EMOTION REGULATION:

A THEME IN SEARCH OF DEFINITION

Ross A. Thompson

Interest in emotion regulation has burgeoned in recent years because

it builds on several recent trends in the study of emotional development
(Thompson, 1993). First, after more than a decade of research emphasizing
the growth of discrete emotions and their consequences for sociopersonality
functioning, researchers have realized that the specific emotion indexes only
part of the rich individual variability that exists in emotional behavior. In
addition, individuals display variations in the intensity, persistence, modula-
tion, onset and rise time, range, and lability of and recovery from emotional
responses. These "emotion dynamics" (Thompson, 1990) constitute signifi-
cant response parameters that are influenced by emotion regulation pro-
cesses. In a sense, while the discrete emotion may "play the tune" of a
person's emotional response, these emotion regulation processes signifi-
cantly influence its quality, intensity, timing, and dynamic features and thus
significantly color emotion experience.
Second, emergent views of emotion underscore its biologically adaptive
and psychologically constructive features in contrast with earlier portrayals
of emotion, which emphasized its disorganizing, irrational, or stressful side
(e.g., Barrett & Campos, 1987; Malatesta, 1990). Emotional arousal has, of
course, the capacity to either enhance or undermine effective functioning,
and emotion regulation processes are important as they enlist emotion to
support adaptive, organized behavioral strategies. Emotion regulation is
relevant, for example, to effective social strategies with peers (e.g., Rubin
& Rose-Krasnor, 1986), successful cognitive performance in tasks involving

1 am grateful to Alice Ganzel, Megan Gunnar, and Kathy Stansbury for helpful,
critical readings of an earlier draft of this essay and to Pamela Cole for stimulating ex-
changes about these issues.

ts
NATHAN A. FOX, ED.

delay, inhibition, or the pursuit of long-term goals (e.g., Mischel & Mischel,
1983), and the management of stressful experiences at home (Cummings,
Pellegrini, Notarius, & Cummings, 1989). In a sense, students of emotional
development have moved beyond the realization that discrete emotions are
biologically adaptive to the awareness that emotional responses must also
be flexible (rather than stereotypical), situationally responsive (rather than
rigid), and performance enhancing (rather than over- or underarousing)
and must change quickly and effectively in order to adapt to chang-
ing conditions if they are to support organized, constructive functioning
in higher organisms. This is where emotion regulation processes often
enter in.
Third, newer portrayals of emotional development also emphasize the
socialization of emotion as a significant constituent of emotional develop-
ment. Through processes ranging from selective reinforcement and model-
ing of expressions of emotion to emotion-focused discourse, the social con-
text not only fosters greater "emotional competence" (Saarni, 1990) in
developing individuals but also channels emotional behavior in directions
that meet the expectations of the "emotion culture" (Gordon, 1989) in which
those individuals develop. As a consequence, emotion experience derives
from an interaction between biologically based emotive processes and the
socialized monitoring, evaluative, and regulatory processes by which emo-
tion experience is interpreted and managed in culture-specific ways. More-
over, as developing individuals become more skilled at regulating arousal,
emotion and its expression can become better integrated into the child's
growing repertoire of strategic behavior in social contexts. Emotion regula-
tion is central both to the socialization process and to its developmental
outcomes.
Finally, emotions theory is currently also concerned with individual
differences in personality and social functioning and the central role that
emotive processes play in these differences. Whether the focus is on individ-
ual differences in infant-mother attachment and the emotional biases that
they reflect (Malatesta, 1990; Thompson, 1991; see also Cassidy, in this
volume), variations in behavioral inhibition and their origins in the self-
regulation of emotion (see Fox, in this volume; Kagan, in this volume), or
the interpretive processes underlying the emotional reactions of aggressive
children to their peers (Dodge, 1991a), researchers are exploring aspects
of emotion regulation and dysregulation that guide social and personality
processes. The common theme underlying these studies is how emotional
arousal comes to mean different things to different individuals (e.g., why
anger is empowering to some people, disorganizing for others, and to be
denied or avoided for others), and emotion regulation is a significant com-
ponent of these individual differences.
 THE DEVELOPMENT OF EMOTION REGULATION

These current trends in theory and research on emotional development

have provided an auspicious beginning to the study of emotion regulation.
But, as so often happens with auspicious beginnings, recent enthusiasm
for the study of emotion regulation has outpaced attention to some basic
definitional and conceptual issues. This is because many researchers share
a common intuitive understanding of what is meant by emotion regulation or
of the distinguishing features of the "optimal" self-management of emotion,
but these implicit formulations have tended to obscure the heterogeneity
of emotion regulation processes, the complexity of their development, their
links to significant social relationships and to the "emotion culture" that we
share, and the challenges of identifying the origins and correlates of these
regulatory processes. My purpose in this essay is to highlight these complexi-
ties by (re)considering the following questions: How should we define emo-
tion regulation? What is regulated in the regulation of emotion? How do
emotion regulation strategies fit into the complex fabric of social interac-
tion? What are the predictable correlates of individual differences in emo-
tion regulation? In a sense, my goal is to "unpack" the concept of emotion
regulation to better elucidate its component processes.

DEFINING EMOTION REGULATION

The power of our shared, implicit notions of what constitutes emotion

regulation is reflected, perhaps, in how frequendy papers on emotion regu-
lation (including my own) lack a clear definition of this phenomenon. Yet
behind this apparent consensual understanding is considerable diversity in
the underlying portrayals of emotion regulation provided by different re-
Searchers and theorists. Does emotion regulation pertain exclusively to the
inhibition of emotional reactions, for example, or does it also include the
maintenance or enhancement of emotional behavior? Is emotion regulation
primarily an issue of emotion 5«//-management, or is the management of
emotional reactions by others also included? Does emotion regulation pri-
marily influence the discrete emotion that one experiences, or rather its
quality (e.g., its intensity, speed of onset, or persistence)? Is emotion regula-
tion primarily concerned with the management of expressions of emotion
or the underlying arousal processes leading to those expressions—or both?
There is surprising diversity in the ways that different researchers an-
swer these questions in their implicit formulations of emotion regulation.
^y own answers are indicated in the following definition:

Emotion regulation consists of the extrinsic and intrinsic processes re-

sponsible for monitoring, evaluating, and modifying emotional reac-
NATHAN A. FOX, ED.

tions, especially their intensive and temporal features, to accomplish

one's goals.

Included in this definition are several characterizations of emotion regula-

tion processes.
First, consistent with Masters (1991), emotion regulation can involve
maintaining and enhancing emotional arousal as well as inhibiting or subdu-
ing it. It is natural that theorists emphasize emotion inhibition in a culture
that, like ours, values this characteristic: in everyday circumstances, emotion
regulation skills are most often enlisted to dampen emotional arousal (espe-
cially negative emotion). But, even in our culture, strategies of emotion
management are often used to maintain or enhance emotional arousal, such
as when children intensify their anger to stand up to a bully who is also
feared (see Miller & Sperry, 1987), or when they "feel sorry for themselves"
when unjustly treated, or when adults ruminate on feelings of guilt, anger,
or shame in response to social injustice. And children and adults frequently
enlist strategies to heighten positive arousal (e.g., by reenacting pleasant or
humorous experiences), sometimes to manage negative affect. In these and
other instances, the enhancement of emotional arousal (not just its display)
serves important strategic purposes, and this becomes the goal of the emo-
tion regulation process.
Second, emotion regulation encompasses not only acquired strategies
of emotion 5e//-management but also the variety of external influences by
means of which emotion is regulated. This is because a considerable amount
of emotion regulation occurs through the interventions of others. In in-
fancy, for example, caregivers devote considerable effort to monitoring,
interpreting, and modulating the arousal states of young offspring—in
other words, regulating their emotions. As offspring mature, parents use
direct interventions as well as indirect strategies (e.g., coaching response
alternatives) not only to maintain emotional well-being in their children but
also to socialize emotional behavior so that it accords with cultural expecta-
tions concerning feelings and their expression (Saarni, 1990). Moreover,
parent-offspring relationships and other significant social ties affect the de-
mands for emotion regulation and the efficacy of strategies for managing
arousal that children acquire in the context of these close relationships.
And, as adults, we frequently manage the emotions of others by extending
a sympathetic ear or using humor in a frustrating situation. Taken together,
therefore, the development of the skills required to manage one's own emo-
tion occurs in a social context (both proximate and cultural) that significantly
shapes children's management of arousal through external regulatory in-
fluences.
Third, although emotion regulation sometimes affects the discrete

28
 THE DEVELOPMENT OF EMOTION REGULATION

emotion experienced by an individual (e.g., the arousal of guilt or shame

rather than anger when unfairly accused), more commonly it affects the
intensive and temporal features of that emotion. In other words, aspects of
emotion management subdue (or enhance) the intensity of experienced
emotion, retard (or speed) its onset or recovery, limit (or enhance) its persis-
tence over time, reduce (or increase) emotion range or lability, and affect
other qualitative features of emotional responding. Because of this, new
strategies for the study of emotion regulation are required that are sensitive
to these intensive and temporal features of emotional responding, even when
the discrete emotion itself is unaffected. Several new methodological ap-
proaches meant to accomplish this are outlined at the conclusion of this essay.
Finally, emotion regulation must be regarded functionally, that is, in
terms of the regulator's goals for a particular situation. These goals may be
diverse and changing, and, as 1 argue later, they include far more than
simply maintaining a positive disposition in oneself or another. Indeed, an
understanding of individual differences as well as developmental changes in
emotion regulation may hinge on an appreciation of the goals for managing
emotion that are motivating regulatory efforts, and this makes goal attain-
ment a central definitional feature of emotion regulation. To be sure, histor-
ically there have been important problems with functionalist approaches
like this one to psychological processes (circular reasoning among them:
theorists can easily infer goals that are consistent with their behavioral analy-
sis), but these problems should instill caution rather than avoidance of such
an approach to the study of emotion regulation. Indeed, contemporary
interest in emotion regulation is consistent with emergent functionalist ap-
proaches to emotional development that are currently enlivening emotions
research (e.g., Barrett & Campos, 1987).
This is an inclusive definition o£ emotion regulation, and among the many
issues not addressed here are distinguishing (if necessary) emotion regula-
tion from related processes like defense mechanisms and display rules, artic-
ulating the relations between the self-regulation of emotion and the devel-
opment of other self-regulatory capacities, and clarifying the inferential,
interpretive, and social-cognitive constituents of the self-regulation of emo-
tion. Necessary also are process models that distinguish operationally such
other elements of emotion regulation as situation appraisal, goal selection,
strategy choice, and outcome monitoring (cf. Garber, Braafladt, & Zeman,
1991). These are formidable challenges, but recognizing them as such steers
us away from prematurely assuming that either clear or consensual defini-
tions of emotion regulation currently exist. Most developmental researchers
(myself included) may "know" emotion regulation when they see it, but
considerably more is required if we are to develop a clear and comprehen-
sive definition of this phenomenon.
NATHAN A. FOX, ED.

WHAT IS REGULATED?

Perhaps the most central definitional quandary for the study of emotion
regulation concerns the issue of what is regulated when we consider the
management of emotion.' Because emotion is a multifaceted phenomenon
(involving physiological arousal, neurological activation, cognitive appraisal,
attention processes, and response tendencies), there are diverse avenues
toward the management of emotion, and consideration of these avenues
reveals that the term emotion regulation does not refer to a unitary phenome-
non but is rather a broad conceptual rubric encompassing a range of loosely
related processes.

Neurophysiological Constituents

At the core of these processes are the systems of nervous system organi-
zation that have evolved to regulate arousal (including emotional arousal)
through the interplay of excitatory and inhibitory mechanisms. Many of
these systems are functionally immature at birth, and their progressive mat-
uration and consolidation not only foster greater behavioral and emotional
self-control in the early years of life but also permit greater susceptibility
to extrinsic regulatory influences and allow for the enlistment of emotion
in aid of strategic behavioral processes. Moreover, individual differences in
the reactivity of some of these systems reflect both biologically and experien-
tially based processes that can influence personality and social functioning.
Thus, one answer to the question. What is regulated? concerns the control
of underlying arousal processes through maturing systems of neurophysio-
logical regulation.
Any explication of these maturing neurophysiological systems must be
both complex and incomplete, partly because the subcortical and cortical
systems affecting emotional arousal are mutually interconnected and are
intimately linked with other neurophysiological systems, including those
governing cognition and vegetative regulation, in ways that researchers are
just beginning to understand (see, e.g.. Fox & Fitzgerald, 1990). Current
advances in research in the neurosciences, and in developmental neuro-
physiology especially, will, one hopes, contribute to a clearer future picture
of these regulatory processes. However, early developmental changes in
emotion regulation can be linked to at least two neurophysiological advances
in the first year of life that provide a foundation for more complex regula-
tory processes with growing maturity.

' I am grateful to Stephen Porges for having raised this provocative question during
the working meeting on emotion regulation on which this Monograph is based.
 THE DEVELOPMENT OF EMOTION REGULATION

First, the diffuse excitatory processes underlying organismic arousal

decline in lability throughout the first year. This is partly due to postnatal
changes in the functioning of the hypothalamic-pituitary-adrenocortical sys-
tem that governs reactions to stress and uncertainty (as discussed in Gunnar,
1986, and Stansbury and Gunnar, in this volume) and maturational changes
in parasympathetic regulation as indexed by vagal tone (Izard et al., 1991;
Porges, 1991; Porges, Doussard-Roosevelt, and Maiti, in this volume). As a
consequence, organismic arousal gradually becomes more graded as well
as emotionally and motivationally more complex with increasing age. The
declining lability of organismic arousal during the early postnatal months
also aids caregivers' efforts to manage the emotion of offspring as well as
enhancing the effects of other emergent, internal controls on emotion.
Second, cortical inhibitory controls over arousal emerge gradually dur-
ing infancy, although some do not become fully functional until long after
birth. By about 2-4 months, for example, the growth of rudimentary fore-
brain inhibitory centers, changes in the organization of attention processes,
and other neurophysiological advances are manifested in behavioral state
changes (e.g., the emergence of more sustained attention and more regular
sleep-wake patterns, greater regularity and control of behavioral state, the
progressive disappearance of neonatal reflexes) as well as emotional changes
(e.g., an increase in exogenous smiling and a capacity for laughter) and
growing awareness of and emotional responsiveness to contingent stimula-
tion (Emde, Gaensbauer, & Harmon, 1976; Rothbart, Ziaie, & O'Boyle,
1992; Watson & Ramey, 1972). According to Dawson (in this volume) and
others (e.g.. Fox, 1991, in this volume), by about 9-10 months maturation
of the frontal lobe and its links to response inhibition (cf. Diamond, 1988)
fosters the capacity for arousal management and efforts to cope with emo-
tionally arousing events (see also Tucker & Frederick, 1989). These changes
may also underlie the growing vitality and complexity of emotional behav-
ior, such as the enhanced speed and intensity of emotional reactions
(Thompson, 1990), the use of emotional behaviors to share and affect the at-
tention and affective states of others (e.g., Scaife & Bruner, 1975; Stern,
1985), and the growth of emotional blends and other complex affective states.
Other cortical processes are also implicated in the growth of emotion
regulation capacities, and the end of the first year is not the terminus of
the neurophysiological changes that influence emotion regulation capacities
(cf. Kinsbourne & Bemporad, 1984). There remains, therefore, a consider-
able research agenda ahead of us if we are to elucidate the neurophysiological
substrates of emotion management capacities during the early years of life.
Individual differences in nervous system reactivity related to these
arousal regulatory processes have also generated considerable recent inter-
est. Consistent with the views of some temperament formulations that por-
tray individual differences in temperament in terms of variations in emo-

31
NATHAN A. FOX, ED.

tionality and self-regulation (e.g., Rothbart & Derryberry, 1981), findings

from several researchers suggest that differences in various neurophysiolog-
ical systems mediating emotion emerge early and are related to broader
features of social and personality functioning. In his longitudinal study of
behaviorally inhibited and uninhibited young children, for example, Kagan
and his colleagues have argued that inhibited children have a generally
lower threshold of reactivity in limbic structures mediating fear and defense
and that these differences can be found in both physiological and behavioral
measures through early childhood (see, e.g., Kagan, Reznick, & Snidman,
1988; Kagan & Snidman, 1991b; and, more generally, Kagan, in this vol-
ume). Fox and Calkins have compared groups of infants who differed on
similar indices of behavioral inhibition early in infancy and found predict-
able later differences on measures of emotion regulation and attachment
(see Calkins & Fox, 1992; Fox & Calkins, 1993; and, more generally. Cal-
kins, in this volume; Fox, in this volume).
Taken together, these findings suggest that early emerging individual
differences in physiological reactivity and regulation are related to varia-
tions in emotionality and emotion regulation. These differences are not
immutable, however; both Kagan and Fox point out that they can be modi-
fied by caregivers' socialization efforts as well as other experiences that alter
organismic reactions to stress and challenge (cf. Dienstbier, 1989). Conse-
quently, not only confirming the neurophysiological bases for these individ-
ual differences in emotional reactivity and emotion management but also
denoting sources of continuity and change in functioning remains a signifi-
cant future research task.
In sum, capacities for emotion regulation and self-management are
based, in part, on neurophysiological constituents that unfold during the
first year and provide the basis for more complex forms of emotion manage-
ment in later years. What is regulated is, in part, the neurophysiological
processes underlying emotional arousal and its management.

Attention Processes

Another way that emotion can be regulated is by managing the intake

of emotionally arousing information. Attention processes assume an emo-
tionally regulating function from very early in life. According to Rothbart
(Rothbart, Posner, & Boylan, 1990; Rothbart et al., 1992), maturational
changes in the neurophysiological organization of visual control between 3
and 6 months of age permit the infant to shift attention between stimulus
events voluntarily, in contrast with the "obligatory attention" observed at
younger ages. Not only does this enable the infant to disengage visually
from emotionally arousing events (which becomes more commonly ob-

32
 THE DEVELOPMENT OF EMOTION REGULATION

served at this time, e.g., during episodes of mother-infant play; Gianino &
Tronick, 1988), but it also enables parents to use visual distraction as an
emotion regulation strategy with very young offspring. Consistent with this
view is Rothbart's report of an association between individual differences
in visual disengagement and soothability in young infants together with a
general increase in soothability between 3 and 6 months of age (see Rothbart
et al., 1992).
With increasing age, the regulation of emotion through the manage-
ment of attention processes becomes more complex. The redirection of
attention is commonly enlisted by caregivers as a means of regulating emo-
tion in children, for example, during threatening or stressful events, when
caregivers may focus attention on positive features of the experience, dis-
tract the child during the event itself, or limit the child's knowledge of
potentially upsetting information (Miller & Green, 1985). With the assis-
tance of caregivers, young children can also sometimes regulate their emo-
tions themselves, using such attention management strategies as covering
their eyes or ears in emotionally arousing situations, removing emotionally
evocative stimuli, or leaving the situation altogether (e.g., Altschuler & Ru-
ble, 1989). Indeed, such approaches to the self-regulation of emotion are
among the earliest strategies observed in young children: attention-based
strategies have been observed in 4- and 5-year-olds in the presence of adults
arguing (Cummings, 1987), and such strategies can be understood and artic-
ulated by even younger children (e.g., one 28-month-old was reported as
saying, "I scared of the shark. Close my eyes"; Bretherton, Fritz, Zahn-
Waxler, & Ridgeway, 1986). In situations involving delayed rewards, redi-
rection of attention away from the reward while awaiting permitted access
is a behavioral strategy commonly observed in children between the ages of
2 and 6 years (Mischel & Mischel, 1983; Vaughn, Kopp, Krakow, Johnson,
& Schwartz, 1986).
As children acquire more complex, psychologically oriented concepts
of emotion, their strategies for the self-regulation of emotion increasingly
involve the internal redirection of attention, for example, thinking pleasant
thoughts during a distressing or frightening experience or self-coaching
that focuses on positive outcomes (e.g.. Band & Weisz, 1988). Even young
children are aware that the intensity of emotion experience tends to wane
over time as people cease to think about emotionally arousing events (Har-
ris, Guz, Lipian, & Man-Shu, 1985) and that "behavioral distraction"—such
as doing something else that takes your mind off emotionally arousing cir-
cumstances—can help you manage your emotions (Altschuler & Ruble,
1989). In one study, 8- and 13-year-olds in an English boarding school
knew, for example, that thinking about positive aspects of the situation
could help alleviate feelings of loneliness when away from home (Harris &
Lipian, 1989), and, according to American children of the same ages, inter-

33
NATHAN A. FOX, ED.

nal distraction helps while waiting during a delay task (Mischel & Mischel,
1983). Knowledge of these internal attention management strategies pro-
vides a very effective means of regulating one's own emotions because they
can be used in situations where escape or avoidance of emotionally arousing
stimuli is impossible. In these situations, what is regulated is the focus of
attention and the intake of information that affects one's emotional con-
dition.

Construals of Emotionally Arousing Events

At other times, emotion is regulated through other components of in-

formation processing. Rather than restricting the intake of emotionally
arousing information, individuals emotionally self-regulate by altering their
interpretations or construals of this information. This is classically illus-
trated by defense mechanisms that reduce anxiety and other negative emo-
tions through denial, projection, rationalization, repression, etc., with the
result that construals of reality are altered and emotion is thereby managed
(cf. Case, Hayward, Lewis, & Hurst, 1988). Other examples are children
who reinterpret the outcomes of scary stories in more emotionally satisfying
ways (e.g., "He didn't really die; he just got frightened and ran away") or
who think "it's just pretend" when listening to a sad account (cf. Meerum
Terwogt, Schene, & Harris, 1986).
Like most of the avenues to emotion regulation discussed in this essay,
children's construals of emotionally arousing situations are often the target
of extrinsic regulatory efforts, especially when offspring encounter poten-
tially stressful or challenging experiences (Miller & Green, 1985). Before
intrusive medical examinations, for example, parents may liken the proce-
dures to "tickling," or they may exaggerate their delight during a carnival
ride when the child appears frightened or look positive when children are
approached by unfamiliar but harmless adults. Such strategies are poten-
tially risky if they present children with a construction of current experience
that is significantly different from the child's own or if they regularly con-
tribute to the development of dysfunctional inferential biases (as may hap-
pen with the offspring of parents with affective disorders; see Cicchetti,
Ganiban, & Barnett, 1991; Zahn-Waxier & Kochanska, 1990). But they pro-
vide a significant means for emotional arousal to be externally managed by
altering the child's construction of the event when emotionally arousing
experiences are unavoidable.
Children also create their own interpretive constructions of emotionally
arousing experiences, of course, and these constructions commonly focus
on the achievement or frustration of personal goals and inferences concern-
ing the causes of success or failure that can have powerful emotional conse-

34
 THE DEVELOPMENT OF EMOTION REGULATION

quences (e.g., Graham & Weiner, 1986; Stein & Levine, 1989; Stein & Tra-
basso, 1989; Thompson, 1987a, 1989). Not surprisingly, kindergartners are
aware that goal substitution is a reasonable response to feelings of sadness
or anger that have been evoked by the frustration of their initial goals,
probably because goal substitution has consequences for emotion manage-
ment (Stein 8c Trabasso, 1989). For example, a child who learns that a
parent does not have time for a bedtime story might decide that playing a
game or listening to music is just as good, and this reinterpretation of the
event can help regulate feelings of sadness or dismay. Altering one's causal
attributions for emotionally arousing events is another way that reconstruals
can have emotionally regulatory consequences (e.g., "Tommy probably
didn't mean to knock over my tower"), and this strategy is often used by
adults, although there have been no studies of children's understanding of
this approach to the management of their emotion experiences. In sum,
when these kinds of internal or extrinsic regulatory strategies are employed,
what is regulated is one's interpretations of emotionally meaningful infor-
mation.

Encoding of Internal Emotion Cues

Another answer to the question. What is regulated? is that individuals

commonly manage their encoding of internal cues of emotional arousal
when they regulate emotion. In other words, emotional arousal is managed
not only by reinterpreting the circumstances eliciting emotion but also by
reinterpreting the internal indicators of emotional arousal, such as rapid
heart rate, increased breathing rate (or shortness of breath), perspiration,
and other concomitants of emotional arousal. One manner of managing
stage fright, for example, is to regard these physiological cues as the ordi-
nary accompaniment to appearing before an audience rather than as signals
of impending dysfunction. Similarly, children who can more easily channel
emotional arousal into adaptive social functioning have perhaps learned to
regard their internal cues of arousal as facilitating (e.g., empowering) their
goals, while children who are more easily undermined by heightened emo-
tion may perceive these internal cues of arousal as reflections of their incom-
petence or inadequacy. At present, however, we have little knowledge of
bow children understand these internal cues of emotional arousal or their
management. Nor do we know how parents influence these constructions
of offspring.

Access to Coping Resources

Emotion regulation also occurs by enhancing one's access to coping

resources, both material and interpersonal. In this sense, what is regulated

35
NATHAN A. FOX, ED.

is the availability of external support for managing emotional arousal.

Adults turn to friends and family for advice when anxious, comfort when
bereaved, and a cool head when angry, and young children are aware of
the benefits of eliciting nurturance from others when experiencing negative
emotion (Masters, Ford, & Arend, 1983; McCoy & Masters, 1985). In these
cases, access to coping resources is enhanced by seeking familiar and trusted
social partners, and this mode of emotion regulation begins early in life.
Indeed, the "secure base behavior" commonly observed in infants who en-
counter threatening or stressful circumstances with their caregivers reflects
the extent to which access to interpersonal coping resources can assist in
emotion management from a very early age. The social expectations under-
lying a secure infant-mother attachment suggest that this perceived access
can have broader consequences for socioemotional functioning (see also
Cassidy, in this volume). Moreover, parents also enlist the aid of material
coping resources (e.g., a favorite toy, blanket, or book) to assist their off-
spring in coping with emotional demands.
With increasing age, access to coping resources as an aspect of emotion
regulation becomes more planned and strategic. Friends are sought out for
their emotional support and understanding (Gottman & Mettetal, 1986),
and peers may be selected as confidants who have been especially sympa-
thetic on previous occasions. Indeed, when others are expected to be sup-
portive, children may sometimes enhance the intensity of their expressed
emotion in order to foster a desired reaction from another that supports
one's own emotionally self-regulatory efforts (Dunn & Brown, 1991; Saarni,
1992). For example, a child may exaggerate the hurt that she experiences
when tripping and falling to elicit sympathy rather than derision from peers
who are looking on as well as to avoid embarrassment, and this helps her
feel better. By age 6, children are aware that emotion displays can be altered
to mislead onlookers about the quality of one's distress (Harris & Gross,
1989), but we know little about their capacity strategically to manipulate
their emotion displays for purposes of emotion regulation, and this consti-
tutes another important research task. As an aspect of emotion regulation,
however, enhancing access to coping resources—especially interpersonal
resources—can entail the strategic as well as the incidental use of social
partners and material resources.

Regulating the Emotional Demands of Familiar Settings

Another answer to the question. What is regulated? is that emotion

regulation commonly involves predicting and controlling the emotional re-
 THE DEVELOPMENT OF EMOTION REGULATION

quirements of commonly encountered settings. That is, emotion experience

is managed as one selects and creates living circumstances that have manage-
able emotional demands.
Like other modes of emotion management, regulating the emotional
demands of familiar settings is one way that parents extrinsically manage
the emotion experience of offspring. They often restrict or expand the
opportunities for emotional arousal experienced by young offspring by con-
trolling the emotional demands of caregiving routines and other common
experiences (e.g., frequency of parent-child separations, promptness of re-
sponding to distress cries, etc.). In doing so, parents take into account both
their own child's temperamental strengths and vulnerabilities and the social-
ization demands of the emotion culture. Thus, one parent may use some-
what nondemanding or permissive childrearing practices in light of her
child's proneness to distress but will nevertheless increment expectations for
emotional tolerance as the child's increasing age permits greater emotional
self-control.
With increasing maturity, however, individuals become more capable
of selecting or constructing environmental settings in light of their self-
perceived needs and characteristics, including consideration of the emo-
tional demands with which they are comfortable (Lerner & Busch-
Rossnagel, 1981). Within certain limits, for example, social relationships are
chosen, home and workplace settings crafted, and commitments scheduled
to create desirable incentives, supports, and expectations, including man-
ageable emotional requirements. For young children, this may involve mak-
ing choices as simple as playing alone or with congenial peers rather than
selecting competitive games; for older people, the range of choices can be
much broader. Although this necessarily leads in idiosyncratic directions
according to individual constellations of personality and temperament,
there are also some developmental trends in the kinds of emotional de-
mands that adults integrate into the environmental and interpersonal set-
tings that they select.
Carstensen (1991; Fredrickson & Carstensen, 1990) has noted, for ex-
ample, that older adults select settings and relationships that ensure man-
ageable and predictable socioemotional demands, maximize positive affect,
and conserve physical energy. Their capacity to regulate the emotional de-
mands that they experience is also reflected in their ability to restructure
their lifestyles to accommodate unexpected emotional needs (e.g., clearing
the schedule for a mid-afternoon "time-out"), escape from a given situation
(e.g., by a spontaneous or planned retreat), or renew or strengthen ties to
supportive individuals (e.g., by more frequent visits, calls, or letters to a
sibling or an offspring). These ways of structuring and restructuring one's
lifestyle can have emotionally managing functions at all ages.

37
NATHAN A. FOX, ED.

Selecting Adaptive Response Alternatives

A final answer to the question. What is regulated? is that emotion regu-

lation commonly involves expressing emotion in a manner that has satisfac-
tory consequences—in other words, that is concordant with one's personal
goals for the situation. For adults, this might entail enlisting anger in a
search for solutions or a persuasive argument rather than insults or physical
attack. For a preschooler, this might involve insisting that a peer who has
destroyed a block tower help reconstruct it rather than angrily attacking
the perpetrator or using language to negotiate a parent's demands rather
than erupting into angry crying. In each case, it is not just that emotion is
regulated in order to achieve personal goals (e.g., reconstructing the tower,
finding a solution) but that the expectation of goal achievement also facili-
tates emotion regulation because of the prospect of beneficial outcomes. In
this respect, the availability of satisfactory response alternatives can promote
emotion management by offering modes of emotion expression that have
predictably satisfactory outcomes for the individual. By implication, emo-
tion regulation is undermined when there is a very limited range of re-
sponse possibilities or, alternatively, when existing options are perceived to
lead consistently to undesirable outcomes.
Students of display rules have noted, of course, that individuals often
minimize or intensify the expression of felt emotion and that doing so com-
monly has emotionally managing consequences (cf. Saarni, 1990). There
are, of course, situations in which managing emotional responses has little
effect on underlying emotion experience, especially when arousal is strong
and salient. Moreover, certain modes of expressing emotion (such as venting
anger) can undermine rather than enhance emotion regulation. The strong
connections between emotional feeling and emotional responding suggest,
however, that emotion regulation is often best accomplished by altering
how one expresses emotion—especially when there are available means of
conveying emotion that can help accomplish one's goals. In these situations,
emotion management occurs because emotion has potentially satisfactory
outlets.
This analysis implies that, as young children acquire a broader reper-
toire of modes for expressing emotion (owing to developmental changes in
expressive capabilities as well as a broadened behavioral repertoire), their
capacities for emotion regulation are likely to be enhanced as this repertoire
is strategically employed. For example, in her longitudinal home-
observational study of young children, Kopp (1992) noted that crying
peaked in frequency late in the second year, with a progressive decline in
crying throughout the third and fourth years. This trend replicates observa-
tions by earlier investigators, and Kopp has suggested that crying declines
at this time because this is when language emerges as a significant alternative

38
 THE DEVELOPMENT OF EMOTION REGULATION

means for expressing emotion and emotion-related experiences (indeed,

older children in her study showed many more verbal refusals and "off-task
negotiations" in circumstances where younger children simply fussed). In
many circumstances, language can express emotion more effectively than
crying while at the same time accomplishing situational goals, and this may
account for its enhanced utility during the preschool years.
With increasing age, the range of expressive alternatives for emotion
broadens, of course, but the expression of emotion also begins to be chan-
neled in directions that the emotion culture finds acceptable. This is be-
cause, on the basis of the socialization practices of parents and other authori-
ties, children acquire emotion schemas that, among other things, guide their
predictions of the consequences of expressing various emotions in certain
situations (cf. Saarni, 1990). For example, on the basis of the verbal (e.g.,
"Use words rather than hitting") and behavioral guidance of socialization
agents, they learn the consequences of responding to an aggressive peer by
taking revenge, crying, tattling to adults, or asserting themselves. In doing
so, they can more thoughtfully evaluate these alternatives in terms of their
relative suitability for accomplishing personal goals in particular circum-
stances.
The reactions of others are, of course, especially important in evaluat-
ing the predictable consequences of different modes of emotional re-
sponding for accomplishing one's goals. Because of this, some reactions will
be better suited to certain social settings than others and will thus better
advance one's regulation of emotion. Recourse to an adult when frustrated
by a peer may be praised by a preschool teacher but regarded negatively
by a parent. Similarly, crying may be maladaptive for toddlers in some
settings (e.g., when used to resist the mother's request to clean up toys) but
accomplish valuable strategic ends in others (e.g., when calling attention to
sudden danger or an older sibling's aggression). In short, the most adaptive
means of expressing an emotion is often situation specific rather than trans-
situational and is based on the child's expressive repertoire, the demands
of the setting, the goals of the child, and the values of social partners.
This can be a complex calculus (as most adults know from experience with
emotionally charged situations) and suggests not only that selecting an adap-
tive response alternative can be complicated but also that this facet of emo-
tion regulation is tied to the growth of social cognition and social compe-
tency in childhood.

Implications

There are clearly diverse developmental pathways toward emotion reg-

ulation, deriving both from the efforts of external agents to manage the

39
NATHAN A. FOX, ED.

emotions of children and from the child's growing capacity to self-regulate

(for an insightful analysis, see also Calkins, in this volume). These are built
on emerging neurophysiological foundations for arousal regulation and in-
clude controlling attention processes, altering construals of emotionally
arousing situations, modifying the encoding of internal emotion cues,
strengthening access to coping resources, regulating the emotional demands
of familiar settings, and selecting adaptive modes for expressing emotion.
Each of these pathways provides different answers to the definitional ques-
tion. What is regulated?
Competent emotion regulation can thus involve any of these processes,
taken individually or in combination. Because we know very little about
developmental changes in the use of these strategies and their consequences
for children's management of emotion, the tasks of constructing a develop-
mental analysis of the growth of emotion regulation skills and identifying
the origins of individual differences in capabilities for emotion management
are complex indeed. In embarking on these tasks, for example, it is impor-
tant to understand the diverse constituents of developmental growth in
emotion regulation skills. To what extent does growth in emotion manage-
ment skills derive (a) from children's developing awareness of the need for
emotion regulation, (b) from a growing repertoire of emotion regulation
strategies (if so, which ones emerge at which ages?), (c) from enhanced
strategic knowledge of the potential utility of different regulatory ap-
proaches in different situations, (d) from growing flexibility in substituting
one regulatory approach for another, (e) from an emerging capacity to
adapt regulatory strategies to different contexts and situational demands,
and (/) from enhanced skills at evaluating (or predicting) the relative suc-
cess of different regulatory approaches?
New inquiries may also be made concerning the nature and origins of
individual differences in emotion regulation among children of a given age.
To what extent do these derive (a) from relying on different regulatory
strategies, (b) from using strategies that differ in their predictable success,
(c) from using strategies that are situation specific rather than transsitua-
tional (i.e., that vary in their flexibility), (d) from using an impoverished as
opposed to a rich repertoire of alternative regulatory strategies, (e) from a
limited understanding of the conditions in which different emotion regula-
tion approaches are useful, suitable, or potentially successful, and (/) from
situationally specific success in emotion regulation that is not generalized?
These questions and related inquiries constitute an important agenda for
future research on emotion regulation and will likely provide considerable
insight into processes of emotional development in general as they contrib-
ute to an awareness of how emotion changes with increasing age beyond
the unfolding of a capacity for discrete emotions.
At the very least, however, the heterogeneity of these emotion regula-

40
 THE DEVELOPMENT OF EMOTION REGULATION

tion processes cautions against regarding the development of emotion regu-

lation as a homogeneous growth process, instead underscoring that emotion
regulation is a conceptual rubric that encompasses a variety of behavioral
strategies, each with likely different developmental timetables and experien-
tial origins

EMOTION REGULATION AND SOCIAL INTERACTION

The development of emotion regulation has become a central interest

of functionalist theorists, who believe that emotion is constituted by the
ongoing transactions between individuals and their environments. Within
this view, "families" of emotion are crucial to social signaling, communica-
tion of needs, enhancement of affiliational ties, self-defense, and other im-
portant goals. Strategies of emotion regulation enlist emotion to achieve
these goals. The functionalist analysis suggests that social encounters pro-
vide the most salient contexts for exercising skills of emotion management
and that the efficacy of these skills depends significantly on the responses
of social partners and the demands of the social setting as they are pertinent
to one's goals. In this view, exploring how the development of strategies of
emotion regulation fits into the fabric of social relationships is thus an im-
portant task in elucidating the functional significance of processes of emo-
tion management and control.
The social context affects emotion regulation in a variety of ways. One
obvious way is that social partners regulate our emotions from early in life.
Generalizing from studies of rats, Hofer (in this volume) illustrates the
powerful effects that relationships might have on regulating physiological
homeostasis and the emotions associated with early attachment and bond-
ing. In humans, a major task of successful parenting is managing and guid-
ing the emotion experience of offspring. This occurs not only through
direct intervention to relieve distress, fear, frustration, and other negative
emotions (cf. Gekoski, Rovee-CoUier, & Carulli-Rabinowitz, 1983; Lamb &
Malkin, 1986) but also through modeling and selective reinforcement of
positive emotion expression (e.g., Malatesta-Magai, 1991), the direct induc-
tion of emotion through such processes as affective contagion, empathy,
and social referencing (e.g.. Stern, 1985; Thompson, 1987b; Walden, 1991),
verbal instruction about emotion and emotion regulation strategies (Dunn
& Brown, 1991; Miller & Sperry, 1987), and the control of opportunities
for emotional arousal through the organization of caregiving demands and
the environment of early development (e.g., early independence training,
quality of out-of-home care, etc.). In these and many other ways, caregivers
extrinsically manage emotion experience through the emotional demands
NATHAN A. FOX, ED.

that they impose on young children and the interpersonal supports that
they provide for containing emotional arousal within manageable limits.
With the child's increasing age, the emotional demands of caregivers
and the strategies that they use for managing the emotions of offspring
evolve in accordance with the child's growing repertoire of emotions and
developing skill at emotion management and the changing demands of the
emotion culture (e.g., Lewis & Michalson, 1983; Miller & Sperry, 1987;
Saarni, 1990). Parents have fairly clearly defined expectations for the emo-
tional behavior of their offspring that change with situational demands and
the child's developing capabilities for the self-management of emotion, and
they use a broadened range of direct and indirect influence strategies to
socialize the child's emotional behavior. But direct efforts to regulate an-
other's emotion are not limited to socialization processes in childhood. As
adults, we seek to manage the emotions of others by extending a sympa-
thetic ear or reassurance, emphasizing the consequences of neglected re-
sponsibilities, using humor in a distressing situation, and in other ways alter-
ing another's experience of emotionally relevant events.
These influences are especially important in the context of meaningful
and long-standing social relationships. Such relationships are important not
just because partners can have mutual, long-term effects on the arousal and
management of emotion but also because of the emotional dimensions of
the relationships themselves and the social expectations that they engender.
Because attachment figures, friends, parents, spouses, offspring, and sig-
nificant others constitute valuable interpersonal resources for coping with
emotion, expectations concerning their accessibility, helpfulness, and sensi-
tivity can significantly enhance—or undermine—the capacity to manage
emotional arousal. A child can more easily cope with a distressing experi-
ence because of the anticipated understanding provided by a parent, a
friend, or a sibling (as suggested by the research reported by Field, in this
volume). Conversely, the anticipation of an uncaring or a denigrating re-
sponse might cause children to restrict the range or vitality of their expres-
sions of emotion in the presence of such partners or to experience difficulty
in coping with strong arousal.
As Cassidy (in this volume) has commented, individual differences in
adult attachment representations as well as infant attachments may be asso-
ciated with distinct styles of emotion regulation that entail minimizing or
enhancing different emotions in interaction with attachment figures. In
other words, on the basis of expectations concerning the partner's availabil-
ity and sensitivity, infants as well as adults may learn to disguise or enhance
the expression of feelings in a relationship (cf. Thompson & Lamb, 1983).
While it is certainly true that there are multiple catalysts underlying these
differences in attachment and attachment representations, social expecta-
 THE DEVELOPMENT OF EMOTION REGULATION

tions may be one important influence on the development of strategies for

regulating emotion and emotion displays within these relationships.
In addition to the expectations engendered by close relationships, social
partners also influence emotion regulation by affecting the interpretation
of emotionally arousing situations and the coping resources that are avail-
able. They may reinforce attributional styles that enhance or inhibit certain
emotions (e.g., "It's not your fault!") and may foster certain coping responses
by direct instruction or modeling, by providing instrumental or material
assistance, or by offering counseling and emotional sustenance (cf. Miller
& Sperry, 1987). For example, offspring who regularly observe parents
suppress emotion displays, perhaps in conjunction with verbal comments
(e.g., "We don't have to fly off the handle!"), are likely to internalize such
strategies as first-resort approaches to managing their own emotion experi-
ences. Furthermore, these relationships may themselves impose emotional
demands that can undermine as well as enhance effective regulatory efforts.
While offering significant support, close relationships are often simulta-
neously sources of emotional stress or turmoil that can also affect efforts to
manage emotion (Thompson, 1992).
In sum, the development of emotion regulation is well integrated into
the fabric of social relationships not only because of the direct and indirect
ways that people seek to manage emotion in others but also because of the
social expectations generated by close relationships with friends and rela-
tives (especially expectations of support and understanding), their influence
on how individuals interpret emotionally arousing situations and the re-
sources that are available to them, and the emotional dimensions of these
relationships, including the demands as well as the support that they
provide.
Recent research in developmental psychopathology provides informa-
tive, albeit distressing, illustrations of the diversity and importance of the
influence of relationships (Thompson, Flood, & Lundquist, in press). Young
offspring of parents with affective disorders are at heightened risk of emo-
tion regulation problems owing not only to the caregiver's limited availabil-
ity as a source of emotional support but also to the adult's modeling of
negative attributional styles and use of discipline practices that enhance the
child's feelings of responsibility and helplessness (Zahn-Waxler & Kochan-
ska, 1990). Parent-child relationships shape not only children's construals
of emotionally arousing situations but also their resources for regulating
emotion: these offspring (like their parents) have difficulty devising appro-
priate strategies for modifying emotion and lack confidence in the efficacy
of these strategies (Garber et al., 1991). Children from homes characterized
by marital conflict show a heightened sensitivity to distress and anger that
is manifested in excessive guilt and diminished coping with adult arguments

43
NATHAN A. FOX, ED.

(Cummings, Zahn-Waxler, & Radke-Yarrow, 1984; Cummings et al., 1989;

Katz & Gottman, 1991). The effects on emotion regulation of the emotional
demands of distressed caregivers are manifested most clearly in children
maltreated by their parents, who sometimes respond with depressed affect,
heightened lability, or marked anger (Gaensbauer & Sands, 1979). Thus,
diverse facets of significant relationships—expectations of helpfulness or
insensitivity, the partner's influence on the interpretation of emotionally
arousing situations, instruction in and modeling of strategies for managing
emotion, and the emotional requirements of these relationships them-
selves—can have important effects on the development of emotion regula-
tion strategies.
Thus far, we have been considering how relationships affect the devel-
opment of emotion regulation. But it is important to note that the reverse
is also true: emotion regulation strategies can significantly influence the
course of social interaction and the development of social relationships. In
a series of well-known investigations. Dodge and his colleagues have found
that aggressive children tend to be deficient processors of social cues—
especially when they feel threatened—and consequently construe hostile
intent in ambiguous or uncertain social encounters with peers (e.g.. Dodge,
1991a; Dodge & Somberg, 1987). But it is quite likely that deficiencies in
emotion management as well as social information processing contribute
to their social dysfunction: under threatening circumstances, the affective
salience of social cues, their interpretation, and the thoroughness of one's
search for and evaluation of alternative response options are all likely to be
affected by skill at self-regulating emotion. Children who can "keep their
cool" when threatened may be better able to think carefully about the situa-
tion and devise competent and successful response strategies. In a sense,
emotion management may be both a contributor to and a result of the
quality of social information processing that leads to successful or unsuccess-
ful encounters with peers.
Needless to say, the development of the skills involved in emotion man-
agement is affected by a panoply of significant relationships in varied set-
tings: in the context of parent-child relationships, out-of-home care settings,
peer interactions, school settings, and, later, workplace associations, strate-
gies for emotion self-regulation are fashioned and refined. A number of
important questions arise from such an analysis. Given the challenges to
"emotional competence" in troubled parent-offspring relationships noted
above, for example, can children acquire more successful strategies of emo-
tion management through extrafamilial support, such as with a teacher or
with peers? Are there qualitatively different kinds of emotion regulation
strategies—or skills for different emotional demands—acquired in peer as
opposed to family contexts? Finally, how do variations in the "emotion cul-
ture" (Gordon, 1989) observed cross-nationally affect the interpretation of

44
 THE DEVELOPMENT OF EMOTION REGULATION

emotion experience and the requirements for the self-regulation of emo-

tion? These and other questions are fascinating catalysts for the study of
the development of emotion regulation, and especially of individual differ-
ences in regulatory capacities, in the context of close relationships.

THE CORRELATES OF INDIVIDUAL DIFFERENCES

IN EMOTION REGULATION

Current enthusiasm for the study of emotion regulation doubtlessly

also derives from its practical applications. The study of emotion regulation
provides an arena within which problems of social competence and incom-
petence, behavioral self-control, and even intellectual and cognitive func-
tioning can be regarded in a new light. By characterizing these differences
as partly a function of individual differences in emotion regulation pro-
cesses, researchers not only contribute new ideas about the origins of these
social and cognitive differences but also begin to identify new intervention
and remediation strategies. Implicit in these efforts is the view that differ-
ences in emotion regulation skills can be reliably identified and that a coher-
ent formulation of adaptive or "optimal" emotion regulation can be framed
as a guide to intervention efforts.
The preceding discussion indicates why this is both a worthwhile goal
and a compellingly challenging one. Because emotion regulation encom-
passes heterogeneous developmental processes, individual differences in
emotion regulation are likely to occur along multiple dimensions rather
than on a single axis. Individuals likely vary, for example, in their knowl-
edge of the need for emotion regulation in specific situations, their aware-
ness of alternative strategies, their flexibility in applying different regulatory
strategies, and other components of emotion control. There is no necessary
reason why individuals should exhibit deficiencies in all aspects of emotion
regulation or in all situations; individual patterns of skill, difficulty, and
compensation may be the rule. This makes the tasks of identifying the
nature of these individual differences and of designing effective interven-
tion approaches considerably more complex.
Adding to this complexity is the need to define clearly rather than
intuitively what we mean by optimal emotion regulation (or, on the other
hand, emotion dysregulation). As Cole, Michel, and Teti (in this volume) have
noted, various clinical approaches emphasize different facets of optimal
regulation, and most are difficult to operationalize. In general, optimal emo-
tion regulation could be defined for either clinical or research purposes as
a process or an outcome. Many formulations of emotion regulation regard
optimal regulation in terms of its outcomes: the individual is capable of
keeping emotions under sufficient control to allow for interpersonal relat-

45
NATHAN A. FOX, ED.

edness and sociability, prosocial initiatives when appropriate, sympathy to-

ward others, personal assertiveness when needed, and/or other indices of
successful functioning. Effective emotion regulation is believed to be an
ingredient of these behaviors, and signs of "emotion dysregulation" are
commonly perceived in the absence of these capacities. But, in other re-
spects, optimal emotion regulation can be regarded as a process: the enlist-
ment of strategies that permit flexibility, quick reappraisals of emotionally
provoking situations, access to a broad range of emotions, and efficient goal
directedness. In this respect, emotion regulation is defined in terms of the
quality of emotion that results, regardless of its other behavioral conse-
quences.
The problem with these formulations is that the construct optimal emo-
tion regulation is so broadly defined that it becomes confounded with intuitive
values of what a well-functioning personality is like. Like the construct of
ego control that characterized earlier research on personality development,
emotion regulation is often regarded by contemporary researchers as a
stable component of personality functioning with broad manifestations in
diverse behavioral domains, with those who are "optimally" regulated show-
ing many positive sociopersonality characteristics that avoid the excesses of
either under- or overregulation. In these circumstances, it is helpful to
remember that optimal emotion regulation (or emotion dysregulation) is often
better defined by the demands of the immediate social situation and the
goals of the individual than as a global, personological construct. What is
"optimal" may vary for different individuals, in different situations, and
with different goals.
Consider, for example, the case of a child who gets angry at a peer
who has wronged her. In that situation, does "optimal" emotion regulation
involve retaliation, or tattling to an authority, or avoiding the perpetrator,
or insisting on the perpetrator's apology, or crying—or some combination
of these responses? I suspect that this depends on many factors that are
specific to this situation, such as the child's power relative to that of the
wrongdoer, the values of the adults to whom the child might turn, the
behavior of other children in the setting, and the overarching values of
the sociocultural milieu (cf. Miller & Sperry, 1987). I suspect that "optimal"
emotion regulation also depends on the child's goals for that situation. In
this example, these goals might include reestablishing a sense of personal
well-being, ensuring that the wrongdoing does not recur, reestablishing
good relations with the perpetrator, restoring a sense of esteem within the
peer group—or some combination of these. In other words, emotion man-
agement does not necessarily involve diminishing unpleasant affect (al-
though it may); depending on the child's goals, anger might be enhanced (to
stand up to the perpetrator), modulated (to enlist the assistance of friends in
 THE DEVELOPMENT OF EMOTION REGULATION

self-defense), or blended (to provoke an adult's intervention through salient

distress). In a sense, the "optimal" choice depends on the child's goals for
a given situation.
Similar questions can be raised about other features of individual varia-
tions in the self-regulation of emotion. In certain circumstances, for exam-
ple, well-regulated emotion sensitizes the child to the emotions of others
and fosters an appropriate emotional response to them. In these situations,
the optimal self-regulation of emotion is likely to be associated with empathy
and prosocial initiatives (cf. Eisenberg & Fabes, 1992b). In other circum-
stances, however, this is a potentially dysfunctional outcome, such as when
the child witnesses domestic violence (and is thus at considerable risk by
intervening prosocially) or has a parent with an affective disorder (in which
case empathy may be disorganizing). Given differing circumstances, optimal
emotion regulation processes may yield different behavioral outcomes
indeed, one might regard a child's manifest distress or avoidance as a more
"optimal" self-regulatory strategy than one that yields emotional engage-
ment in a parent's personal turmoil.
Children do, of course, acquire characterological styles of emotion man-
agement that become increasingly important facets of successful or dysfunc-
tional aspects of personality functioning with increasing age. But a prema-
ture research focus on identifying and labeling these styles may cause
researchers to miss noticing the social-contextual processes and personal
goals that help define what optimal regulation is in that context. Researchers
may also neglect the fact that children can be effective managers of their
emotions in some situations (e.g., encounters with peers) and not in others
(e.g., sibling interactions) or that they may effectively accomplish some emo-
tional goals and not others. Moreover, in situations involving extreme emo-
tional demands (such as child maltreatment, marital conflict, or a parent's
psychopathology), "optimal" emotion regulation may be manifested in be-
havior that looks very different from what a more typical, well-functioning
personality manifests. Finally, a research emphasis on the social-contextual
constituents of emotion regulation will also sensitize researchers to the stan-
dards by which members of a culture (and subculture) define optimality.
In the future, I suspect that individual differences in emotion regula-
tion will be defined much less globally and in a manner that is far more
situationally specific than is presently the case, with careful attention to the
nature of the child's goals, other developmental capacities, and the contex-
tual demands that the child faces. This will contribute, 1 hope, to a develop-
mental picture in which individual patterns of compensation and specializa-
tion, rather than "optimality" and "dysregulation," characterize our
portrayals of developing individuals. And, in this context, I suspect that
future research will link the growth of emotion regulation to the growth of

47
NATHAN A. FOX, ED.

self-understanding and of social cognition as these processes are jointly

involved in the child's construction of emotion and its functions in social
contexts.

NEW APPROACHES TO THE STUDY OF EMOTION REGULATION

Definitions of psychological processes are closely allied with measure-

ment strategies, so it is perhaps appropriate that this essay close with a
brief overview of new methodological approaches to the study of emotion
regulation that have been explored in my lab with students and colleagues.
The purpose in doing so is not to propose that these approaches are neces-
sarily useful for all research purposes but rather to contribute to the variety
of research methods that are currently available to researchers in this bur-
geoning field of study.
Consistent with the definition of emotion regulation discussed earlier, our
methodological strategy has focused on measuring the dynamic features of
emotional responses observed in the infants we have studied. This is because
some of the most informative features of emotional arousal entail not varia-
tions in discrete emotions (which often covary) but rather changes in the
temporal and intensive features of emotion that reflect the appraisal and
regulatory processes related to emotion and that, in turn, influence many
of the functional properties of expressions of emotion (such as the reactions
that they elicit from social partners). We call these response features emotion
dynamics because, while the discrete emotion may "play the tune" of an
individual's response, emotion dynamics (like the dynamic markings on a
musical score) significantly influence quality, intensity, timing, and modula-
tion and thus significantly color emotionality. These response parameters
have long interested students of infant temperament and child clinical re-
searchers (especially those concerned with emotion regulation), but they
have seldom been effectively operationalized in studies of early emotional
development. This is the task that we have undertaken, together with study-
ing the meaning and correlates of individual differences and developmental
changes in these dynamic features of emotional responding.
We use either continuous time-sampled ratings or on-line temporal
assessments of facial and vocal measures of emotion to index response pa-
rameters like the latency of the response (i.e., time from the onset of the
eliciting stimulus until onset or peak of the emotional response), rise time
(i.e., time from the onset of the emotional response until peak intensity is
achieved), persistence (i.e., duration) of the emotional reaction, and recovery
(i.e., time from the terminus of the eliciting stimulus until emotional re-
sponses reach a neutral baseline). More broadly, we also examine the range
and lability of emotional responsiveness as well as indexing the intensity of

4a
 THE DEVELOPMENT OF EMOTION REGULATION

emotional reactions over short- and longer-term periods (see Thompson,

1990).
Needless to say, these response dynamics are multidetermined: many
factors are influential in shaping the latency, intensity, and other character-
istics of an emotional response, some of which involve emotion regulation
processes. Consequently, we have examined these response parameters in
the context of specific hypothesis-testing studies in which we have sought
to predict group differences in these emotion dynamics on the basis of
characteristics of the infants themselves or of their experiential history that
might contribute to differences in emotion regulation. Such an approach
addresses the study of emotion regulation strictly in terms of its influence
on emotional response parameters, somewhat independently of the broader
consequences for psychosocial functioning.
In one study, for example, the emotion dynamics of a sample of Down
syndrome (DS) infants in the Strange Situation procedure were compared
with those of a sample of typical infants observed twice in this procedure:
once when their ages were equivalent to those of the DS sample and also
earlier, when their mental age was more comparable to that of the DS
sample. Our purpose was to evaluate whether the emotion dynamics of
Down syndrome infants were uniquely different from those of the compari-
son sample or could instead be interpreted in terms of the DS infants'
cognitive lags (Thompson, Cicchetti, Lamb, & Malkin, 1985). Our results
indicated that the emotional responses of the DS sample differed signifi-
cantly from those of the typical sample at each age: Down syndrome infants
showed diminished emotion intensity, a decreased emotion range, limited
lability, a more prolonged latency to distress onset during separation epi-
sodes, but a quicker recovery during reunions compared with typical infants
regardless of whether they were matched for age or cognitively comparable
to the DS sample.
These differences suggest that the organization of emotional behavior
for DS infants is unique owing both to their physiological difficulties in
modulating arousal and to the cognitive retardation that blunts the speed
and efficiency of their appraisals of situations and events. Both physiological
and cognitive factors regulate the emotional behavior of DS (and typical)
infants via attention, interpretive, as well as physiological processes. This
conclusion has had, in turn, important implications not only for our under-
standing of how DS infants manage to address the socioemotional chal-
lenges of early growth but also for intervention and parent education (cf.
Cicchetti, 1990).
In other studies, we have examined whether the difficulties in alertness,
physiological stability, and arousal modulation commonly observed in pre-
mature babies would be manifested in differences (compared to full-term
infants) in their regulation of emotional arousal at the end of the first year.

49
NATHAN A. FOX, ED.

We discovered that the emotion dynamics of relatively healthy preemies

(i.e., without compromising medical complications) were no different than
those of typical infants at 12 months (Frodi & Thompson, 1985) but that
medically compromised premature infants showed significantly different
dynamics in their emotional behavior in the second year that reflected their
continuing difficulties modulating arousal (Stiefel, Plunkett, & Meisels,
1987). Such findings not only underscore the importance of considering
the long-term influences of early biological insult on emotion regulation
but have also indicated how the quality of infant-caregiver interactions can
support healthy emotional responsiveness in both healthy and medically
compromised premature infants.
In these studies as well as others (e.g., Connell & Thompson, 1986;
Thompson, Connell, & Bridges, 1988; Thompson & Lamb, 1984), we have
examined differences in emotion dynamics in the context of attachment
and found that variations in emotionality assume a central role in the nature
and stability of the differences between securely and insecurely attached
infants. These differences in emotion dynamics have temperamental as well
as nontemperamental origins (cf. Belsky & Rovine, 1988) and are likely
shaped by the ongoing features of caregiver-infant interaction that affect
other components of attachment-system functioning (cf. Thompson &
Lamb, 1983). This conclusion is consistent with emerging views of how early
caregiving contributes to the development of "emotional biases" in infants
(Malatesta, 1990) as well as of the role of temperamental individuality in
emotion regulation.
More recently, we have explored developmental changes in the dynam-
ics of emotion in a longitudinal study of infants observed in mother-infant
play, an encounter with a stranger, and a separation episode at ages 6, 9,
and 12 months (Thompson, 1990). This half year of life is a period of
striking changes in emotion regulation owing to rapid cognitive advances
that affect emotion appraisals (e.g., the growth of means-ends understand-
ing and of evocative memory skills), neurophysiological maturation in the
frontal lobes related to arousal regulation, the consolidation of social expec-
tations for familiar partners, the emergence of social referencing as a means
of construing emotionally arousing events, and the growth of self-propelled
locomotion, which alters the child's transactions with the social and nonso-
cial world in that it offers a new potential for goal attainment and feelings
of self-efficacy and frustration (Campos, Kermoian, & Zumbahlen, 1992).
Consequently, we expected that across this developmental transition infants
would show progressively enhanced emotion intensity, growing persistence
in their emotional responses, and decreased response latency and rise time
in their reactions—indicators, in short, of growing emotional "vitality."
This is precisely what we found across both positive and negative emo-
tion elicitors. That is, regardless of whether infants were positively or nega-

$0
 THE DEVELOPMENT OF EMOTION REGULATION

tively aroused, they showed greater speed and intensity in their emotional
reactions from 6 to 12 months owing, in part, to the regulatory changes in
emotionality outlined above. This emotional "vitality" appears to assume an
important role in the nature of the infant's transactions with the social world
during this period (cf. Tronick, 1989).
However, even though these developmental changes in emotionality
were consistent across positive and negative situations, the valence of the
reaction did make a difference for the organization of these dynamic fea-
tures. To summarize briefly, negative emotional arousal in infants was char-
acterized by a biologically based "emergency reaction" in which high distress
intensity was accompanied by a short latency, long persistence, but pro-
longed rise time to provide a rapid mobilization of the baby's resources in
the face of threat and to evoke a preemptory response from adult listeners.
By contrast, high positive emotional arousal involved a longer latency, a
shorter escalation to peak intensity, and much less persistence, fostering a
more sustained appraisal of situational events and a capacity to respond to
changes in those events but also an ability to become quickly engaged in
positive stimulation. In sum, these findings indicate significant develop-
mental changes in emotion dynamics from 6 to 12 months against a back-
ground of a consistent organizational structure of these dynamic features,
as one might predict on the basis of an appreciation of both the biologically
adaptive and the psychologically flexible features of the emotion regulation
processes discussed in this essay.
These studies underscore what the literature reviewed here also attests:
that continued study of the development of emotion regulation processes
is compelling because of the potential insights that it can contribute to our
understanding of emotional growth. With greater conceptual clarity con-
cerning the meaning of emotion regulation and its constituents and conse-
quences, many of these potential contributions are bound to be realized.

SUMAAARY

Contemporary interest in emotion regulation promises to advance iin-

portant new views of emotional development as well as offering applications
to developmental psychopathology, but these potential contributions are
contingent on developmentalists' attention to some basic definitional issues.
This essay offers a perspective on these issues by considering how emotion
regulation should be defined, the various components of the management
of emotion, how emotion regulation strategies fit into the dynamics of social
interaction, and how individual differences in emotion regulation should
be conceptualized and measured. In the end, it seems clear that emotion
regulation is a conceptual rubric for a remarkable range of developmental

51
NATHAN A. FOX, ED.

processes, each of which may have its own catalysts and control processes.
Likewise, individual differences in emotion regulation skills likely have
multifaceted origins and are also related in complex ways to the person's
emotional goals and the immediate demands of the situation. Assessment
approaches that focus on the dynamics of emotion are well suited to eluci-
dating these complex developmental and individual differences. In sum, a
challenging research agenda awaits those who enter this promising field of
study.

52