
Jones & Pansa, 1979; Ulrich & Harms,
Waxer, 1974). For example, one study
found that depressed people showed fewer facial
expressions in response to positive stimuli (but
not to negative stimuli) than schizophrenic
patients without flat affect and nonpatient
controls (Berenbaum & Oltmanns, 1992).
Social disturbances. Empirical studies con-
sistently find that depression is marked by
disturbed relationships and social interactions
(for reviews, see Barnett
Gotlib, 1988; Gotlib,
Hokanson & Rubert, 1991). Specifically,
both dysphoric and clinically depressed individu-
als have been found to have fewer social skills
(e.g., Youngren & Lewinsohn, 1980), fewer
close relationships (e.g., Billings & Moos, 1985;
Brown & Harris, 1978; Gotlib & Lee, 1989),
less elaborated social networks (Gotlib, 1992),
less rewarding relationships (Hokanson, Loewen-
stein, Heden, & Howes, 1986; Joiner, 1996;
Joiner, Alfano, & Metalsky, 1992; Joiner &
Metalsky, 1995; Nezlak, Imbrie, & Shean,
fewer social contacts (Gotlib & Lee,
but see Nezlak et al., 1994), less social
support (e.g., Joiner, 1997; Lara, Leader, &
Klein, 1997), and more marital problems and
family arguments (e.g., Brown & Harris, 1978;
Gotlib & Hooley, 1988; Gotlib & Lee, 1989;
Monroe, Bromet, Connell, & Steiner, 1986; see
Beach, Smith, & Fincham, 1994, for a review).
Moreover, a number of these interpersonal
deficits remain stable across periods of depres-
sion and remission (Barnett & Gotlib, 1988;
Gotlib & Lee, 1989) and are predictive of future
symptomatology and course (Billings & Moos,
George, Blazer, Hughes, & Fowler, 1989;
Joiner, 1997; Joiner et al., 1992; Joiner &
Metalsky, 1995), even when controlling for
initial levels of depressed mood, neuroticism,
number of previous depressive episodes, and a
prior history of dysthymia (Lara et al., 1997).
Linking emotional and social disturbances.
The above findings suggest that depression is
marked by low levels of positive affect, high
levels of negative affect, diminished facial and
vocal expressivity, and troubled interpersonal
relationships. Our social-functional approach
suggests a number of lines of inquiry and
predictions concerning the relations between
emotional and social disturbances.
Although not yet fully tested, heightened
negative affect among depressed individuals
likely communicates information about the state
of interpersonal relationships that may further
damage those relationships. Several studies are
consistent with findings linking negative moods
and emotions to assessments of reduced relation-
ship satisfaction (e.g., Keltner et al., 1993). For
example, depressed individuals have been found
to be more pessimistic in expectations about
their current and future social relationships than
nondepressed individuals (e.g., Hokanson &
Rubert, 1991) and to perceive family relation-
ships as less supportive (Billings & Moos,
Studies have also shown college students
(particularly male participants) with depressive
symptoms who repeatedly seek reassurance
from their roommates to be more likely to be
rejected by their roommates (e.g., Joiner, 1996;
Joiner et al., 1992). In addition, to the extent that
positive affect reflects engagement with the
environment (e.g., Watson, 1988), we would
expect lowered levels of positive affect among
depressed people to be associated with fewer
initiations of social interaction. Although this
remains an empirical question, it can be
prospectively studied by assessing changes in
positive (and negative) affect before, during, and
after a depressive episode and then measuring
the extent to which these changes are linked to
increases or decreases in social interaction and
the extent to which these individuals derive
pleasure from interpersonal interactions and
relationships. An alternative approach would be
to predict changes in positive and negative affect
from changes (i.e., quantity, quality) in interper-
sonal relationships.
Studies that have documented a robust
relationship between marital dissatisfaction and
depression (Beach et al., 1994) provide addi-
tional evidence related to the informative and
evocative functions of emotion. Specifically,
vocally expressed negative affect (e.g., Smith,
Vivian, & O'Leary, 1990), self-reports of
negative affect (e.g., Gottman & Levenson,
Levenson, Carstenson, & Gottman, 1994),
and vocal expression of specific negative
emotions, such as contempt, anger, and fear
(e.g., Gottman &
1989), during marital
problem solving interactions have been linked to
decreases in marital satisfaction. Beach and
Fincham (1994) have hypothesized that individu-
als who are higher in negative affectivity may be
more likely to evidence negative communica-
tion patterns in marital interactions and have
greater marital dissatisfaction. Based on these
findings, we would predict that negative affect
may be an important mediator of the link
between depression and marital discord. How-
ever, it is likely that increases in positive affect
may also be necessary to improve marital
communication in couples where one partner is
depressed. Thus interventions aimed at chang-
ing emotion and mood may indirectly affect
close relationships. By contrast, interventions
aimed at improving marital relationships may
also be beneficial for treating depression. In fact,
a number of studies have shown that behavior
marital therapy is not only effective for treating
troubled marriages, but it is also effective at
reducing depressive symptomatology in couples
where one partner is depressed (see Beach et al.,
for a review).
Other recent evidence linked to the evocative
functions of emotion indicates that depression
elicits a number of predominantly negative
emotions in others, including depression (e.g.,
Joiner, 1994; Joiner et al., 1992). For example, a
recent study found that negative social interac-
tions between caregivers and depressed individu-
als predicted caregivers' reports of depression
and distress. Moreover, positive interactions did
not buffer the relationship between caregiver
demand and distress (Rauktis, Koeske, &
Tereshko, 1995). Several studies have found that
brief interactions with a depressed person elicit
feelings of distress, anxiety, depression, and
anger, even among strangers (e.g., Coyne, 1976;
Marks & Hammen, 1982). However, consistent
with the evidence concerning distress and
sympathy, individuals instructed to be "helpers"
to depressed people displayed support and
expressions of concern (Marks & Hammen,
Sacco, Milana, & Dunn, 1985). Continued
exposure to a depressed person, however, is
likely to induce more negative than positive
emotions, particularly in the context of close
relationships (Coyne, 1976; Joiner, 1994; Joiner
etal., 1992).
Finally, the tendency for depressed people to
display few positive expressions and to be
unexpressive in general suggests that they will
not provide positive incentive cues for others'
social behavior and, more generally, will fail to
provide important signals about emotional state,
intentions, and objects or events in the environ-
ment to interaction partners. Evidence from the
developmental literature suggests that depressed
caregivers' limited or inappropriate facial and
vocal emotional displays may have direct effects
on infants' learning, behavior, and emotional
regulation. For example, studies examining
interactions between depressed mothers and
their infants have shown that mothers often
exhibit flat or negative facial and vocal expres-
avoid eye contact, and are less likely to be
playful and attentive (e.g., Cohn & Cambell,
Field, 1995). In addition, infants of
depressed mothers have been found to be less
playful, be less active, display less positive
affect, and display more expressions of sadness
and anger (e.g., Field, 1995; Field, Healy,
Goldstein, & Guthertz, 1988; Pelaez-Nogueras
et al., 1994; Pickens & Field, 1993). Infant-
directed speech is believed to play a particularly
important role in promoting the development of
emotion regulation and attentional skills among
infants, and a recent study has shown that the
vocalizations of mothers with depressive symp-
toms failed to promote associative learning in
4-month-old infants (Kaplan, Bachorowski, &
Zarlengo-Strouse, in press).
Emotional disturbances. Two of the more
prominent emotional features of schizophrenia
include flat affect (a lack of outward expression
of emotion) and anhedonia (the inability to
experience pleasure in situations that normally
evoke pleasure). Experimental investigations
using emotionally evocative stimuli have found
that schizophrenic patients are less facially
expressive than nonpatients in response to
emotional films (Berenbaum & Oltmanns, 1992;
Kring, Kerr, Smith, & Neale, 1993; Kring &
Neale, 1996; Mattes, Schneider, Heimann, &
Birbaumer, 1995) and cartoons (Dworkin, Clark,
Amador, & Gorman, 1996) and during social
interactions (Krause, Steimer, Sanger-Alt, &
Wagner, 1989; Mattes et al., 1995), but report
experiencing the same or greater amount of
emotion and exhibit the same or greater amount
of skin conductance reactivity as nonpatients
(Kring, Germans, & Earnst, 1997; Kring &
Earnst, 1998; Kring & Neale, 1996).
Blanchard and colleagues have studied the
linkage among anhedonia, reports of experi-
enced emotion in response to emotional stimuli,
and reports of positive and negative affect
among schizophrenic patients. These studies
have documented the relationship of anhedonia
to lower levels of state-positive affect in
response to emotional films (Blanchard, Bel-
lack, & Mueser, 1994), low levels of trait-
positive affect and high levels of trait-negative
affect, and poorer social adjustment, which was
defined by interpersonal contacts, leisure, and
romantic involvement (Blanchard, Mueser, &
Bellack, in press).
Taken together, these findings suggest that