You are on page 1of 17

12 / Attachment in Adolescence

Diana S. Rosenstein and Harvey A. Horowitz

The family experience of those who grow up and process, organization and activity, differentia-
to become relatively stable and self-reliant is tion and integration, continuity and discontinuity,
characterized not only by unfailing parental sup-
port when called upon but also by a steady but stability and change.
timely encouragement toward increasing auton- 2. The relational and contextual: Development pro
omy, and by the frank communication by par- ceeds within a relational matrix; it is interactive
ents of working models—of themselves, of [the] and contextual. In human development, the indi-
child, and of others—that are not only tolerably
valid but are open to be questioned and revised.
vidual is seen as emerging or differentiating from
—BOWLBY, 1973, pp. 322-323 within a matrix of relationships and remaining in-
terdependent within relational contexts through
Although attachment theory has been central to out the life span.
3. The constructivist and metaphoric: Development
the understanding of the social development of
is a co-construction of meaning and knowledge
infants and young children for the past 30 years, through a coordination of actions, affective com-
the vicissitudes of attachment in the adolescent munication, language, awareness, and shared ex-
years did not receive comparable attention until perience. All human categories, including "reality"
relatively recently (Kobak & Sceery, 1988; Weiss, and "objectivity," are metaphoric, constructed
1982). In fact, it was commonly thought that ado- from experience, constrained by biology, and
lescence was a period of weakening attachment to finding stability in the consensual domain of hu
the family, with increased separation from parents man meaning systems.
and growing autonomy. It is true that direct and 4. The cybernetic and recursive: Development is a
immediate parental influence is less apparent dur- living process with systemic and holistic properties
ing adolescence than in earlier stages of childhood, embodying complex circular causality, self-refer
ence and self-generation (autopoiesis), and part-
but adolescence is a time when relationships with
whole relation and coherence.
parents come again to the fore and are heightened
in intensity compared to preadolescence. The By understanding these aspects of attachment
emergence of emotionally charged issues such as theory, it is possible to comprehend the paradox
separation, identity formation, and autonomy, of the increasingly autonomous adolescent who is
while confronting the adolescent with difficult de- firmly embedded in and profoundly influenced by
velopmental tasks, present opportunities for fur- attachment to parents.
ther differentiation and integration of the individ-
ual within the context of relationships.
Attachment theory is a developmental theory,
and it is grounded in a perspective that views Review of Attachment Theory
development as a goal-directed process of change,
during which new competencies and adaptive pat-
terns emerge from the reorganization of previous BOWLBY: THE ATTACHMENT BEHAVIORAL SYSTEM
patterns, structures, and competencies. According AND INTERNAL WORKING MODELS
to Werner (1957), this process "proceeds from an
initial stage of globality and lack of differentiation Bowlby (1958, 1969/1982, 1973, 1980) devised
to increasing states of articulation, differentiation attachment theory as a combination of features of
and hierarchical integration" (p. 126). This devel- ethology, control systems theory, cognitive sci-
opmental perspective is defined by four basic as- ence, and psychoanalytic theories. Following the
sumptions: British Object Relations school, Bowlby postu-
lated the primacy of social affectional ties between
1. The dialectical and paradoxical: Development is a child and his or her caregiver, independent of the
a dynamic process that embodies both structure secondary reinforcement of caregiving practices.


Using ethological explanations, the child's tie to tive development, mastery, and self-reliance are
the mother subserved a biological and adaptive fostered. A mother's (or other primary caregiver's)
function in an evolutionary context. The goal of ability to create a secure base works jointly with
enhanced survival of the child is mediated by the her ability to foster appropriate levels of autonomy
proximity of the child to the caregiver to accom- in her infant. As cognitive ability develops, the
plish the caregiver's protective and nurturing abili- toddler becomes less dependent on immediate
ties, which fulfills a complementary evolutionary sensory experiences and is able to use cognition
function. From these basic premises, Bowlby at- to ascertain maternal availability, cope with sepa-
tempted to explain childhood psychopathological rations, and explore the environment. What Bow-
disorders, reactions to separation from parents, lby (1969/1982) termed the "goal-corrected part-
and mourning as examples of behaviors designed nership" emerges. The toddler is now able to
to increase proximity to the caregiver, thereby engage in communication with the mother about
enhancing survival. As the child developed, the her availability and can negotiate shared plans
easily observed behavioral aspects of attachment about the maintenance of security in her absence
were enhanced and supplanted by the emotional (e.g., "Mommy will be in the kitchen. You play
and cognitive aspects that subserved the same here and get me if you need me.") Quality of
adaptive function of increasing access to the care- attachment in the preschool years should, there-
giver and hence survival. fore, include aspects of child reciprocity, perspec-
Drawing on control systems theory, Bowlby tive-taking, management of relationships, and em-
posited an attachment behavioral system that pathy (Crittenden, 1992, p. 211).
could be activated when access to the caregiver Also accompanying the toddler's increasing cog-
was threatened. This system coexisted in recipro- nitive and linguistic sophistication is Bowlby's no-
cal interplay with the exploratory system, which tion of an "internal working model." In this model
accomplished exploration of the environment and the history of the child's relationship to caregivers,
the development of autonomy and competence. particularly information about the caregiver's
Only when the child was confident in his or her availability and quality of responsiveness, and
access to the caregiver did the child feel confident complementary aspects of the child's self are cog-
to engage in exploration. Ainsworth (Ainsworth, nitively represented. Such a model would guide
Blehar, Waters, & Wall, 1978) called this the se- expectations and attitudes about current and fu-
cure base phenomenon, and reformulated the goal ture relationships and organize an individual's be-
of attachment behavior as both proximity to the havior in all emotionally significant relationships.
caregiver and emotional security in the availability Although they are updated by ongoing experience,
of the caregiver. Thus, three criteria define a rela- working models become resistant to change be-
tionship as fulfilling the function of attachment: yond early childhood because they develop pre-
(1) the need for ready access to the attachment verbally, are affectively laden, and tend to be rein-
figure—"accessibility"; (2) the desire for the prox- forced by ongoing experience. Attachment,
imity to the attachment figure in times of stress— therefore, is one means of conceptualizing and
"proximity seeking under conditions of threat"; describing a basic intrapsychic organization that
and (3) diminished anxiety in the company of the is a "tolerably accurate reflection" (Bowlby, 1973)
attachment figure and increased anxiety when the of actual relationship experiences. This view of
attachment figure is unavailable—"secure base internal working models as affectively charged
and felt security" (Weiss, 1982). representations of self and important others is very
During the first year of life, the infant's attach- close to psychoanalytic formulations of self and
ment behaviors become organized around main- object representations.
taining access to a specific, preferred figure, the Bowlby's models were based on actual experi-
primary caregiver (typically the mother). The goal ence. He saw that the competence of the caregiver
of attachment is achieved when the infant is suffi- directly affects the quality of caregiver-child rela-
ciently near the attachment figure and attachment tions. Children whose caregivers are sensitive, at-
behaviors such as crying or clinging are not ex- tuned, and accepting expect to continue to be
pressed. The infant is then free to explore the treated in this way. In contrast, inconsistency or
environment, and through this exploration, cogni- insensitivity in parental responsiveness encour-
12 / Attachment in Adolescence

ages children to develop defenses (in Bowlby's from the cradle to the grave, are happiest when
terminology, "defensive biases") that allow the life is organized as a series of excursions, long
children to cope with the painful affects they expe- or short, from the secure base provided by our
rience and simultaneously to maintain access to attachment figure(s)" (1988, p. 62).
the caregiver. Bowlby postulated that such defen- Weiss (1982) echoes much of Bowlby's theoriz-
sively biased models were even more enduring ing regarding the role of attachment in adoles-
than models based on security and confidence cence and beyond. The three basic criteria for
in the caregiver. Thus, defensively biased models identifying a relationship as an attachment rela-
formed the initial stages of defensive structures, tionship—accessibility, proximity seeking, and a
which would ultimately lead to distortions in per- secure base—can be found in "relationships of
sonality and psychopathology. central emotional importance" (p. 174) through-
out the life span. Several important distinctions
exist between childhood and adult attachments,
however. The first concerns the asymmetry inher-
Despite the focus in developmental psychology ent in childhood attachment relationships, with
on the nature and function of attachment in infants the attachment figure viewed not only as a safe
and toddlers, Bowlby viewed attachment as devel- haven but as a source of other forms of caregiving.
oping throughout the life span. Indeed, one of his Another aspect of this asymmetry is that the at-
most important contributions was to point out that tachment figure may be imbued with powers be-
an increased need and desire for closeness in times yond the range of the child, such as greater compe-
of stress is normal rather than a sign of maladap- tence, wisdom, or strength. In adolescence, one
tive dependency. However, how these attachment of the central transformations revolves around the
functions are played out beyond early childhood deidealization of the attachment figure (Bios,
was less well studied. While Bowlby strongly af- 1967) and hence the more objective recognition of
firmed the existence of attachment, especially the that figure's abilities. With the parent-adolescent
secure base phenomenon, in adolescence and relationship put on a more equal footing, attach-
adulthood, he did not consider that it changed in ment figures come to be used to foster the adoles-
quality. He explained the fact that older children cent's own capacity to master challenges, as allies.
and adolescents tolerated longer and more distant Weiss (1982) describes parents of adolescents as
separations from attachment figures on the basis "attachment figures in reserve" (p. 176).
of their advancing cognitive and ego development, Increased adolescent autonomy is implied in
which in turn permitted greater intemalization of this now more equal relationship with parents,
self-regulatory abilities previously supplied by the making the adolescent less susceptible to anxiety-
attachment figure and allowed for substitution of triggered activation of the attachment system.
symbolic methods of proximity seeking (e.g., Whereas in infancy and early childhood, lack of
phone calls, letters, photographs) for the physical physical access to the attachment figure when de-
closeness required by the young child. Changes sired is highly affectively arousing or even disorga-
in the hierarchy of attachment figures also occur nizing, adolescents can utilize more symbolic ac-
beginning in adolescence (Bowlby, 1969/1982; cess to attachment figures and do not experience
1988). Romantic partners replace parents as pri- major disruptions when access is denied. In con-
mary attachment figures, and attachment is di- trast to younger children, adolescents seek out
rected toward groups and institutions. Adolescents longer intervals away from parental supervision or
who have had inadequate attachments to their direct contact with parents. Weiss (1982) views
parents may form relationships with other, more the process of relinquishment of the parents as
responsive adults, to replace these deficient rela- attachment figures not as a gradual one but as one
tionships. Although the adolescent may be able in which the ongoing relationship is punctuated
to tolerate, or more correctly seeks and welcomes, by more frequent and longer separations with
increased physical and psychological distance briefer resurgences of attachment seeking. An
from early attachment figures, he or she is able attachment/autonomy dialectic proceeds; during
to do so only with the knowledge that a secure these separations, adolescents test out their own
base awaits a call. In Bowlby's words, "All of us, competence, seek out new attachment figures, and


increasingly come to rely on their own ability to ministrations, such as holding or reassurance.
function autonomously. The new attachment fig- These infants have a working model of their
ures are usually peers for younger adolescents and mother as "sensitively responsive" to their needs,
romantic partners for older adolescents. Attach- which refers to the caregiver's capacity to under-
ment is not forsaken, since parents are "in re- stand the needs of the infant accurately and re-
serve," and new attachment relationships take on spond to those needs. Securely attached infants
the same properties as the original ones with expect that their mothers will see their distress
parents. and also will respond to their signals about how
to end that distress.
Secure toddlers and school-age children are not
distressed by brief separations from parents and
Mary Ainsworth, an American developmental affectionately and confidently initiate conversa-
psychologist and colleague of Bowlby's, set out to tion or contact with dieir parents on reunion.
provide "a normative account of the development Their conversations are distinguished by fluidity
of attachment during the first year of life . . . and and a sense of ease. When asked to fantasize about
. . . an examination of individual differences in separations from parents, their descriptions show
the qualitative nature of the attachment" (Ains- an openness to the topic, an ease with both positive
worth, 1989, p. 709). From home observations and negative aspects of the situation, an ability to
of mother-infant pairs and a laboratory paradigm become absorbed in other activities during the
called the Strange Situation (Ainsworth & Wittig, separation, and the anticipation of a happy re-
1969), in which the separation and reunion behav- union.
ior of infants with their mothers is observed, three Secure (autonomous) adolescents and adults
patterns of attachment were identified, one secure value attachment relationships and regard attach-
and two insecure (avoidant and ambivalent) (Ains- ment-related experiences as pleasurable. They are
worth, et al., 1978). Although her goal was only relatively independent and objective regarding
to identify patterns in infancy, the same basic pat- any particular experience or relationship. Their
terns have been utilized to recognize behavioral view of parents is coherent and consistent. Typi-
patterns in older children, adolescents, and adults cally, in discussing their relationships with their
that are the developmentally reorganized and parents, they are fluent and relaxed. They seem
transformed enactments of working models. Sub- at ease with the topic of relationships with parents
sequent research has described a fourth, insecure and therefore able to reflect objectively on experi-
(disorganized) pattern (Main & Solomon, 1986, ences with them.
1990). The same four patterns of attachment have Avoidant /Dismissing Attachment: A second
been found in infants, toddlers (Crittenden, 1992), group of infants (20 to 25% of middle-class sam-
school-age children (Main, Kaplan, & Cassidy, ples) is characterized as having insecure-avoidant
1985), adolescents (Kobak & Sceery, 1988; Ro- attachments. These infants rarely show overt dis-
senstein & Horowitz, 1993), and adults (Main & tress over separation. On reunion they are indiffer-
Goldwyn, 1985-1991). Each attachment pattern ent to their mother's return and literally avoid
corresponds to a working model that guides be- interaction with her, usually turning their atten-
havior on the part of the child as well as the tion to toys, exploration of the room, or the com-
mother. The insecure models involve defensive pany of the stranger. As with the other types of
biases leading to characteristic defensive behav- attachment, this avoidant behavior develops in the
iors. The attachment types and their behavioral context of a particular land of mother-infant rela-
and defensive correlates are described next. tionship. These infants have been emotionally and
Secure/Autonomous Attachment: Most infants physically rejected, particularly in times of distress
(65-75%) in American middle-class samples are (Ainsworth et. al., 1978). The avoidant infant has
secure. Secure infants acknowledge maternal ab- created an adaptative strategy to deal with rejec-
sence on separation, either through protest or di- tion and to regulate his or her affective response
minishment in the quality of play, but greet the to this rejection. The mother's rejection of the
mother with pleasure on reunion. If distressed, infant's attachment behavior leads to anger and
the infant is calmed by the mother's return or anxiety within the infant. Avoidance is a compro-

12 / Attachment in Adolescence

mise between alienating the attachment figure for physical contact with their mothers with angry
with angry demands for attention and maintaining rebuffs of their mothers. They fail to be soothed
an intolerably angry, rejected position in the ab- by their mothers' attentions and therefore cannot
sence of a comforting attachment figure. Avoid- use their mothers as a secure base. These infants
ance serves to cut off the affective arousal before have experienced inconsistent sensitivity from
it is experienced as distress. This allows the infant their mothers (Ainsworth et al., 1978), leading
to regulate affect, thereby maintaining some orga- them to become uncertain of the mother's avail-
nization during times of distress, and to maintain ability. In order to elicit and maintain the mother's
some comforting physical proximity to the attach- attention, the infant intensifies attachment behav-
ment figure. iors and vigilance regarding her whereabouts,
Avoidant toddlers and children show the same leading to a pattern of exaggerated emotionality
pattern of indifference to relationships and height- and angry rebuff of maternal soothing and anxiety.
ened interest in the environment. Avoidant (dis- Ambivalent toddlers and preschoolers show in-
missing) adolescents and adults either dismiss the appropriate or exaggerated emotional expressions
importance of relationships or dismiss the extent and chronic low-level dependency on adults, and
of the impact of die relationship on the self. They are socially incompetent with peers. Ambivalent
idealize their parents or portray negative experi- (preoccupied) adolescents and adults appear
ences with their parents as normal. Frequently highly conflicted and in an ongoing cycle of fruit-
they lack memories of childhood; if negative mem- less reflection on relationships with parents and
ories are present, they regard themselves as unaf- disavowal of interest in maintaining these relation-
fected by them. Often they highly value achieve- ships. Identity diffusion is common. The "actively
ment, self-reliance, personal strength, or cunning. preoccupied" adolescent shows extreme anger to-
Sometimes they cite these qualities as rationaliza- ward parents and blames them for his or her prob-
tions for the lack of effect of negative experiences lems, together with placating attitudes toward the
on the self. The dismissing organization functions parents. "Passively preoccupied" adolescents os-
defensively to exclude from awareness any infor- cillate between positive and negative characteriza-
mation that may evoke attachment behaviors and tions of parents, or give descriptions of their par-
hence make a person vulnerable to being rebuffed, ents that are so vague that they are nearly
resulting in painful affects of anger and sadness. incomprehensible. Preoccupied individuals' bids
To make a bad situation worse, if these feelings for attachment are partially or inconsistently met,
are revealed, die parents are not able to respond and die behavior of parents of such adolescents
to them in a way that alleviates the adolescent's is such that it prevents disengagement from the
distress, nor does the adolescent have sufficient relationship. Patterns of hysteric, obsessional, and
self-regulatory capacity to acknowledge and toler- borderline defenses are therefore common in
ate negative affect. The result is that negative these adolescents.
affects can be neither displayed nor tolerated in Disorganized/Unresolved Attachment: Infants
order to achieve mastery over threatening or frus- in the fourth category, whose attachment has been
trating situations. Hence, denial, falsification of characterized as "disorganized," do not possess a
affective expression, and displacement of aggres- coherent and functional strategy for regulating
sion are used as the primary defenses of dismiss- their distress on separation and therefore engage
ing individuals. in behaviors that seem inexplicable or contradic-
Ambivalent/Preoccupied Attachment: The third tory in intent or function (Main & Solomon, 1990).
attachment classification, the insecure-ambivalent Behaviors seen in these infants include "ap-
attachment, is the rarest organization in infancy proaching with the head averted, stilling or freez-
(less than 10% of Ainsworth's middle-class sam- ing of movement with a dazed expression, walking
ple). It is characterized by cyclical efforts to gain backwards toward the mother, calm, contented
security from the attachment figure and avoidance play suddenly succeeded by distressed angry be-
of that figure. These infants are distressed by novel havior" (Main & Solomon, 1990, p. 122). In low-
environments and cling to their mothers. On sepa- risk (middle-class) samples, infant disorganization
ration they are excessively distressed and uncon- has been found to be associated with the mother's
trolled. On reunion these infants alternate bids unresolved mourning for her own attachment fig-


ure (Ainsworth & Eichberg, 1991). According to create emotional and physical distance from the
Main and Hesse (1990), the mother's lack of reso- family of origin. Distancing was viewed as a neces-
lution of mourning leads to frightened and fright- sary component of the separation-individuation
ening behavior by the mother with her own infant. task for the adolescent, in the service of enhanced
The infant is placed in the intolerable position of individuation. These processes have been held to
having the source of his or her distress and haven result ultimately in the adolescent's self-reliant
from that distress embodied in the same person— adaptation, with internalized affective and cogni-
the attachment figure. In high-risk samples (im- tive controls, and a replacement of relationships
poverished individuals; individuals with abusive, with parents with peer and romantic relationships.
neglectful, or psychopathological attachment fig- The end result of the separation-individuation
ures), infant disorganization is also strongly related process is a cohesive identity and sense of self,
to maternal lack of resolution of trauma, typically allowing for independent functioning. This view
parental history of childhood abuse or neglect has long-standing roots in the psychoanalytic liter-
(Carlson, Cicchetti, Barnett, & Braunwald, 1989). ature (A. Freud, 1958), with the most recent and
Disorganized (unresolved) attachment in ado- influential voice articulated by Bios (1967). Ac-
lescents and adults results from unresolved re- cording to Bios, individuation results from the
sponses to trauma in childhood (typically physical adolescent's relinquishment of dependence on in-
or sexual abuse) or from unresolved mourning of fantile parental introjects. Separation, defined as
an attachment figure lost when the individual was a distinct sense of self as a separate person from
young. In both cases, unresolved individuals can- the parents, then occurs. The adolescent thereby
not move beyond the events in question to form becomes capable of taking over functions that
an abstract understanding of their effect. They were previously performed by parental ego sup-
continue to experience disorganization and disori- ports and parental introjects. Bios describes this
entation, as manifest in irrational thought pro- process as the "second separation-individuation,"
cesses surrounding the traumatic or loss event, likening it to the separation-individuation process
unfounded fear, unfounded guilt, and continuing that occurs in toddlerhood (Mahler, Pine, &
disbelief that the traumatic events occurred. Bergman, 1975), through which a sense of self,
While this category frequently is viewed as a sepa- as distinct from mother, is first established. The
rate attachment organization, it can occur in ado- rekindled need for separation-individuation in ad-
lescents or adults who had developed any of the olescence is prompted by puberty and the urge
other types of attachment organizations prior to toward sexual object choice outside of the family
the trauma or loss. In such persons, both the su- this entails. Hence failures in the separation-indi-
perimposed unresolved attachment and the pre- viduation process can lead to difficulty in ac-
existing attachment type should be discernible. cepting sexuality and in the development of in-
Preliminary studies have shown that unresolved timacy.
attachment is associated with dissociative experi- With the growing attempt to understand the
ences in adults (Main, van IJzendoorn, & Hesse, role of attachment beyond its origins in infancy
1993). and toddlerhood, the prevailing view of adoles-
cence as a progressive movement away from emo-
tional reliance on parents and family and toward
complete self-reliance has come into question. Ac-
Attachment and the Tasks cording to attachment theory, the attachment sys-
tem is active throughout the life span, although
of Adolescence reorganized in its behavioral expression to meet
the needs of the individual at every point in devel-
opment (Sroufe & Waters, 1977). The first separa-
tion-individuation process in toddlerhood does not
dissolve the child's attachment relationship when
Until recently, the dominant view of an adoles- a sense of self is established. Rather, attachment
cent's relationship with parents emphasized the is reorganized in the context of developmental
adolescent's developmentally appropriate need to transformations. Attachment behaviors in tod-

12 / Attachment in Adolescence

dlerhood continue to be organized around mainte- Beyond the secure base functions provided by
nance of security, although incorporating the tod- parents of adolescents, parental encouragement
dler's capacity for understanding the goals and of appropriate levels of autonomy also is associated
intentions of the attachment figure in order to with adolescent attachment security. For parents
negotiate the attachment figure's availability. Sim- to encourage autonomy and at the same time pro-
ilarly, attachment is maintained in adolescence, vide support, they must intuitively understand the
albeit with transformed behavioral expressions needs of their adolescent children. This empathic
and transformations in the contexts in which it sensitivity is analogous to that needed for infants
occurs. That is, adolescents maintain attachment and toddlers. Just as maternal sensitivity to her
relationships with parents, although attachment infant predicts security of the infant's attachment,
behaviors may be emitted less frequently, with so, too, does parental sensitivity to the adolescent's
less intensity and urgency, and increasingly come developmental needs predict attachment security.
under the control of higher-order cognitive pro- Not only are attachment behaviors in adolescence
cesses such as abstraction, formal reasoning, and redefined in light of the organizational perspec-
metacognitive monitoring. tive, but so are reciprocal parental behaviors. Sen-
Just as attachment functions early in life to pro- sitive parents of infants and toddlers maintain
vide security to the vulnerable child, so it functions close contact and provide room for exploration.
in adolescence to provide security under instances Sensitive parents of adolescents maintain them-
of heightened vulnerability, fear, and stress. Just selves as available for emotional contact while ac-
as in the preschool years, more and more self- tively promoting and supporting the adolescent's
regulatory functions are assumed by the child, appropriate efforts at self-reliance, self-gover-
secure in the knowledge of parental availability if nance, individual identity, and autonomy. Ryan
needed, so in adolescence, more and more self- and Lynch (1989) underscore this view:
regulatory functions typical of adulthood are as-
sumed by the adolescent, in the context of parental We suggest that individuation during adolescence and
availability if needed. Rice (1990) cites the contin- into young adulthood is facilitated not by detachment
gent necessity of parents and contingent aware- . . . but rather by attachment where the latter is appro-
ness of attachment longings as one of the reasons priately conceptualized. . . [as] those relationships that
are optimal and appropriate to a person's developmental
that links between attachment and adolescent ad- level, thereby enabling the person to function with a
justment have been so variable: degree of coherence and integrity that would not other-
wise be possible. In adolescence and young adulthood,
[the link] waxes and wanes during one's development. gratifying relationships with parents would involve emo-
It may be that a stronger association between attach- tional closeness and a sense of support within a context
ment and adjustment occurs prior to important develop- of encouragement for one's efforts at individuation and
mental transitions (e.g., admission to college, graduation autonomy. . . . Indeed, individuation is not something
from college). Once the transition is made, the adoles- that happens from parents but rather with them. (Pp.
cent may rely on other sources to help him or her 340-341)
adjust (e.g., peer groups, counselors). Since multiple
attachment figures are possible, the adolescent may at
certain times and in certain situations utilize attachment ATTACHMENT AND PEER GROUP RELATIONS
relationships with someone other than parents. (Pp.
534-535) Early adolescence is a time during which the
peer group takes on prominence and in which
Parents continue to serve as a "secure base" susceptibility to peer influence is at its height.
for their adolescent children. They serve both as Traditional accounts have interpreted the peer
proximal sources of aid in times of need or distress group as supplanting parental influence and pro-
and as distal sources of potential aid. The assur- viding support for adolescent independence and
ance of their availability frees the adolescent to individuation in the service of the first steps to-
utilize his or her capacities to explore the environ- ward separation from parents. The peer group
ment. Hence, as the attachment behavioral system further functions as a normative reference group
is deactivated, the exploratory behavioral system in which adolescents come to evaluate their own
can come in to play. behavior, increase social skills, and consolidate


their identity. Bios (1967) views these extrafamilial romantic partner when a committed relationship
influences as essential in the consolidation of iden- is established.
tity, a process that cannot take place solely within The sequence of this developmental progres-
the confines of the family. Steinberg and Sil- sion can be altered by the quality of the adoles-
verberg (1986) take an essentially consistent view, cent's relationship with parents. Varying types of
that normative development consists in "a trading insecure attachment can lead to individual differ-
of dependency on parents for dependency on ences in the desire for and adequacy of parents
peers rather than straightforward and uni-dimen- as attachment figures, and this in turn affects rela-
sional growth in autonomy" (p. 848). Autonomy tionships with peers. Late adolescents classified
is then reserved for a later stage in adolescent in traditional attachment groups (i.e., secure, dis-
development, when the support of the peer group missing, preoccupied) had parents who remained
is no longer vital. prominent as attachment figures when compared
Ryan and Lynch (1989) have questioned this with peers and romantic partners (Smith &
belief in the relative functions of parents and peers George, 1993). This was especially true for the
in the lives of early adolescents. They cite data secure and preoccupied groups, the latter group
that show that those adolescents who are most being characterized by its enmeshment with par-
reliant on peer support are doing so in a compen- ents. Dismissives were relatively more reliant on
satory manner to make up for feelings of rejection peers than were the other two groups. The unre-
and detachment from parents. Therefore, those solved adolescents did not rely on any individuals
adolescents who trade emotional support on par- for attachment functions; their isolation from inti-
ents for emotional support on peers are insecurely mate contact was a striking feature of their rela-
attached to parents. Adolescents who are secure tional lives.
with parents are also secure with peers. Armsden
and Greenberg (1987) similarly showed the more
predominant effect of parental attachment secu-
rity on adolescent adjustment over and above the
Empirical Studies of
effect produced by peer attachment security, Adolescent Attachment
highlighting the causal role of parental attachment
quality in producing adequate peer relationships.
But how does attachment theory account for
the increasing interest in and importance of peers
in adolescence? Attachment provides a model that A large body of empirical work has linked the
underscores the continued importance of parents quality of adolescent relationships to their overall
as attachment figures in adolescence, both in the emotional adjustment. (See Armsden and
ongoing influence on adolescents and in the effect Greenberg, 1987, and Rice, 1990, for reviews of
of internalized representations of parents. A time- this literature.) A much smaller number of investi-
table for the relinquishment of the immediate as- gators have attempted to narrow the construct of
pects of parental influence is suggested, as peers "relationships" to Bowlby's ideas of attachment
and ultimately romantic partners supplant parents relationships. In keeping with the perspective that
on the hierarchy of attachment figures. Hazan views attachment as an organizational construct
(1991) has most closely studied this issue and pos- whose expressions vary both behaviorally and in-
its a developmental sequence in which four basic trapsychically with development, adolescent at-
attachment functions—proximity seeking, separa- tachment researchers have attempted to delineate
tion protest, safe haven, and secure base—shift the central functions of attachment in adoles-
in object from preadolescence to early adulthood. cence. Most common, and in keeping with find-
In preadolescence, all four functions are directed ings of non-attachment-based studies linking rela-
toward parents. Proximity seeking is the first func- tionships and emotional adjustment, are studies
tion to shift to peers in early adolescence. By late that explore the relative contributions of peer and
adolescence, only the secure base function is re- parent attachment to adolescent social and emo-
tained by parents, and this, too, will shift to a tional development, especially social competence,

12 / Attachment in Adolescence

self-esteem, identity formation, and emotional ad- have allowed greater predictive power from their
justment. Investigators have hoped to show either prior attachment status.
prospectively or concurrendy that adolescent ad- In a meta-analytic review, Rice (1990) con-
justment is facilitated by positive attachment to cluded, however, that the association among ado-
parents, in which parental availability in times of lescent attachment security, social/emotional ad-
stress, parental support, and warmth are most justment and identity formation did indeed hold:
highly associated with adolescent social and emo- "There appeared to be a consistent positive associ-
tional adjustment. Most studies show that a secure ation between attachment and measures of social
relationship to parents continues to have a positive competence, self-esteem, identity, and emotional
effect on adjustment well into adolescence (Arms- adjustment. Negligible correlations emerged be-
den & Greenberg, 1987; Papini, Roggman, & tween attachment and measures of college adjust-
Anderson, 1991; Ryan & Lynch, 1989) and that ment" (p. 534). He offers two speculations about
the quality of the adolescent's relationship with inconsistencies in results. The first is based on
parents is a more important determinant than peer differences in measurement instruments. No
relationships on the adolescent's adjustment. studies used multiple measures of attachment or
These studies also show that secure adolescents identity, which would have allowed for convergent
not only feel accepted by parents, but seek out and discriminant validation of the attachment and
involvement with them. identity measures. Using more assessment tools
While the studies just reviewed agree on the to measure each of these constructs may have
association among attachment, social competence, revealed more definitively that certain aspects of
and interpersonal functioning, Rice (1990), in a identity are related to certain aspects of attach-
review of the correlative adaptations of adolescent ment. The second speculation involves time of
attachment, cites several inconsistent findings, assessment. Rice posits that
particularly with regard to identity formation
(Lapsley, Rice, & FitzGerald, 1990; Quintana & the association between attachment and indices of ad-
Lapsley, 1987). Since identity integration was justment waxes and wanes during one's development.
viewed by Erikson (1968) as the essential task It may be that a stronger association between attach-
ment and adjustment occurs prior to important develop-
of adolescent development, failure to find a link
mental transitions (e.g., admission to college, graduation
between attachment and identity development from college). Once the transition is made, the adoles-
undercuts the premise of attachment as basic to cent may rely on other sources to help him or her adjust
all processes of social and emotional development. (e.g., peer groups, counselors). (Pp. 534-535)
Nevertheless, some aspects of attachment have
been found to be related to identity. Lapsley et al. While the studies just reviewed allow for greater
(1990) did show that communication with parents specificity of prediction by using constructs from
(one aspect of attachment) was correlated with attachment theory, they all allow predictions only
personal and social identity. Kroger and colleagues regarding the secure/insecure dimension of at-
(Kroger, 1985; Kroger & Haslett, 1988) related tachment. None of these studies allows for a de-
attachment style to identity status, defined ac- scription of the emotional and interpersonal diffi-
cording to a developmental model, including an culties specific to each insecure attachment type.
identity crisis leading either to foreclosure, com- Several other studies have remedied this defi-
mitment, moratorium, or diffusion. While se- ciency. Resnick (1991) reported a correspondence
curely attached adolescents were the most likely between security and self-esteem, and additionally
to be committed and those who were insecurely demonstrated that individuals classified as preoc-
attached most likely to be foreclosed, no predic- cupied showed the lowest rating of all groups on
tions could be made about the other two identity attachment to peers. Jenkins and Fisher (1989)
statuses. Because of the cross-sectional nature of showed that avoidant individuals reported poor
the studies, it is possible that moratorium or dif- relations with both parents and viewed their moth-
fused individuals were simply delayed in acquiring ers as unresponsive, troubled, and demanding,
commitment or foreclosure and that knowing how while ambivalent individuals reported poor rela-
they eventually resolved their identity crisis might tions with their mothers and positive ones with


their fathers, whom they viewed as understanding and high levels of social support from friends and
and sympathetic. Based on attachment classifica- family. Parents were viewed as loving and avail-
tion, Bartholomew and Horowitz (1991) made able. Kobak and Sceery (1988) conclude that lov-
predictions about aspects of late-adolescent self- ing and supportive experiences with parents di-
esteem (self-confidence, self-reliance, positive rectly lead to the ability to report childhood
versus negative view of self) and aspects of rela- experiences with clarity, coherence, and balance
tionships to others (level of involvement in rela- in the presentation of both positive and negative
tionships, emotional expressiveness in relation- evaluations. The group dismissing of attachment
ships, trust). Secures are positive about both self rated themselves as lonelier, with little support
and others, allowing them to maintain close rela- from family, but they had as little distress as the
tionships without loss of autonomy, to express in- securely attached group. Memory for childhood
terpersonal warmth, and to exhibit coherence and was poor, presumably protecting these individuals
thoughtfulness in discussing relationships. Dis- from the negative affect associated with self-
missives have high self-regard and low regard for acknowledged parental rejection and lack of love.
others, predisposing them to high self-confidence, Peers viewed dismissives as more hostile than any
independence, and self-reliance, but they lack in- attachment group and more anxious than secures.
terpersonal trust and closeness and have restricted Preoccupied adolescents rated themselves as
emotional expressiveness. Others view them as highly distressed, with poor dating skills, although
hostile. By contrast, preoccupied individuals have with high support from parents. They saw parents
a negative view of self and positive view of others. as loving but role-reversing, leaving them vulnera-
They are highly emotionally expressive and self- ble to incoherence and continuous efforts to gain
disclosing, highly dependent and caregiving, over- support from parents. Others viewed them as
involved in relationships, and incoherent in their more anxious than secure individuals. Taken to-
discussion of relationships. Others view them as gether, these results suggest deficits in the regula-
exploitable. The fourth category, unique to this tion of negative affect in insecure individuals and
system but paralleling some aspects of the infant differential strategies for the regulation of nega-
disorganized classification, is fearful. These indi- tive affects leading to differential deficits in self-
viduals have negative evaluations of both self and awareness, symptomatic expression, and social
others. They avoid closeness because of fear of competence. Secure individuals are accurate in
rejection, distrust others, and are personally inse- self-awareness and socially competent with little
cure. They have low levels of romantic involve- distress. Dismissing and preoccupied individuals
ment and no reliance on others yet lack self-con- distort self-awareness for defensive purposes. Dis-
fidence. They exhibit the most social problems, missives exclude from awareness information that
including lack of assertiveness and social inhi- would lead them to confront the anger generated
bition. by parental rejection. Preoccupied individuals
While this study represents a significant ad- cannot move beyond a cyclical state of intensive
vance in predicting the particular problems in so- distress, promoting efforts to gain parental sup-
cial relations and self-concept associated with each port that fail, thereby increasing distress, espe-
attachment type, it does not answer questions re- cially anxiety.
garding the etiology of the types or the dynamic
underpinnings that perpetuate dysfunctional rela-
tionships. Such deficiencies are remedied by an
exemplary study conducted by Kobak and Sceery Less well explored than the relation of attach-
(1988). Attachment was conceptualized as (among ment to social competence is the relation of family
other functions) a theory of affect regulation, functioning to attachment status. This line of re-
which incorporates the idea of defensive bias (Cas- search grows from a tradition in family theory
sidy & Kobak, 1988; Kobak, 1986). An under- relating patterns of family functioning to ego de-
standing of the styles of affect regulation led to velopment (Bell & Bell, 1982; Hauser et al., 1984),
predictions about representation of self and oth- identity formation and role-taking skill (Cooper,
ers. In a college (late-adolescent) sample, securely Grotevant & Condon, 1983), resistance to psycho-
attached adolescents self-reported little distress pathology (Wynne, Jones, & Al-Khayyal, 1982),

12 / Attachment in Adolescence

and academic and social competence (Baumrind, while preoccupied adolescents were equally likely
1991). Since each of these areas of adolescent to come from families rate high and low in respon-
development has been shown to be affected by sivity. These results underscore the psychological
the attachment status of the adolescent, it stands detachment inherent in the families of dismissing
to reason that adolescent attachment should be adolescents and the confounding of factors of em-
directly affected by the quality of family function- pathy and promotion of individuation in the fami-
ing. The results of three studies, to be reviewed, lies of preoccupied adolescents.
show this to be the case. All three studies use an Kobak and colleagues (Kobak, Cole, Ferenz-
analysis of observed family interactions as their Gillies, Fleming, & Gamble, 1993) took the per-
main source of data. This strategy connects two spective of the ongoing utilization and develop-
important strands in family research, what Reiss ment of emotion regulation to describe parent-
(1989) refers to as the individual's subjective or adolescent communication. In a problem-solving
"represented" family and the interactive or "prac- task designed to place parent and adolescent in
ticing" family. The represented family is the family moderate conflict (i.e., discussion of an area of
as the developing individual understands or cogni- disagreement in their relationship), communica-
tively represents it; this is the concept that has tion patterns between parent and adolescent were
been explored by attachment research. The term found to depend on attachment classification. Ad-
"practicing family" refers to the current organiza- olescents with secure attachment engaged in less
tion and interactional processes of the family. This dysfunctional anger and less avoidance of problem
aspect of the family has been the focus of both solving, and maintained balanced assertiveness
family therapy and parenting practices research. with their mothers. Adolescents with a dismissing
Reiss (1989) emphasizes that both the represented attachment style employed high levels of dysfunc-
and the practicing families are important regula- tional anger, and the interactions were char-
tory contexts of adolescent development that need acterized by higher levels of maternal dominance.
to be understood in conjunction with one another Kobak et al. (1993) interpret these findings as
but most often are studied in isolation. In the exemplifying the dysfunctional aspects of insecure
studies that follow, the characteristics of the repre- attachment on the attainment of joint goals even
sented family, defined as the adolescent's attach- into adolescence. In keeping with Bowlby's notion
ment organization, and of the practicing family, of the "goal-corrected partnership" in tod-
described by family interaction processes, were dlerhood (1969/1982), in which mother and child
found to be linked and to be associated with im- reciprocally set goals and plans and carry them
portant aspects of adolescent development. out in the service of increasing autonomy for the
In our own research, the represented family child, adolescents with secure attachments are
for each adolescent was assessed by the Adult able to balance individual and relational needs,
Attachment Interview. The practicing family was while those with insecure attachments are not.
assessed by observing families of psychiatrically Insecure attachment can compromise the adoles-
hospitalized adolescents using the Beavers Inter- cent's ability to either individuate from parents or
actions Scales, drawn from the Beavers-Tim- maintain close relations with them.
berlawn Model of family functioning. Factor anal- Drawing on similar constructs regarding the
ysis of the subscale scores yielded two distinct balance of autonomy and relatedness, Allen and
factors: responsivity and behavioral control. Re- Hauser (1993) examined continuities between ad-
sponsivity refers both to the degree of empathy olescent-family interactions and attachment orga-
and connectedness within the family as well as nization 11 years later, when the adolescents had
the intentional facilitation of individuation or psy- become young adults. Again using an adolescent-
chological autonomy. Parental control taps the de- parent interaction revolving around resolving a
gree to which parents set clear standards for be- disagreement, both normal and psychiatrically
havior and maintain those standards. Results hospitalized adolescents and their families were
showed an association between attachment and observed. Results showed that coherence in the
family responsivity (Reimer, Overton, Steidl, Ro- young adults' descriptions of their attachment re-
senstein, & Horowitz, 1996). Dismissing adoles- lationships were predictable from qualities of au-
cents came from families rated as less responsive tonomy and relatedness in their families' interac-


tions when they were adolescents. These disorders; overt disclosure of symptomatic dis-
predictions could be obtained both from adoles- tress; and avoidant, dependent, schizotypal, and
cents' behavior toward their parents and from par- dysthymic personality traits. Sex differences in
ents' behavior toward their adolescents. However, both diagnosis and attachment classification were
the one insecure attachment indicator in young found, with males more likely to be dismissing
adulthood that could be predicted from parent- and conduct disordered or substance abusing and
adolescent interaction (passivity of thought pro- females more likely to be preoccupied. We inter-
cesses) could be predicted only on the basis of preted these findings in the context of the emer-
adolescents' behavior toward parents and not vice gence of styles of regulating distress from working
versa. These adolescents were inhibited in the models, which evolve in the course of develop-
development of autonomy by personalizing dis- ment into styles of adaptation and defense. Ulti-
agreements with parents or by failing to recant mately these styles coalesce into personality traits
their positions. and symptomatology. Therefore, in adolescence,
distinctive personality traits and attachment orga-
nizations are well established. Symptomatology,
CLINICAL IMPLICATIONS personality, and attachment organization are inte-
OF ATTACHMENT CLASSIFICATION grally related, since diey result from the same
From die standpoint of psychopathology, links context in development.
between attachment insecurity and increased Adolescents whose attachment organization
symptomatology have been shown consistently for was dismissing relied on a strategy of defensive
adolescents. Insecure adolescents are more de- exclusion from awareness of information that por-
pressed, anxious, resentful, and alienated (Arms- trayed attachment relationships in a negative light.
den & Greenberg, 1987; Armsden, McCauley, This strategy is shared by adolescents with exter-
Greenberg, Burke, & Mitchell, 1990) and are nalizing disorders, that is, conduct and substance
more likely to engage in problem drinking abuse disorders, and by those with narcissistic and
(Hughes, Francis, & Power, 1989; Kwakman, antisocial personality disorders and traits. The per-
Zuiker, Schippers, & deWuffel, 1988) or drug nicious effect of this exclusion from awareness of
abuse (Allen, Hauser & Borman-Sporrell, 1996). negatively charged information is inferred from
The links between depression and preoccupied the individual's self-defeating behavior (e.g., ag-
attachment (Cole-Detke and Kobak, 1996; Kobak, gressiveness that is interpersonally alienating or
Sudler, & Gamble, 1991) and eating disorders runs afoul of the law, or substance abuse that
and dismissing attachment (Cole-Detke & Kobak, compromises health and academic functioning).
1996) have been demonstrated. Adam, Sheldon- The widely observed affective regulatory function
Keller, and West (1996) have demonstrated that of substance abuse can be viewed from an attach-
adolescents unresolved with respect to trauma or ment theoretic perspective as a means to cut off
loss were at greater risk for suicide attempts. Doz- awareness of negative affects surrounding attach-
ier (1990) showed that insecure attachment was ment, thereby maintaining an idealized view of
associated with treatment noncompliance in se- attachment figures and dismissing personal dis-
verely disturbed late adolescents. tress. Symptomatic behavior, such as substance
Rosenstein and Horowitz (1996) have examined abuse or the aggression seen in a conduct-disor-
the relationships among attachment classification, dered adolescent, thus may be seen as reflecting,
psychopathology, and personality traits in a group in a displaced manner, the anger generated in
of psychiatrically hospitalized adolescents. These the adolescent by ongoing parental rejection or
adolescents were overwhelmingly insecure (97%) intrusion, coupled with the adolescent's failure to
in their current state of mind with respect to at - acknowledge his or her own anger.
tachment. The adolescent group dismissing of at- By contrast, adolescents in the preoccupied
tachment was associated with conduct and sub- group were extremely sensitive to difficulties in
stance abuse disorders; denial of psychiatric their attachment relationships and overwhelmed
symptomatology; and narcissistic, antisocial, and by negative perceptions of parents. Preoccupied
histrionic personality traits. The group preoccu- adolescents were unduly sensitive to their own
pied by attachment was associated with affective distress and characteristically exaggerated their af-

12 / Attachment in Adolescence
feet in order to elicit comfort from the attachment to the study of attachment; the literature con-
figure, in a manner that blocked autonomy. Thus verges on the view that security of attachment
they were likely to have an affective disorder and predicts more competent social and emotional
personality traits that emphasized exaggerated functioning throughout childhood and adoles-
emotionality and overt conflict, such as histrionic, cence. Insecure attachment forms a major risk
borderline, or dependent traits. factor in the development of impoverished or con-
flictual relationships, negative mood states, and
psychopathology. Empirically, the relation of each
type of attachment insecurity to specific symptom-
Conclusion atic patterns in childhood has been inconsistent.
Not until adolescence can the latent effect of at-
tachment insecurity, coupled with the renewed
A knowledge of attachment theory is important press for the adolescent to separate from parents,
to the understanding of adolescent development. specify the pattern of psychopathology and char-
There is a significant body of knowledge related acter.
Adam, K. S., Sheldon-Keller, A. E., & West, M. (1996). peer attachment in early adolescent depression. Jour-
Attachment organization and history of suicidal be- nal of Abnormal Child Psychology, 18, 683-697.
havior in clinical adolescents. Journal of Consulting Bartholomew, K., & Horowitz, L. M. (1991). Attach-
and Clinical Psychology, 64, 264-272. ment styles among young adults: A test of a four-
Ainsworth, M. D. S. (1989). Attachments beyond in- category model. Journal of Personality and Social
fancy. American Psychologist, 44, 709-716. Psychology, 61, 226-244.
Ainsworth, M. D. S., Blehar, M. C, Waters, E., & Wall, Baumrind, D. (1991). Parenting styles and adolescent
S. (1978). Patterns of attachment: A psychological development. In R. Lerner, A. C. Peterson, & J.
study of the Strange Situation. Hillsdale, NJ: Law- Brooks-Gunn (Eds.), The encyclopedia of adolescence
rence Erlbaum. (pp. 746-758). New York: Garland Press.
Ainsworth, M. D. S., & Eichberg, C. G. (1991). Effects Bell, D. C., & Bell, L. G. (1982). Family climate and
on infant-mother attachment of mother's unresolved the role of the female adolescent: Determinants of
loss of an attachment figure or other traumatic experi- adolescent functioning. Family Relations, 31,
ence. In P. Marris, J. Stevenson-Hinde, & C. Parkes 519-527.
(Eds.), Attachment across the life cycle, (pp. 160- Bios, P. (1967). The second individuation process of
183). New York: Routledge. adolescence. Psychoanalytic Study of the Child, 22,
Ainsworth, M. D. S., & Wittig, B. A. (1969). Attachment 162-186.
and the exploratory behavior of one-year olds in a Bowlby, J. (1958). The nature of the child's tie to his
strange situation. In B. M. Foss (Ed.), Determinants mother. International Journal of Psycho-analysis,
of infant behavior (Vol. 4, pp. 113-136). London: 39, 350-373.
Methuen. Bowlby, J. (1973). Attachment and loss: Vol. 2. Separa-
Allen, J. P., Hauser, S. T., & Borman-Spurrell, E. tion, anxiety and anger. New York: Basic Books.
(1996). Attachment theory as a framework for under- Bowlby, J. (1980). Attachment and loss: Vol. 3. Loss.
standing sequelae of severe adolescent psychopathol- New York: Basic Books.
ogy: An 11-year follow-up study. Journal of Con- Bowlby, J. (1982). Attachment and loss: Vol. 1. Attach-
sulting and Clinical Psychology, 64, 254-263. ment (2nd ed.). New York: Basic Books. (Originally
Allen, J. P., & Hauser, S. T. (1996). Autonomy and published 1969.)
relatedness in adolescent-family interactions as pre- Bowlby, J. (1988). A secure base. New York: Basic
dictors of young adults' states of mind regarding at- Books.
tachment. Manuscript submitted for publication. Carlson, V., Cicchetti, D., Bamett, D., & Braunwald,
Armsden, G. C., & Greenberg, M. T. (1987). The Inven- K. (1989). Disorganized/disoriented attachment rela-
tory of Parent and Peer Attachment: Individual dif- tionships in maltreated infants. Developmental Psy-
ferences and their relationship to psychological well- chology, 25, 525-531.
being in adolescence. Journal of Youth and Adoles- Cassidy, J., & Kobak, R. R. (1988). Avoidance and its
cence, 16, 427-454. relation to other defensive processes. In J. Belsky & T.
Armsden, G. C, McCauley, E., Greenberg, M. T., Nezworski (Eds.), Clinical implications of attachment
Burke, P. M., & Mitchell, J. R. (1990). Parent and (pp. 300-323). Hillsdale, NJ: Lawrence Erlbaum.