Content uploaded by Mario Mikulincer
Author content
All content in this area was uploaded by Mario Mikulincer on May 19, 2016
Content may be subject to copyright.
THE ATTACHMENT BEHAVIORAL
SYSTEM IN ADULTHOOD:
ACTIVATION, PSYCHODYNAMICS,
AND INTERPERSONAL PROCESSES
Mario Mikulincer
Phillip R. Shaver
My life’s work has been directed to a single aim. I have observed
the more subtle disturbances of mental function in healthy and
sick people and have sought to infer—or if, you prefer it, to
guess—from signs of this kind how the apparatus which serves
these functions is constructed and what concurrent and mutually
opposing forces are at work in it.
—Sigmund Freud (1961/1930, p. 208)
As my study of theory progressed it was gradually borne in upon
me that the field I had set out to plough so lightheartedly was no
less than the one that Freud had started tilling sixty years earlier,
and that it contained all those same rocky excrescences and
thorny entanglements that he had encountered and grappled
with—love and hate, anxiety and defense, attachment and loss.
—John Bowlby (1982, p. xxvii)
Psychology has changed its focus several times during its relatively short
history—psychoanalysis, behaviorism, cognitivism, and neuroscience being
prominent examples. Within such conceptual traditions, an emphasis on the
individual and the individual’s selfish motives often gives way, over time, to
an emphasis on social interrelatedness. Freud’s psychoanalytic theory of
drives or instincts, for example, evolved into later theories of ‘‘object
relations’’ (a misleading name for social relations) and ‘‘intersubjectivity.’’
Behaviorism focused first on the individual’s drives and the role of drive-
reduction in learning, but then gradually made room for social learning. The
‘‘cognitive revolution,’’ aimed originally at such topics as perception
53
ADVANCES IN EXPERIMENTAL Copyright 2003 Elsevier Science (USA).
SOCIAL PSYCHOLOGY. VOL. 35 All rights reserved.
0065-2601/03 $35.00
of letters and memory for words, generated the field of social cognition that
deals with person perception and social memory. More recently, neurosci-
ence has bred cognitive neuroscience, which in turn is generating social
cognitive neuroscience.
These days, no one doubts that the human mind/brain is a social machine.
By evolutionary standards, the large human cerebral cortex (including the
very large frontal lobes and special modules for language) developed
surprisingly rapidly, primarily to serve increasingly complex interpersonal
functions (cooperative hunting and child-rearing, for example; identification
of reliable mates and allies as well as detection of cheaters and scoundrels).
Assuming, as we do, that major contributors to the different traditions in
psychology all had valuable insights and solid reasons for pursuing their
particular angles on human experience and behavior, we wish to combine
them to form a powerful, integrative theory of the social mind. In so doing,
we take as primary guides the British psychoanalyst John Bowlby and his
North American research collaborator Mary Ainsworth. These investigators
combined key insights of psychoanalysis, ethology, developmental psych-
ology, and cognitive psychology to create a theory of ‘‘attachment,’’ or
emotional bonding, that has been tested and elaborated in literally
thousands of studies over the past 30 years. Both Bowlby and Ainsworth
were unusually talented analytic thinkers and clear writers who respected
the scientific method. Bowlby was notable for the breadth and openness of
his mind, Ainsworth for her keen observational skills and methodological
creativity. Together they laid a foundation for studying close interpersonal
relationships across the lifespan.
Our work can be better understood if we explain briefly how we came,
individually, to attachment theory and then began to influence each other
and, eventually, work together. Both of us were attracted to psychoanalytic
theory as undergraduates, despite the hard knocks it has always taken from
diverse critics. Anyone who opens his mind to what goes on in real people’s
lives, and anyone who reads novels or poems or watches artful films, realizes
that the issues raised by psychoanalysts, beginning with Freud, are
important: sexual hunger and attraction; romantic love; the development
of personality, beginning in the context of infant–caregiver relationships;
negative emotions such as anxiety, sadness, jealousy, guilt, shame, and anger
and their role in intrapsychic conflict and psychopathology; intergroup
hostility and war.
When we first began studying experimental social psychology, it seemed
rather tame and superficial compared to psychoanalysis. But its strong
point, and the weak point of psychoanalysis, was a collection of powerful
and creative research methods. Psychoanalytic theorists seemed capable
of endlessly proliferating and debating hypothetical constructs and
54
MIKULINCER AND SHAVER
processes without being constrained by operational definitions, sound
psychometrics, or replicable research evidence. Both of us began our careers
as experimental researchers pursuing existing topic areas (stress and learned
helplessness in Mikulincer’s case, self-awareness theory and fear of success
in the case of Shaver), but our interest in psychoanalytic ideas never abated.
When Bowlby’s books began to appear, we realized it was possible for a
psychoanalytic thinker to amass empirical evidence for important psycho-
analytic propositions, and that psychoanalytic theory could be altered and
elaborated based on empirical research. Ainsworth’s development of a
laboratory ‘‘strange situation’’ assessment procedure, which allowed her to
classify infants’ attachment patterns and relate them systematically to home
observations of infant–parent interactions, added to our confidence that
research on an appropriately social and evolutionary version of psychoana-
lytic theory was possible.
In the mid-1980s, Shaver was studying adolescent and adult loneliness
(e.g., Rubenstein & Shaver, 1982; Shaver & Hazan, 1984) and noticing both
that attachment theory could be useful in understanding loneliness (e.g.,
Weiss, 1973) and that patterns of chronic loneliness were similar in certain
respects to the insecure infant attachment patterns identified by Ainsworth
and her colleagues (Ainsworth, Blehar, Waters, & Wall, 1978). Building on
this insight, one of Shaver’s doctoral students, Cindy Hazan, wrote a
seminar paper suggesting that attachment theory could be used as a broad
framework for studying romantic love or ‘‘romantic attachment,’’ as they
called it in their initial article on the topic (Hazan & Shaver, 1987).
That paper caught the eye of Mikulincer, who had become interested in
attachment theory while studying affect-regulation processes related to
learned helplessness, depression, combat stress reactions, and posttraumatic
stress reactions in Israel. He noticed similarities between (1) certain forms
of helplessness in adulthood and attachment-figure (e.g., parental) un-
availability in infancy; (2) intrusive images and anxiety in the case of
posttraumatic stress disorder and the anxious attachment pattern described
by Ainsworth et al. (1978) and Hazan and Shaver (1987); and (3) avoidant
strategies of coping with stress and the avoidant attachment pattern
described by these same authors. In 1990, Mikulincer, Florian, and
Tolmacz published a study of attachment patterns and conscious and
unconscious death anxiety, one of the first studies to use the preliminary
self-report measure of attachment style devised by Hazan and Shaver
(1987).
From that point on, both of us continued to pursue the application of
attachment theory to the study of adults’ emotions, emotion-regulation
strategies, and close relationships, noticing that we were both interested in
the experimental study of what might be called attachment-related
THE ATTACHMENT BEHAVIORAL SYSTEM 55
psychodynamics: the kinds of mental processes, including intense needs,
powerful emotions, and defensive strategies, that had captivated both Freud
and Bowlby. In recent years we have pooled our efforts to work out a more
carefully articulated statement of our theoretical ideas. In the present
chapter we summarize what we have learned to date, placing primary
emphasis on our own ideas and research (in line with the policy of the
Advances series) rather than attempting to provide a comprehensive review
of the entire adult attachment literature.
The chapter is organized as follows: We begin with an overview of basic
theoretical concepts, which will be essential for understanding the remainder
of the chapter. We then present an integrative model of attachment-system
dynamics in adulthood, a model that has both guided our research and been
affected by it. These two theoretical sections of the chapter constitute
roughly a third of its length, which is perhaps unusual in a chapter designed
to summarize research. In the present case, the research makes better and
deeper sense if the reader understands the theory. The subsequent sections
of the chapter summarize the diverse empirical studies motivated by the
theory and review a variety of unresolved conceptual and empirical issues
that provide an agenda for future research.
I. Basic Theoretical Concepts
We begin by introducing basic concepts in Bowlby’s (1973, 1980, 1982/
1969) attachment theory, briefly covering both the normative (general)
aspects of what he called the attachment behavioral system and individual
differences in what social psychologists call ‘‘attachment style’’ (Levy &
Davis, 1988). These styles, first noticed and delineated by (Ainsworth, 1967;
Ainsworth et al., 1978), are thought to emerge over the course of
development from countless interactions between the attachment system
and its interpersonal environment.
A. NORMATIVE COMPONENTS OF THE ATTACHMENT
BEHAVIORAL SYSTEM
1. What Is a Behavioral System?
Bowlby (1982/1969) defined a behavioral system as a species-universal
program that organizes an individual’s behavior in functional ways—i.e., in
ways that increase the likelihood of survival in the face of particular
56
MIKULINCER AND SHAVER
environmental demands. A behavioral system is an inborn, preset program
of the central nervous system that was ‘‘designed’’ by evolution via natural
selection. It governs the choice, activation, and termination of behavioral
sequences that produce a predictable and generally functional change in the
person–environment relationship. (The term ‘‘behavioral system’’ therefore
implies the involvement of actual behavior and behavior generation,
although it points theoretically to the central neural program we are
emphasizing.) Each behavioral system (e.g., attachment, caregiving, explor-
ation, affiliation) follows a particular predictable pattern of activation and
termination in almost all members of a species, a pattern that does not
depend on learning opportunities.
Conceptually, a behavioral system has six components or aspects: (1) a
specific biological function that increases the likelihood of an individual’s
survival and reproductive success; (2) a set of contextual activating triggers;
(3) a set of interchangeable, functionally equivalent behaviors that consti-
tutes the primary strategy of the system for attaining a particular goal state;
(4) a specific set-goal—the change in the person–environment relationship
that terminates the activation of the system; (5) the cognitive operations
involved in the functioning of the system; and (6) specific excitatory or
inhibitory neural links with other behavioral systems.
2. Biological Function of the Attachment System
According to Bowlby (1982/1969), the biological function of the
attachment behavioral system is to protect a person (especially during
infancy) from danger by ensuring that he or she maintains proximity to
caring and supportive others (attachment figures). In his view, natural
selection favored maintenance of proximity to attachment figures [what
Bowlby (1973) called ‘‘stronger and wiser’’ caregivers], because it increases
the likelihood of survival and eventual reproduction on the part of members
of a species born with immature capacities for defense from predators and
other dangers. Because infants require a long period of protection, they are
born with a tendency to seek proximity to others who can provide care and
support.
Although the biological function of the attachment behavioral system is
most critical during the early phases of life, Bowlby (1988) assumed the
system is active over the entire life span and is manifested in thoughts and
behaviors related to seeking proximity to attachment figures in times of
need. In his view, proximity seeking is a behavioral adaptation to
evolutionary pressures that can contribute to adjustment and health
throughout the lifespan. This function of the attachment system is most
relevant during stressful periods or traumatic experiences, wherein the
THE ATTACHMENT BEHAVIORAL SYSTEM 57
support and comfort given by attachment figures enhance coping and
adjustment.
During infancy, primary caregivers (usually one or both parents) are the
main attachment figures. In adulthood, however, a wide variety of
relationship partners can act as attachment figures, including parents,
friends, and romantic partners. Moreover, groups, institutions, and abstract
or symbolic figures (e.g., God) can become targets of proximity seeking
in times of need. There are also context-tailored attachment figures, who
are sources of support only in specific milieus: teachers and supervisors
in academic settings; therapists in therapeutic settings; and managers in
organizational settings. Attachment theory conceptualizes attachment
figures as forming a hierarchical network, with the primary attach-
ment figures being those with whom the individual maintains long-term
and strong affectional bonds (parents, close friends, spouses).
3. The Attachment System’s Activating Triggers
Originally, Bowlby (1982/1969) claimed that the attachment system is
activated by environmental threats that endanger a person’s survival.
Encounters with such threats create a need for protection from other people
and automatically activate the attachment system. When no threat is
present, there is no need to seek care from others and no proximity-seeking
tendency is activated, at least not for the purposes of protection. (A person
may seek proximity to others for the purpose of some other behavioral
system such as affiliation or sexual mating.) In subsequent writings, Bowlby
(1973) extended this reasoning and proposed that the attachment system is
also activated by ‘‘natural clues of danger’’—stimuli that are not inherently
dangerous but that increase the likelihood of danger (e.g., darkness, loud
noises, isolation)—as well as by attachment-related threats such as
impending or actual separation from, or loss of, attachment figures. In his
view, a combination of attachment-unrelated sources of threat and lack of
access to an attachment figure compounds distress and triggers the highest
level of attachment-system activation.
4. The Primary Attachment Strategy
According to Bowlby (1982/1969), proximity seeking is the natural and
primary strategy of the attachment behavioral system when a person has a
perceived need for protection or support. This strategy consists of a wide
variety of behaviors that have a similar meaning (the seeking of proximity)
and serve similar adaptive functions (protection from threats). Among these
behaviors, one can find signals (interaction bids) that tell a relationship
58
MIKULINCER AND SHAVER
partner an individual is interested in restoring or maintaining proximity;
overt displays of negative emotion (e.g., anger, anxiety, sadness) that drive
the relationship partner to provide support and comfort; active approach
behaviors that result in greater physical or psychological contact, including
what Harlow (1959) called ‘‘contact comfort’’; and explicit requests for
emotional or instrumental support. According to Bowlby (1982/1969), not
all of these behaviors are likely to be manifested in every threatening
situation. Rather, they are part of a repertoire of behaviors from which an
individual can ‘‘choose’’ (consciously or unconsciously) the most adequate
means for attaining protection in a given situation.
In adulthood, the primary attachment strategy does not necessarily lead
to actual proximity-seeking behavior. In fact, this strategy can be manifested
in the activation of mental representations of relationship partners who
regularly provide care and protection. These cognitions can create a sense of
safety and security, which helps a person deal successfully with threats. That
is, mental representations of attachment figures can become symbolic
sources of protection, and their activation can establish what might be called
symbolic proximity to supportive others. Of course there are times when
these strategies are insufficient and even adults seek actual proximity to
attachment figures.
5. Set-Goal of the Attachment System
Bowlby (1982/1969) viewed the attainment of actual or perceived
protection and security as the set-goal of the attachment system, which
normally terminates the system’s activation [see also Sroufe & Waters
(1977), who introduced the term ‘‘felt security’’]. Bowlby also enumerated
the provisions attachment figures should supply in order to facilitate the
attainment of this set-goal (see also Hazan & Shaver, 1994). First,
attachment figures should be responsive to the individual’s proximity-
seeking attempts in times of need. Second, these figures should provide a
physical and emotional safe haven—i.e., they should facilitate distress
alleviation and be a source of support and comfort. Third, attachment
figures should provide a secure base from which the individual can explore
and learn about the world and develop his or her own capacities and
personality while feeling confident that care and support will be available if
needed. When these provisions are supplied, a person feels secure and safe,
and proximity seeking for the purpose of protection and care is terminated.
6. Cognitive Substrate of the Attachment System
According to Bowlby (1982/1969), the attachment system operates in a
complex goal-corrected manner; that is, the individual evaluates the
THE ATTACHMENT BEHAVIORAL SYSTEM 59
progress his or her behaviors are making toward achieving the set-goal and
then corrects these behaviors to produce the most effective action sequence.
In our view, this flexible, goal-directed and goal-corrected adjustment of
attachment behavior requires at least three cognitive operations: (1)
processing of information about the person–environment relationship,
which involves the monitoring and appraisal of threatening events and one’s
own inner state (e.g., distress, security); (2) monitoring and appraisal of the
attachment figure’s responses to one’s proximity-seeking attempts; and (3)
monitoring and appraisal of the viability of the chosen behaviors in a given
context, so that an effective adjustment of these behaviors can be made, if
necessary, in accordance with contextual constraints.
Bowlby (1973, 1982/1969) also stressed that the goal-corrected nature of
attachment behavior requires the storage of relevant data in the form of
mental representations of person–environment transactions. He called these
representations working models and seemed to intend the word ‘‘working’’
to carry two senses: (1) the models allow for mental simulation and
prediction of likely outcomes of various attachment behaviors (that is, they
can provide dynamic, adjustable, context-sensitive representations of
complex social situations); and (2) the models are provisional (in the sense
of ‘‘working’’ drafts—changeable plans). Bowlby (1982/1969) distinguished
between two kinds or components of working models: ‘‘If an individual is to
draw up a plan to achieve a set-goal not only does he have some sort of
working model of his environment, but he must have also some working
knowledge of his own behavioral skills and potentialities’’ (p. 112). That is,
the attachment system, once it has been used repeatedly in relational
contexts, includes representations of attachment figures’ responses
(working models of others) as well as representations of the self ’s efficacy
and value (working models of self ). These working models organize a
person’s memory about an attachment figure and him- or herself during
attempts to gain protection in times of need (Main, Kaplan, & Cassidy,
1985).
7. Interplay between the Attachment Behavioral System and Other
Behavioral Systems
Encountering threats and experiencing disruptions in felt security activate
the attachment behavioral system, which in turn inhibits the activation of
other behavioral systems and prevents or hinders engagement in nonattach-
ment activities (Bowlby, 1982/1969). Under conditions of threat, people turn
to others as providers of support and comfort rather than as partners for
exploratory, affiliative, or sexual activities. Moreover, at such times they are
likely to be so self-focused (so focused on their need for protection) that they
60
MIKULINCER AND SHAVER
lack the mental resources necessary to attend empathically and altruistically
to others’ needs and engage in caring behavior. Only when relief is attained
and a sense of attachment security is restored can the individual deploy
attention and energy to other behavioral systems and engage in nonattach-
ment activities. Because of this reciprocal relation between the attachment
system and other behavioral systems, the attainment of attachment security
fosters engagement in nonattachment activities such as exploration, sex,
caregiving, and affiliation, and allows the individual to distance from
an attachment figure with the belief that this figure will be available if
needed.
8. Summary
According to Bowlby (1973, 1980, 1982/1969), the attachment system is
an inborn regulatory device that has important implications for personality
and interpersonal behavior. The system is activated by perceived threats and
dangers, which cause the threatened individual to seek proximity to
protective others. The attainment of proximity and protection results in
feelings of relief and security as well as positive mental representations of
relationship partners and the self. Bowlby (1988) viewed the optimal
functioning of this behavioral system as extremely important for the
maintenance of emotional stability and mental health, development of a
positive self-image, and formation of positive attitudes toward relationship
partners and close relationships in general. Moreover, because optimal
functioning of the attachment system facilitates relaxed and confident
engagement in nonattachment activities, it contributes to the broadening of
a person’s perspectives and skills as well as the actualization of his or her
unique potentialities. To borrow a term from another theoretical tradition,
humanistic psychology (e.g., Maslow, Rogers), attachment security is not
only an important prerequisite for healthy love but also a major stepping
stone to self-actualization.
B. INDIVIDUAL DIFFERENCES IN THE OPERATION OF THE
ATTACHMENT SYSTEM
1. The Role of Actual Interactions with Attachment Figures
According to Bowlby (1973), optimal functioning of the attachment
system depends on the availability of one or more attachment figures in
times of need and on their sensitivity and responsiveness to the individual’s
proximity-seeking behaviors. As Cassidy (1999) noted, ‘‘whereas nearly all
THE ATTACHMENT BEHAVIORAL SYSTEM 61
children become attached (even to mothers who abuse them, Bowlby,
1956), not all are securely attached’’ (p. 7). The quality of interactions
with attachment figures in times of need is, according to attachment
theory, the major source of individual differences in attachment-system
functioning.
On the one hand, when an attachment figure is available, sensitive, and
responsive to an individual’s proximity-seeking attempts, the attachment
figure facilitates the restoration of emotional equanimity and the attainment
of a sense of attachment security. As a result, the individual leaves these
satisfying interactions with increased confidence in proximity seeking as an
effective regulatory strategy, heightened trust in others’ availability and
responsiveness, and increased confidence in his or her own resources for
dealing with stress. Moreover, these interactions reinforce the individual’s
sense that the world is a safe place populated by people of good will, which
allows him or her to form rewarding relationships with others and engage
freely in nonattachment activities.
On the other hand, when the attachment figure is not physically or
emotionally available in times of need or is not sensitive or responsive to the
individual’s proximity-seeking attempts, the functioning of the attachment
system is disrupted and its set-goal is not attained. In these cases, the
individual does not experience distress alleviation. Rather, the distress
originally elicited by the encounter with threats is now compounded by
serious doubts that safety can be attained, that the world is a safe place, that
others can be trusted, and that the self has the resources necessary to
manage stress. Needless to say, the resulting sense of vulnerability and
uncertainty can interfere drastically with a whole range of life activities.
2. Secondary Attachment Strategies
Negative interactions with an unavailable and unresponsive attachment
figure have an additional immediate effect on attachment-system function-
ing. These interactions signal that the primary attachment strategy is failing
to accomplish its regulatory goal and that pursuing this strategy is
exacerbating rather than diminishing distress and insecurity. As a result,
the individual is forced to continue to deal with unregulated distress while
searching for alternative strategies of affect regulation to replace the
primary attachment strategy. Main (1990) labeled these alternatives
secondary attachment strategies. They are, one might say, the best a person
can do under dreadful circumstances.
Attachment theorists (e.g., Cassidy & Kobak, 1988; Main, 1990) have
delineated two major secondary attachment strategies that involve either
hyperactivation or deactivation of the attachment system. Hyperactivating
62
MIKULINCER AND SHAVER
strategies are a kind of ‘‘fight’’ (or as Bowlby, 1982/1969, said, ‘‘protest’’)
response to the frustration of attachment needs. A hyperactivating
individual does not give up his or her proximity-seeking attempts, and in
fact intensifies them in order to coerce an attachment figure’s love and
support. This requires that the vigilance, preoccupation, and energization
of behavior typical of short-term bids for proximity on the part of a less
anxious individual be maintained more or less constantly, even in the
face of repeated evidence of attachment-figure unavailability. In contrast,
deactivating strategies are a kind of ‘‘flight’’ reaction to the unavailability of
an attachment figure. The deactivating individual gives up his or her
proximity-seeking efforts, deactivates the attachment system without restor-
ing a sense of attachment security, and attempts to deal with dangers alone.
Bowlby (1982/1969) called this strategy ‘‘compulsive self-reliance.’’ Individ-
uals who rely on deactivating strategies of affect-regulation try to keep their
attachment system down-regulated to avoid the pain and distress caused
by attachment-figure unavailability. The two strategies have a variety of
psychological and interpersonal costs, which we will address later in this
chapter.
3. From Strategies to Mental Representations: Generalized
Individual Differences
Beyond producing immediate individual differences in reliance on a
specific attachment strategy during a given interaction, variations in the
quality of this interaction can also produce more enduring and generalized
individual differences in the functioning of the attachment system.
According to Bowlby (1973), this passage from context-tailored variations
to person-tailored variations is mediated by the cognitive substrate of the
attachment system—working models. Every interaction with an attachment
figure can be incorporated symbolically into working models of the self and
partner, models that allow a person to predict future interactions with the
partner and design new proximity-seeking attempts without rethinking each
one from the beginning. In this way, the internalization of a specific
interaction can affect the functioning of the attachment system in future
interactions with a particular attachment figure, and a strategy choice that
was originally manifested in a specific interaction can be generalized to
other relational episodes.
According to Main et al. (1985), early working models organize a child’s
memory about the self and the partner during attempts to gain security, as
well as the typical outcomes of those attempts: successful attainment of
security or failure of the primary attachment strategy. In this way, a child
can develop working models for successful proximity-seeking episodes, for
THE ATTACHMENT BEHAVIORAL SYSTEM 63
episodes in which the attachment system had to be hyperactivated, and
for episodes in which the system needed to be defensively deactivated. Each
of these models consists of episodic memories of the interaction sequence,
declarative knowledge about the partner’s responses and the efficacy of the
individual’s responses, and procedural knowledge about the ways in which
one responds to such situations and deals with different sources of distress.
In our view, the central organizing factor in a working model is the
attachment strategy used by the individual in a particular relational episode.
These strategies not only shape the procedural knowledge included in the
representations but also bias declarative knowledge about the self and the
attachment figure according to the goal these strategies are intended to
attain (e.g., getting hold of an insufficiently available attachment figure by
hyperactivating the attachment system or avoiding punishment or perpetual
frustration by deactivating the system). Because of such biases, working
models of the self and others reflect only in part the ways the person and the
partner actually behaved in a given interaction. They also reflect the
underlying regulatory action of attachment strategies that shape cognition,
affect, and behavior. That is, working models of self and others are always
blended reflections of what actually happened in a social encounter and
subjective biases resulting from attachment strategies.
Like other mental representations, which are presumably the psycho-
logical manifestations of underlying neural networks, working models form
excitatory and inhibitory associations with one other, and the activation of
one model primes congruent models and inhibits incongruent models. That
is, experiencing or thinking about an episode of security attainment
activates memories of other congruent episodes of successful proximity-
seeking attempts and renders memories of hyperactivation and deactivation
less accessible. With the passage of time and the recurrent retrieval of related
memories, these associative links are strengthened and favor the formation
of more abstract and generalized representations of attachment-system
functioning with a specific partner. In this way, models of security-
attainment, hyperactivation, and deactivation with a specific attachment
figure (relationship-specific models) are created, and they form excitatory
and inhibitory links with models representing interactions with other
attachment figures. With the consolidation of these links, even more generic
working models are formed, global representations of the self and others
across different relationships. The end product of this cognitive generaliza-
tion and consolidation process is the formation of a hierarchical associative
network, in which episodic memories become exemplars of relationship-
specific models, which in turn become exemplars of generic relational
schemas. As a result, with respect to a particular relationship and across
different relationships, everyone has models of security-attainment,
64
MIKULINCER AND SHAVER
hyperactivation, and deactivation and so can sometimes think about
relationships in secure terms and at other times think about them in more
hyperactivating or deactivating terms.
The idea of multiple working models raises questions about the personal
and contextual factors that determine which kind of model is activated and
used to guide expectations, concerns, and behaviors in a particular
interaction with an attachment figure. In our view, this activation depends
mainly on the relative strength of a model, which is determined by the
amount of experience on which it is based, the number of times it has been
applied in the past, and the density of its connections with other cognitive
representations (see also Collins & Read, 1994; Shaver, Collins, & Clark,
1996). At a relationship-specific level, this idea implies that the model
representing the typical interaction with an attachment figure has the
highest chance of being activated in subsequent interactions with the same
person. At a generic level, the model that represents interactions with major
attachment figures (e.g., parents, spouse) becomes the most easily activated
and chronically accessible attachment-related representation.
In addition to the history of attachment interactions, features of the
current situation also contribute to the activation of a particular working
model. For example, clear-cut contextual cues concerning a partner’s love,
availability, and supportiveness can activate models of security attainment.
In addition, working models can be invoked by a person’s current motives
(e.g., wishing to distance from a partner) or by his or her current mood
(Shaver et al., 1996). In our view, the chronically accessible model coexists
with less typical working models in the memory network, and these models
can be activated by contextual factors or a person’s inner state.
The consolidation of a chronically accessible working model is the most
important psychological process that explains the passage from context-
tailored variations in attachment-system functioning to person-tailored
variations. Given a fairly consistent pattern of interactions with attachment
figures during childhood and adolescence, the most representative models of
these interactions are solidified through thousands of repeated experiences
and increasingly become part of the individual’s developing personality.
Like other mental schemas, these chronically accessible models tend to
operate automatically and unconsciously and are resistant to change. Thus,
what began as representations of specific interactions with a specific partner
become core personality characteristics and tend to be applied in new
situations and relationships where they continue to guide the functioning of
the attachment system.
Bowlby (1979) aptly illustrated how these chronically accessible models
shape a person’s experience. In his words, a person ‘‘tends to assimilate any
new person with whom he may form a bond, such as a spouse, or child, or
THE ATTACHMENT BEHAVIORAL SYSTEM 65
employer, or therapist, to an existing model (either of one or other parent or
of self ), and often to continue to do so despite repeated evidence that the
model is inappropriate. Similarly he expects to be perceived and treated by
them in ways that would be appropriate to his self-model, and to continue
with such expectations despite contrary evidence’’ (pp. 141–142).
4. Measurement of Individual Differences: Attachment Style
Most of the research examining individual differences in attachment-
system functioning has focused on attachment styles—patterns of expect-
ations, needs, emotions, and social behavior that result from a particular
history of attachment experiences, usually beginning in relationships with
parents (Fraley & Shaver, 2000). Attachment style reflects the individual’s
most chronically accessible working models and the typical functioning of
his or her attachment system in a specific relationship (relationship-specific
style) or across relationships (global or general attachment style). As such,
each attachment style is closely tied to working models and reflects the
underlying, organizing action of a particular attachment strategy (primary
or secondary, hyperactivating or deactivating).
As mentioned earlier, these styles were first described by Ainsworth (1967;
Ainsworth et al., 1978) based on her observations of infants’ responses to
separations from and reunions with mother, both at home and in a labora-
tory strange situation. Using this assessment procedure, infants are classified
into one of three attachment categories: secure, avoidant, or anxious/ambi-
valent. Recently, a fourth category, ‘‘disorganized/disoriented,’’ has been
added. It is characterized by odd, awkward behavior during separation and
reunion episodes and random fluctuations between signs of anxiety and
avoidance (Main et al., 1985).
Infants classified as secure seem to hold accessible working models of
successful proximity-seeking attempts and security-attainment. In the
strange situation, they tend to exhibit distress during separation but then
recover quickly and continue to explore the environment with interest.
When reunited with mother, they greet her with joy and affection, respond
positively to being held, and initiate contact with her. During home
observations, mothers of these infants are emotionally available in times of
need and responsive to their infants’ proximity-seeking behavior (Ainsworth
et al., 1978). It seems reasonable to characterize these mothers as a source of
attachment security and as reinforcing reliance on the primary attachment
strategy.
Avoidant infants seem to hold accessible working models of attachment-
system deactivation. In the strange situation they show little distress when
separated from their mother and avoid her upon reunion. In home-based
66
MIKULINCER AND SHAVER
observations, their mothers tend to be emotionally rigid as well as angry at
and rejecting of their infants’ proximity-seeking attempts (Ainsworth et al.,
1978). Anxious infants seem to hold accessible working models of
attachment-system hyperactivation. In the strange situation, they are
extremely distressed during separation and exhibit conflictual responses
toward their mother at reunion (e.g., they may cling one moment and
angrily resist comforting the next, which was Ainsworth’s reason for calling
them ‘‘ambivalent’’). During home-based observations, interactions between
anxious infants and their mothers were characterized by lack of harmony
and lack of caregivers’ consistent responsiveness (Ainsworth et al., 1978).
Mothers of both avoidant and anxious infants seem to thwart security
attainment, thereby fostering the adoption of secondary strategies.
However, whereas avoidant infants deactivate the attachment system in
response to attachment-figure unavailability, anxious infants tend to
hyperactivate the system (Main, 1990; Main et al., 1985).
Disorganized/disoriented infants seem to suffer from a breakdown of
organized attachment strategies (primary, hyperactivating, or deactivating).
They either oscillate between strategies or do something bizarre like lie face
down on the floor without moving or sit under a table, evincing no strategy
at all. This pattern of behavior seems to be due to disorganized,
unpredictable, and discomfiting behavior on the part of attachment figures
who, research shows, are likely to be suffering from unresolved losses or
attachment-related traumas (Hesse, 1999; Lyons-Ruth & Jacobvitz, 1999).
With the extension of attachment research into adolescence and
adulthood, new measures of attachment style were created for use with
those age groups. Based on a narrative approach, Main and her colleagues
(George, Kaplan, & Main, 1985; Main et al., 1985; see Hesse, 1999, for a
review) devised the Adult Attachment Interview (AAI), in which partici-
pants answer open-ended questions about their childhood relationships with
parents. In the AAI, interviewees are classified into three primary categories
that parallel Ainsworth’s infant–mother attachment typology: ‘‘secure’’ (or
free and autonomous with respect to attachment), ‘‘dismissing’’ (of
attachment), or ‘‘preoccupied’’ (with attachment). If a narrative contains
indications of unresolved traumas or losses, it gets a secondary classification
of ‘‘unresolved.’’
AAI narratives are coded as indicative of secure working models if an
interviewee describes positive relationships with parents in a clear, convin-
cing, and coherent manner or if negative relationships are described
coherently with some degree of perspective. The narratives are coded as
indicative of dismissing attachment if the interviewee dismisses the
importance of attachment relationships or idealizes them and provides no
clear examples to support his or her characterizations. This category parallels
THE ATTACHMENT BEHAVIORAL SYSTEM 67
Ainsworth’s avoidant style and reflects the underlying action of deactivating
strategies. The narratives are coded as preoccupied if the interviewee is
entangled in still-intense worries and conflicted feelings about parents, can
easily retrieve attachment-related memories, but has trouble coherently
discussingthemwithoutangeroranxiety.ThiscategoryparallelsAinsworth’s
anxious style and reflects the underlying action of hyperactivating strategies.
Despite the richness of AAI narratives, the interview is expensive to
administer and score and it deals almost exclusively with memories of child–
parent relationships. Given that we are primarily interested in attachment
behavior in the context of adult close relationships and would like to be able
to run large questionnaire and experimental studies in a reasonably short
period of time, our research program relies on simple self-report measures of
attachment style.
The first such measure was constructed by Hazan and Shaver (1987, 1990)
for their studies of romantic attachment. In its original form, the measure
consisted of three brief descriptions of feelings and behaviors in close
relationships that were intended to parallel Ainsworth’s three types of
infants. Participants were asked which of the three descriptions best
characterizes them in close relationships (see Table I). However, subsequent
studies (e.g., Bartholomew & Horowitz, 1991; Brennan, Clark, & Shaver,
1998; Fraley & Waller, 1998) revealed that attachment styles are best
conceptualized as regions in a two-dimensional space, which conceptually
parallels the space defined by two discriminant functions in the summary of
research on infant–mother attachment by Ainsworth et al. (1978, Fig. 10, p.
102). The two dimensions defining the space are called attachment anxiety
and attachment avoidance (Brennan et al., 1998). (See Fraley & Spieker,
2003, for a similar analysis of individual differences among infants in the
strange situation.)
Table I
Original T hree-C ategory Measureof Romantic Attachment Style
a
I am somewhat uncomfortable being close to others; I find it difficult to trust them
completely, difficult to allow myself to depend on them. I am nervous when anyone gets too
close and often, others want me to be more intimate than I feel comfortable being. (Avoidant)
I find that others are reluctant to get as close as I would like. I often worry that my partner
doesn’t really love me or won’t want to stay with me. I want to get very close to my partner and
this sometimes scares people away. (Anxious)
I find it relatively easy to get close to others and am comfortable depending on them and
having them depend on me. I don’t worry about being abandoned or about someone getting too
close to me. (Secure)
a
Hazan and Shaver (1987, 1990)
68 MIKULINCER AND SHAVER
In the two-dimensional attachment-style space (see Fig. 1), what was
formerly called the ‘‘secure style’’ is a region in which both anxiety and
avoidance are low. This region is defined by a sense of attachment security,
comfort with closeness and interdependence, and reliance on the primary
attachment strategy in times of need. What was formerly called the ‘‘anxious
style’’ refers to the region in which anxiety is high and avoidance is low. This
region is defined by a lack of attachment security, a strong need for
closeness, worries about relationships, and reliance on hyperactivating
strategies. What was called the ‘‘avoidant style’’ refers to a region in which
avoidance is high. This region is defined by a lack of attachment security,
compulsive self-reliance, preference for emotional distance from others, and
reliance on deactivating strategies. In the research of Ainsworth et al. (1978),
avoidant infants occupied mainly the region where avoidance is high and
anxiety is low. In adult attachment research, Bartholomew and Horowitz
(1991) drew a distinction between ‘‘dismissing avoidants’’ (high on
LOW AVOIDANCE
HIGH AVOIDANCE
SECURE PREOCCUPIED
DISMISSING-
AVOIDANT
FEARFUL-
AVOIDANT
LOW ANXIETY HIGH ANXIETY
Fig. 1. Diagram of the two-dimensional space defined by attachment anxiety and attach-
ment avoidance. The terms in the four quadrants are Bartholomew’s (1990) names for the four
major attachment styles.
THE ATTACHMENT BEHAVIORAL SYSTEM
69
avoidance, low on anxiety) and ‘‘fearful avoidants’’ (high on both avoidance
and anxiety).
Because several different measures have been used to study adult
attachment, some based on three categories, some on four categories, and
some on two dimensions, it is somewhat difficult to summarize results
of studies on a particular topic without constantly raising the issue of
measurement. Here, we will generally summarize results in terms of ‘‘secure,’’
‘‘anxious,’’ and ‘‘avoidant’’ attachment patterns, unless some feature of the
results requires reference to the distinction between fearful and dismissing
avoidance or to simultaneously high scores on both the attachment anxiety
and attachment avoidance dimensions. The issue of ‘‘fearful avoidance’’ is, in
any case, less likely to arise in normal samples of college students and
community adults. Extremely high scores on both the anxiety and avoidance
dimensions are more common in samples of abused or clinical samples (see
Shaver & Clark, 1994, for a review), where ‘‘disorganized’’ attachment
behavior is a consequence of maltreatment or other contributors to
psychopathology. In most of our studies, the results can be adequately
described in terms of either anxiety or avoidance.
At present, a person’s locations on the two dimensions are measured with
two reliable 18-item Likert scales created by Brennan et al. (1998) based on
factor analyses of previous measures. The two scales are reliable in both the
internal-consistency and test–retest senses and have high construct,
predictive, and discriminant validity (Crowell, Fraley, & Shaver, 1999).
The following items are representative of the avoidance scale: ‘‘I try to avoid
getting too close to my partner,’’ ‘‘I prefer not to show a partner how I feel
deep down,’’ and ‘‘I turn to my partner for many things, including comfort
and reassurance’’ (reverse scored). The following items are taken from the
anxiety scale: ‘‘I need a lot of reassurance that I am loved by my partner,’’
‘‘I do not often worry about being abandoned’’ (reverse scored), and
‘‘I get frustrated if romantic partners are not available when I need them.’’
The two scales were conceptualized as independent and, like the two
discriminant functions reported by Ainsworth et al. (1978), have been found
to be uncorrelated in most studies.
It is important to note that despite the substantial differences in focus
(parent–child vs. adult–adult relationships) and method (brief self-reports
vs. extensive interview transcripts), self-report measures of attachment in
close relationships are related to AAI coding scales (Bartholomew & Shaver,
1998). For example, in a study of over 100 married women, Shaver, Belsky,
and Brennan (2000) found that both avoidance and anxiety scores based on
self-report scales could be predicted from AAI coding scales. Analyses
running in the other direction revealed that every AAI coding scale except
one was predictable from self-report items. These findings imply that scores
70
MIKULINCER AND SHAVER
on the two kinds of measures are related to each other in sensible ways, and
that both are reflections of an underlying sense of attachment security, even
though their differences in focus, method, and interpretation make them far
from identical.
5. Summary
In this section we introduced major individual differences in the
functioning of the attachment system. Although the system is an inborn,
prewired device that presumably operates mainly at a subcortical level and
in a reflexive, mechanistic, and unintentional manner, its activation can lead
to different response strategies (primary, hyperactivating, or deactivating)
depending on both the quality of the current interaction with an actual
attachment figure and internal representations of past interactions with the
same or other attachment figures. These representations (working models
of self and others) presumably operate in a more cortical, reflective,
constructive, and intentional manner (although they can become habitual
and automatic, and can be held out of consciousness by defensive
maneuvers), and can transform context-tailored variations in the activation
of specific attachment strategies to person-tailored variations. That is, they
are a source of within-person continuity in attachment-system functioning.
In the next section, we present a theoretical framework that makes sense of
the cognitive operations and dynamic processes involved in the activation
and functioning of the attachment system in adulthood.
II. An Integrative Model of Attachment-System Dynamics in Adulthood
In attempting to characterize the dynamics of the activation and
functioning of the attachment system in adulthood, we (Mikulincer, Shaver,
& Pereg, in press; Shaver & Mikulincer, 2002a, 2002b) have proposed a
control-system model that integrates recent findings with earlier theoretical
proposals by Bowlby (1973, 1980, 1982/1969), Ainsworth (1991), Cassidy
and Kobak (1988), and Main (1995). This model is a conceptual extension
and refinement of previous control-systems representations of attachment-
system dynamics presented by Shaver, Hazan, and Bradshaw (1988) and
Fraley and Shaver (2000). It includes three components (Fig. 2): the primary
strategy associated with attachment-system activation, the strategy related to
the attainment of a sense of attachment security, and the strategies triggered
by attachment-figure unavailability, which lead to a regulatory failure of the
attachment system. The model also delineates (1) the main goals of each of
THE ATTACHMENT BEHAVIORAL SYSTEM 71
FIRST MODULE
Continue with
ongoing
activities
Signs
of threat?
Activation of the
attachment system
Yes
No
−
+
−
+
Attachment insecurity
(compounding of distress)
SECOND MODULE
Engagement in
nonattachment
activities (e.g.,
exploration,
caregiving)
Attachment security,
security-based strategies
Is
attachment figure
available, attentive,
responsive, etc?
No
Yes
Seeking proximity to external or
internalized attachment figure
Hypervigilance regarding threat-
and attachment-related cues
Hyperactivating strategies
THIRD MODULE
Distancing of
threat- and
attachment-related
cues
Is proximity-seeking
a viable option?
Yes
No
Deactivating
strategies
Fig. 2. Shaver and Mikulincer’s (2002a) integrative model of the activation and dynamics
of the attachment system.
72 MIKULINCER AND SHAVER
these strategies; (2) their declarative knowledge base and procedural rules for
managing interpersonal behavior, coping with stress, and processing
information; and (3) their implications for self-image, social judgments,
mental health, relationship quality, and other behavioral systems.
The first component involves the monitoring and appraisal of threatening
events; it is responsible for activation of the attachment system. The second
component involves the monitoring and appraisal of attachment-figure
availability and is responsible for individual differences in the sense of
attachment security and the psychological correlates and consequences of
the secure attachment style. The third component involves monitoring and
appraisal of the viability of proximity seeking as a way of dealing with
attachment insecurity. This component is responsible for individual
variations in the use of hyperactivating or deactivating strategies and
accounts for the psychological manifestations of attachment anxiety and
avoidance. The model also includes hypothetical excitatory and inhibitory
‘‘neural circuits’’ (shown as arrows on the left-hand side of Fig. 2), resulting
from the recurrent use of hyperactivating or deactivating strategies. These
circuits affect the monitoring of threats and the monitoring of attachment
figures’ availability. All components and circuits of the model can operate
either consciously or unconsciously. Moreover, these components and
circuits can operate either in parallel or in opposite ways at conscious and
unconscious levels. (We will provide concrete examples later in the chapter.)
Our model is sensitive to both context- and person-tailored individual
differences. On the one hand, each component of the model can be affected
by specific contextual factors (e.g., actual threats, information about
attachment-figure availability or proximity-seeking viability), which initiate
a bottom-up process in a person’s working models, activating congruent
attachment strategies and producing immediate changes in the functioning
of the attachment system. On the other hand, each component of the model is
affected by the chronically accessible working model, which biases the
appraisal of threats, attachment-figure availability, and proximity-seeking
viability. These biases are part of a top-down process by which the attachment
system functions in accordance with a person’s chronic attachment style.
Overall, the model gives place to both reality—the current context in which the
attachment system is activated—and fantasy, the cognitive biases resulting
from attachment strategies.
A. ACTIVATION OF THE ATTACHMENT SYSTEM
Following Bowlby’s (1982/1969) ideas, we assume that the monitoring of
unfolding internal or external events results in activation of the attachment
THE ATTACHMENT BEHAVIORAL SYSTEM 73
system when a potential or actual threat is perceived. These triggers include
both physical and psychological threats. They also include both attachment-
unrelated and attachment-related sources of threat. In our view, every
unfolding event perceived by a person as threatening his or her sense of
security, and consequently his or her adjustment and survival, can activate
the attachment system. This activation automatically heightens
the accessibility of attachment-related cognitions, which in turn favor the
seeking of proximity to supportive others. Although this component of the
model represents the normative working of the system, which occurs
regardless of individual differences in attachment history and orientation, it
is still affected by the excitatory circuits of hyperactivating strategies and the
inhibitory circuits of deactivating strategies.
In our model, attachment-system activation depends on the subjective
appraisal of threats rather than their actual occurrence. Although the actual
presence of threats is obviously a very important factor, the individual’s
appraisal of internal or external events as potential or actual threats is the
critical factor in activating the attachment system. This idea fits with Lazarus
and Folkman’s (1984) model of stress and coping, which emphasizes the
critical role played by appraisal processes in the regulation of distress and the
energization of coping efforts. In fact, the attachment system is designed to
cope with threats and regulate distress, and it, like other regulatory devices,
depends on a person’s perception that unfolding events are threats that require
the mobilization of coping efforts.
We view the appraisal of threats as a product of both the actual presence
of threat-related cues and the individual’s forecast or expectations that
unfolding events may have negative effects on well-being, adjustment, and
survival. Although these expectations can follow a rational, conscious risk-
analysis of the unfolding events, excitatory and inhibitory circuits resulting
from secondary attachment strategies can automatically bias these
expectations. That is, the appraisal process is not necessarily conducted at
a conscious, rational level and the person may not even be aware of the
monitoring and appraisal of threats. These cognitive processes can occur at
preconscious levels and be manifested in physiological reactions and the
measurable accessibility—readiness to influence information processing—of
threat-related thoughts.
Another important feature of our model is that inner sources of threat
(e.g., thoughts, imagery) can activate the attachment system. For example,
thoughts about personal death are usually appraised as extremely
threatening (Greenberg, Pyszczynski, & Solomon, 1997) and so can activate
the attachment system. In such cases, the person does not need to actually
encounter the threat: Merely being reminded of death is enough to activate
the system. Again, these thoughts need not be conscious; they can be
74
MIKULINCER AND SHAVER
activated by a stimulus as minimal as the word ‘‘death’’ presented
subliminally (e.g., Mikulincer, Birnbaum, Woodis, & Nachmias, 2000).
These two features—subjective appraisal and the activating power of
threat-related thoughts—are critical for delineating the effects of the neural
circuits associated with secondary attachment strategies on attachment-
system activation. On the one hand, hyperactivating strategies, which
maintain the attachment system in a chronically activated state, include
vigilance with respect to triggers of attachment-system activation and
exaggeration of the threatening aspects of person–environment transactions;
they also include rumination on potential threats, which occupies working
memory. This means anxiously attached individuals can experience
attachment-system activation in the absence of any consensual, external
sign of danger. On the other hand, deactivating strategies, which maintain
the attachment system in a down-regulated or inactive state, foster the
ignoring or dismissal of threatening aspects of person–environment
transactions and the suppression of any threat-related thought that might
activate the attachment system. This means that avoidant individuals can
distance themselves from various sources of threat and keep themselves
from thinking about either their need for protection or the relief that might
be found in the presence of a loving and protective partner.
Once a threat is appraised, attachment-system activation is automatically
manifested in heightened accessibility of attachment-related nodes within the
associative memory network. These nodes include internal representations of
security-enhancing attachment figures; episodic memories of supportive and
comforting interactions with these figures; thoughts related to proximity,
love, and support; and proximity-seeking goals. These preconsciously
activated nodes become ready for use in subsequent information processing
and can shape a person’s state of mind and influence his or her behavioral
plans even before they are consciously formulated. This idea fits with recent
findings from social cognition research showing that accessible cognitive
material can shape a person’s state of mind before he or she recognizes the
material in the stream of consciousness (Wegner & Smart, 1997). It is also
congruent with the ‘‘auto-motive’’ model (Bargh, 1990), which suggests that
goals can be preconsciously activated and this preconscious activation can
automatically guide a person’s behavior without the mediation of conscious
behavioral planning.
This automatic, preconscious accessibility of attachment-related memory
nodes is the normative manifestation of attachment-system-activation in
adulthood. We assume that age and development result in an increased
ability to gain security from internalized representations of security-
enhancing attachment figures without necessarily causing a person to
actually seek proximity to these figures. This cognitive activation maintains
THE ATTACHMENT BEHAVIORAL SYSTEM 75
symbolic proximity to figures who in the past provided security and safety.
That is, the regulatory action of the adult attachment system can occur
intrapsychically without a person’s awareness. However, like Bowlby (1982/
1969), we assume that no one of any age is completely free of dependence on
others. There are situations, such as physical and psychological traumas,
illnesses, or losses, in which symbolic proximity to internalized figures is not
sufficient to provide a sense of security, and in such situations attachment-
system activation leads to proximity-seeking behavior. There are also
developmental stages, such as old age, that tax people’s physical and
psychological resources to the point where it becomes necessary to seek
proximity to and support from others.
Overt manifestations of attachment-system activation in conscious
thoughts, behavioral intentions, and actual behaviors are dependent on
contextual, cultural, and personal factors and do not necessarily follow
directly from the accessibility of attachment-related nodes. Furthermore,
variations in proximity-seeking behavior are more a reflection of the under-
lying action of attachment strategies than of attachment-system activation
itself. In fact, people differing in attachment style differ in their tendency to
seek actual proximity in times of need.
These individual differences are also manifested in the content of the
attachment-related nodes that are automatically activated by threat
appraisal. For people with a history of security attainment and a secure
attachment style, threat appraisal makes accessible thoughts about typical,
positive interactions with attachment figures—i.e., thoughts of proximity,
support, love, comfort, and relief. However, for persons with a painful
history of attachment interactions and an insecure attachment style, threat
appraisal can make accessible many negative attachment-related thoughts
(e.g., thoughts about separation and rejection). These people’s frustrating
attachment experiences may create an associative link in their memory
networks between activation of the attachment system and worries about
separation or rejection, so that the accessibility of these worries increases
every time attachment needs are activated.
In sum, the appraisal of unfolding events as threatening automatically
activates attachment-related nodes in the associative memory network and
preconsciously activates internalized representations of attachment figures
and proximity-seeking goals. This activation can remain at the preconscious
level and be a source of comfort. However, it can also be manifested in
attachment-related conscious thoughts and actual proximity-seeking behav-
iors, depending on contextual factors and the person’s attachment style.
Attachment style is also a source of individual differences in threat appraisal
and the specific content of the attachment-related nodes activated by the
perceived threat.
76
MIKULINCER AND SHAVER
B. ATTACHMENT-FIGURE AVAILABILITY AND
SECURITY-BASED STRATEGIES
Once the attachment system is activated, an affirmative answer to the
question, ‘‘Is the attachment figure available?’’ results in a sense of
attachment security and positive models of the self and others. Attachment-
figure availability also reinforces the perceived effectiveness of the seeking of
proximity and fosters the development of what we call security-based
strategies. The goals of these strategies are to form and maintain close bonds
with others, alleviate distress, and bolster personal adjustment through
constructive, flexible, and reality-attuned mechanisms. Moreover, they
create what we, following Fredrickson (2001), call a ‘‘broaden and build’’
cycle of attachment security, which builds a person’s resources for
maintaining emotional equanimity in times of stress and broadens his or
her perspectives and capacities.
In the long run, repeated episodes of attachment-figure availability have a
powerful and enduring impact on intrapsychic organization and interper-
sonal behavior. At the intrapsychic level, these episodes lead to the
consolidation of security-based strategies as the main method of affect
regulation and of secure working models of self and others as the most
chronically accessible cognitive representations. At the interpersonal level,
these episodes foster the enactment of a secure attachment style in social
situations and close relationships. Security-based strategies and the resulting
‘‘broaden and build’’ cycle of attachment security can be viewed as the core
characteristics of securely attached persons.
As in the system-activation module, the answer to the question about
attachment-figure availability in the second module depends on subjective
appraisal of this availability and can be biased by attachment strategies.
The neural networks or circuits of insecure persons make the positive
appraisal of attachment-figure availability unlikely. Anxious persons’
hyperactivating strategies intensify the vigilant monitoring of attachment-
figure behaviors and slant perceptions in the direction of noticing or
imagining insufficient interest, availability, and responsiveness. As a result,
the likelihood of detecting signs of distance, rejection, and unavailability is
increased, because the attachment figure cannot always be available and
totally at the disposition of the attached person’s needs. Avoidant
individuals’ deactivating strategies interfere with the monitoring of
cues concerning either the availability or unavailability of the attachment
figure, increasing the likelihood that genuine and clear-cut signals of
attachment-figure availability will be missed. In contrast, security-based
strategies facilitate the appraisal of attachment-figure availability—they
reflect a positive view of relationship partners as available and supportive,
THE ATTACHMENT BEHAVIORAL SYSTEM 77
and favor the ignoring or downplaying of episodes of momentary
unavailability.
These cognitive biases are amplified when attachment nodes in the
memory network are preconsciously activated. At the preconscious level,
appraisal of attachment-figure availability depends entirely on the type of
internalized figure, available or unavailable, that is activated. Insecurely
attached persons tend to give a negative answer to the question of
attachment-figure availability, because they have chronic access to
representations of unavailable figures and these representations become
increasingly accessible when attachment nodes are activated. More securely
attached people, however, tend to give a positive answer to this question,
because they have chronic access to representations of available figures and
their attachment nodes are linked with a host of associations and memories
of security-enhancing attachment figures.
Despite these cognitive biases, however, reality is still important in the
appraisal of attachment-figure availability. In our view, the actual presence
of an available attachment figure or contextual cues that activate
representations of available attachment figures (thinking about one of these
figures) can lead people to give an affirmative answer to the question of
attachment-figure availability. These contexts, mainly when they are clear-
cut, personally meaningful, and stable over time and situations, can
counteract even insecure people’s tendencies to negate the availability of
attachment figures, and can set in motion the ‘‘broaden and build’’ cycle
of attachment security. That is, these contextual cues can influence the
appraisal of attachment-figure availability beyond the cognitive biases
resulting from attachment strategies. Accordingly, these cues can activate
security-based strategies even among chronically insecure persons.
Security-based strategies include three core beliefs: optimistic beliefs
about distress management, a sense of trust in others’ availability and good
will in times of need, and a sense of self-efficacy in dealing with threats.
These beliefs are a direct result of positive interactions with available
attachment figures. During these interactions, individuals learn that distress
is manageable and external obstacles can be overcome. Moreover, they learn
about others’ positive responses to attachment behaviors and about the
control one can exert over the course and outcome of threatening events. In
our view, these cognitive acquisitions are the building blocks of a person’s
capacity for dealing with stress.
The procedural knowledge involved in security-based strategies consists
of a set of rules embodied in what Waters, Rodrigues, and Ridgeway (1998)
called the ‘‘secure base script.’’ This hypothetical script is organized around
three regulatory tendencies—acknowledgment and display of distress, the
seeking of intimacy, closeness, and support, and engagement in instrumental
78
MIKULINCER AND SHAVER
problem solving. The ‘‘emotion-focused coping’’ components of this script
(Lazarus & Folkman, 1984)—acknowledging and expressing feelings and
seeking emotional support—work in the service of alleviating distress so
that ‘‘problem-focused coping’’ components—seeking instrumental support
and solving problems—can proceed successfully. Again, these tendencies
seem to stem from prior appraisals of attachment-figure availability and
consequent confirmations that proximity seeking results in protection,
support, and relief of distress. Relatively secure individuals have learned
that acknowledgment and display of distress elicit positive and supportive
responses from others. They have also learned that their own actions are
often able to reduce distress and remove obstacles, and that turning to
others when threatened is an effective route to enhanced coping.
These regulatory strategies are directly manifested in interpersonal
behaviors and close relationships. One basic declarative component of
security-based strategies is the knowledge, or belief, that proximity
maintenance is rewarding and that relationships are an adequate framework
for satisfying one’s needs. As a result, these strategies predispose people to
feel comfortable with intimacy and interdependence, emphasize the benefits
of being together, and organize their interaction goals around the search for
intimacy. Another declarative component of security-based strategies is the
belief in others’ good will, which fosters a sense of trust, gratitude, and
affection toward a relationship partner as well as tolerance for ambiguous or
negative partner behaviors. These strategies also include a positive attitude
toward affect display, which heightens the readiness to self-disclose and
share personal feelings with a relationship partner.
Security-based strategies are also manifested in the process of coping with
stress and the management of negative emotions. Secure individuals’
optimistic and constructive attitude toward life’s difficulties leads them to
adopt what Epstein and Meier (1989) called constructive ways of coping—
active attempts to remove the source of distress, manage the problematic
situation, and restore emotional equanimity without creating negative
socioemotional side effects. These constructive ways of coping consist of
active problem solving as well as transformational strategies that involve
symbolically transforming unsolvable problems into meaningful, growth-
promoting challenges. They also include reliance on close relationships as an
anxiety buffer and the seeking of proximity and support as a means of
coping with stress.
The constructive nature of security-based strategies is enhanced by their
flexibility and reality-attunement. Optimistic beliefs about self-efficacy and
distress management allow people to open their cognitive structures to new,
even threatening, information, and then to flexibly adjust their strategies for
dealing realistically with environmental demands. Experiencing, or having
THE ATTACHMENT BEHAVIORAL SYSTEM 79
experienced, attachment figures as approving allows people to revise
erroneous beliefs without excessive fear of criticism or rejection, thus
avoiding the cognitive and motivational entrapment that results from being
unable to revise beliefs. Furthermore, the building of a person’s resources
for dealing with stress makes it less necessary to rely on maladaptive means
of coping, including primitive defense mechanisms that distort perceptions
and generate interpersonal conflicts.
In our view, the adoption of security-based strategies epitomizes optimal
functioning of the attachment system. Such strategies are expected to
contribute directly to the development of generalized positive representa-
tions of others, the consolidation of a stable sense of self-efficacy and self-
esteem, the maintenance of mental health and adjustment in times of stress,
and the formation of long-lasting, satisfying close relationships. Moreover,
these strategies facilitate the deployment of resources to other behavioral
systems, such as exploration, sex, caregiving, and affiliation, and thereby
contribute to the broadening of a person’s perspectives and capacities. The
sense of attachment security allows people to engage in exploratory
activities, enjoy sex and social interactions, and be attentive and responsive
to others’ needs. Moreover, with confidence that support is available when
needed, people can take risks and engage in autonomy-promoting activities.
In other words, security-based strategies facilitate the development of
autonomy, individuality, and self-actualization.
C. PROXIMITY-SEEKING VIABILITY AND SECONDARY
ATTACHMENT STRATEGIES
Attachment-figure unavailability results in attachment insecurity and
compounds the distress anyone might experience when encountering a
threat. We claim that this painful state forces a decision about the viability
of proximity seeking as a means of self-regulation and the subsequent choice
of a secondary attachment strategy. As in the two modules of Fig. 2 already
described, a decision about proximity-seeking viability depends on
subjective appraisal processes, which are affected by organismic and
contextual cues that will be described later.
1. Hyperactivating Strategies
When proximity seeking is appraised as being a viable option if only
greater effort is expended, people tend to make very energetic, insistent
attempts to attain security—in short, to use hyperactivating strategies
(Cassidy & Kobak, 1988). The main goal of these strategies is to get an
80
MIKULINCER AND SHAVER
attachment figure who is viewed as unreliable or insufficiently available and
responsive to pay attention and provide protection or support. The basic
means for attaining this goal is to maintain the attachment system
chronically activated until an attachment figure is perceived to be available
and to provide safety and security. This strategy involves exaggerating the
presence and seriousness of threats and being highly vigilant regarding
attachment-figure availability or unavailability, because these cues are
highly relevant for security attainment. This constant worry and vigilance
increases the likelihood that a person will notice real or imagined signs of a
relationship partner’s disapproval, waning interest, or impending departure.
In the long run, repeated episodes of attachment-figure unavailability or
nonresponsiveness combined with the appraisal of proximity-seeking
viability lead to the consolidation of hyperactivating strategies as the main
regulatory device. At the interpersonal level, this internal process is
manifested in heightened desires for closeness, merger, and security in close
relationships together with an increased preoccupation with the partner’s
unavailability—the two definitional components of attachment anxiety
(Brennan et al., 1998). We view hyperactivating strategies as the key
underlying characteristics of the anxious attachment style.
Hyperactivating strategies are sustained by negative declarative beliefs
about distress management, others’ good will, and one’s self-efficacy in
dealing with threats. Although these beliefs are developed during negative
interactions with insufficiently available, responsive, or reliable attachment
figures, they also reflect cognitive biases that overgeneralize past attachment
injuries and apply memories about them inappropriately to new situations
and partners. They also result from the amplification of threat appraisals
that maintain the attachment system chronically activated.
At the interpersonal level, hyperactivating strategies manifest themselves
as exaggerations of the primary attachment strategy: constant monitoring of
the relationship partner and strong efforts to maintain proximity. They
consist of overdependence on a relationship partner as a source of comfort;
intense demands for attention and care; strong desire for enmeshment or
merger; attempts to minimize cognitive, emotional, and physical distance
from a partner; and clinging and controlling behaviors designed to elicit
the partner’s affection and support. These strategies also reflect a
malfunction of the primary attachment strategy produced by exaggerated
threat appraisals (i.e., a heightened sense of vulnerability) and exaggerated
appraisals of attachment figures’ unavailability. This malfunction is evident
in anxious persons’ hypersensitivity to rejection cues and worries about
separation and abandonment.
In the process of coping with threats, hyperactivating strategies are
manifested in what Lazarus and Folkman (1984) called ‘‘emotion-focused
THE ATTACHMENT BEHAVIORAL SYSTEM 81
coping’’—hypervigilant attention to internal indications of distress. This
method of coping consists of intensification of negative emotions, mental
rumination on related negative cognitions, self-preoccupation, self-criticism,
and overt displays of distress. As a result, anxiously attached individuals
have free access to threat-related thoughts and emotions and are unable or
unwilling to suppress them.
Intrapsychically, amplification of threat appraisals heightens the chronic
accessibility of negative thoughts and makes it likely that new sources of
distress will mingle and become confounded with old accessible ones. That
is, activation of one cognitive node with a negative emotional tone will
automatically spread to other negatively tinged cognitive nodes, because all
of these cognitions are maintained simultaneously in working memory.
Moreover, this pattern of cognitive activation gives predominance to the
affective implications of the information and favors the organization of
cognitions in terms of simple, undifferentiated features, such as the extent to
which the information is threatening or implies rejection. In this way,
hyperactivating strategies create a chaotic mental architecture that is
constantly pervaded by negative affect.
The heightened accessibility of threat-related thoughts also creates a
mental context that biases the encoding of new information. According to
cognitive theories (e.g., Anderson, 1994), this context facilitates the
encoding of congruent information (e.g., threat-related cues), while
diverting attention away from the encoding of context-incongruent cues.
This context also fosters the creation of excitatory links between threat-
related information and the large number of congruent cognitions that are
highly accessible in the memory network. As a result, this information is
encoded at a deep level of processing, which seems to be further exacerbated
by mental rumination on threat-related material. In contrast, positive
information may be encoded in a shallow way, and then easily be forgotten
or recalled in an inaccurate manner.
Hyperactivating strategies have negative implications for anxious persons’
self-image. Overdependence on a partner as the main source of comfort
interferes with the formation of a sense of confidence in one’s own abilities
and skills. Moreover, the intensification of threat appraisals causes attention
to be directed to self-relevant sources of distress (e.g., thoughts about
personal weaknesses, memories of personal failures), thereby fostering
chronic doubts about self-worth. Frequent activation of negative
cognitions and deep-level encoding of congruent information also
heighten the chronic accessibility of negative self-relevant thoughts, further
encouraging the formation of an undifferentiated negative self-image.
Beyond these automatic processes, more controlled attempts to achieve
some sort of closeness and love can add to a negative self-image—the
82
MIKULINCER AND SHAVER
anxious person may emphasize his or her helplessness and present him- or
herself in a negative or needy way to elicit others’ compassion and support.
Through three convergent processes, hyperactivating strategies encourage
negative appraisals of others: (1) The exaggerated appraisal of attachment
figures’ unavailability contributes to the formation of a negative global view
of others. (2) The frequent activation of negative cognitions primes negative
thoughts about others’ intentions and responses and makes these thoughts
accessible for social judgments. (3) The search for proximity and fusion
leads anxious individuals to project their negative self-views onto
relationship partners in order to create some sort of consensus and illusory
closeness. This projective tendency, which object relations theorists (e.g.,
Klein, 1940) call ‘‘projective identification,’’ results in a negative view of
others and an overestimation of self–other similarity. In this way, a negative
view of others may be guided, in part, by the search for connectedness and a
person’s own negative self-views.
Chronic reliance on hyperactivating strategies also places anxious
individuals at risk for a variety of emotional and adjustment problems.
These strategies impair the ability to regulate negative emotions and lead to
intense distress, which continues even after actual threats subside. As a
result, the person experiences an endless and uncontrollable flow of nega-
tive thoughts and moods, which in turn may lead to cognitive disorgani-
zation and, in certain cases, culminate in psychopathology. Specifically,
hyperactivating strategies can promote chronic attachment-related and
attachment-unrelated anxieties, strong depressive reactions to actual or
potential interpersonal losses, and intrusive symptoms following traumatic
events (e.g., unwanted trauma-related thoughts). Moreover, problems in
emotion control may be manifested in anger outbursts, impulsive behaviors,
and the development of more severe personality disorders.
Hyperactivating strategies also have several negative implications for
relationship quality. First, these strategies lead a person to adopt a
dependent and needy role that impairs the formation of a mature reciprocal
relationship. Second, they lead the person to feel chronically frustrated due
to the unfulfilled need for demonstrations of love and commitment. Third,
hyperactivating strategies contribute to a catastrophic appraisal of
interpersonal conflicts, the perpetuation of the resulting negative affect,
and conflict escalation. Fourth, these strategies may lead partners to feel
abused by a person’s endless demands for security and frequent suspicion
and distrust; engulfed by the person’s desire for merger; and controlled by
his or her clinging behavior and hypervigilance. All of these negative
feelings may lead partners to take distance from the demanding person
and reject his or her bids for proximity, which in turn may intensify the
person’s insecurities and worries. In this way, a self-amplifying dyadic
THE ATTACHMENT BEHAVIORAL SYSTEM 83
cycle of dissatisfaction is created which impairs relationship quality and
stability.
Of course, this cycle also depends on the functioning of the partner’s
attachment system. For example, a securely attached partner can be more
tolerant of a person’s demands for security, and this tolerance may reduce
relational tensions. We will return to this point in a later section in which we
extend our intrapersonal model of attachment-system functioning to a
systemic model that takes into account the attachment systems of two
interacting partners.
Hyperactivating strategies also inhibit the activation of other behavioral
systems. People who rely on these strategies are so focused on the seeking of
protection and security that they have insufficient resources for engaging in
nonattachment activities. Such people egocentrically perceive others as a
source of comfort and are unable to perceive them as partners for
exploration, sex, and affiliation, or as human beings who may also need help
and care. They also cannot attain the relatively calm and secure state of
mind that is necessary for engaging fruitfully in many nonattachment
activities (Weiss, 1998).
2. Deactivating Strategies
When proximity seeking is not perceived as viable, people tend to squelch
their quest for proximity and inhibit the activation of the attachment system
through the use of deactivating strategies (Cassidy & Kobak, 1988). These
strategies include the denial of attachment needs and the pursuit of excessive
self-reliance. They also include downplaying threats and blocking the
monitoring of attachment-figure availability, because every thought about
threats or attachment figures can reactivate the attachment system. In the
long run, repeated episodes of attachment-figure unavailability combined
with a negative appraisal of proximity-seeking viability lead to the
consolidation of deactivating strategies as the main regulatory device. At
the interpersonal level, this internal process is manifested in an avoidant
style. We view deactivating strategies as the underlying core characteristics
of avoidant persons.
Deactivating strategies consist of negative core beliefs about the
relationship partner as a source of protection, combined with positive
beliefs about the self as capable of dealing with threats. During their
interactions with unavailable attachment figures, individuals relying on
deactivating strategies have probably experienced serious doubts about the
management of distress, the partner’s good will, and the self’s capacity for
exerting control over the course and outcome of threatening events.
However, they must suppress any recognition of personal lack of control,
84
MIKULINCER AND SHAVER
because only a powerful and competent self can handle distress alone.
Moreover, such recognition entails the appraisal of the self as a source of
threat and the consequent unwanted reactivation of the attachment system.
As a result, deactivating strategies include heightened perceptions of self-
efficacy.
At the interpersonal level, deactivating strategies represent an inhibition
of the primary attachment strategy. People who chronically rely on
deactivating strategies have two main goals in social interactions and close
relationships: (1) pursuit of distance, control, and self-reliance, and (2)
avoidance of negative emotional states that demand attachment-system
activation. The first goal is manifested in active attempts to maximize
cognitive, emotional, and physical distance from a partner; avoidance of
interactions that demand emotional involvement, intimacy, interdepend-
ence, and self-disclosure; and suppression of attachment-related thoughts
and emotions that imply a sense of relational closeness, cohesion, and
consensus. The second goal is manifested in reluctance to directly or
symbolically confront relational tensions and conflicts; unwillingness to deal
with a partner’s distress and needs for proximity and security; and
suppression of thoughts and emotions related to rejection, separation,
abandonment, and loss.
In coping with threats, deactivating strategies lead to the adoption of
what Lazarus and Folkman (1984) called ‘‘distancing coping’’—cognitive
and behavioral maneuvers aimed at preventing active confrontation with
threats and intrusion of threat-related thoughts into consciousness. These
strategies consist of dismissal of threats; suppression of threat-related
thoughts and feelings; repression of painful memories, diversion of attention
away from the threat-related thoughts; and withdrawal of problem-solving
efforts. They also consist of detachment from challenging and emotional
interactions that potentially constitute sources of threat. These maneuvers
inhibit the experience of distress, exclude threat-related cognitions from
awareness, and thereby reduce the need for attachment-system activation.
Deactivating strategies are also manifested in the organization of the
cognitive system. The inhibitory circuits established as part of these
strategies lower the accessibility of threat- and attachment-related cogni-
tions and inhibit the spread of activation from these cognitions to other,
different cognitions. These circuits also create difficulties in encoding
material that is congruent with the excluded cognitions. Attention is
diverted from threat-and attachment-related information, and when
encoded this information is processed at a shallow level because it has no
strong excitatory association with the accessible elements of the memory
network. As a result, this information is not fully integrated with other parts
of the network and is encapsulated in segregated structures. The resulting
THE ATTACHMENT BEHAVIORAL SYSTEM 85
fractionated cognitive architecture resembles what Bowlby (1980) and
George and West (2001) called ‘‘segregated systems.’’
These procedural rules have important implications for avoidant people’s
self-image. The inhibitory circuits of deactivating strategies divert attention
away from self-relevant sources of distress, and can therefore inhibit the
appraisal of negative aspects of the self, suppress cognitions about personal
weaknesses and imperfections, and repress memories of personal failures.
In this way, deactivating strategies lead to the exclusion of negative
self-aspects from consciousness and a conscious sense of high self-esteem.
This defensive inflation of self-esteem is further reinforced by the
adoption of a self-reliant attitude, which requires enhancement of self-
worth, and by strategic attempts to convince others that one does not need
their support.
This defensive maintenance of a positive self-image seems to be achieved
at the expense of negative beliefs about others. These negative perceptions
result from the inhibitory circuits running from deactivating strategies to the
monitoring and appraisal of attachment-figure availability. These circuits
lead people to divert attention away from any attachment-related infor-
mation, including information about others’ positive traits, intentions, and
behaviors. As a result, this information is processed in a shallow manner, is
easily forgotten or biased, and is not accessible when making social
judgments, which results in the person maintaining a negative and inflexible
image of others. In addition, the defensive exclusion of negative information
about the self and attempts to maintain interpersonal distance facilitate the
projection of the excluded information onto a relationship partner, thereby
reinforcing the basic negative view of others. This projective tendency,
which Freud (1957/1915) called ‘‘defensive projection,’’ results in a negative
view of others and an overestimation of self-other dissimilarity. Thus,
negative perceptions of others related to deactivating strategies may be
guided, in part, by a preference for distance and a desire to view the self
more positively than others.
Deactivating strategies can be a source of emotional and adjustment
problems. Although these strategies lead to the defensive maintenance of a
faca,de of security and calmness, their inhibitory circuits leave the suppressed
distress unresolved and impair the ability to directly confront life’s
adversities. This impairment is particularly likely to be manifested during
prolonged, highly demanding stressful experiences that require active
confrontation of the problem and mobilization of external sources of
support. In these cases, deactivating strategies can produce a sense of
inadequacy in coping with stress, a marked decline in functioning,
constricted affect, and what Horowitz (1982) called ‘‘avoidance-related’’
posttraumatic symptoms (e.g., psychic numbing, behavioral inhibition,
86
MIKULINCER AND SHAVER
counterphobic activities). All of these symptoms reflect the inadequacy of
deactivating strategies under certain conditions and their failure to restore
emotional equanimity. In addition, although deactivating strategies involve
suppressing the conscious experience and display of distress, the distress can
still be indirectly manifested in somatic symptoms, sleep problems, and
other health disorders. Moreover, the negative attitude toward close
relationships can channel the unresolved distress into feelings of hostility,
loneliness, detachment, and estrangement from others.
Deactivating strategies also have negative implications for relationship
quality and stability. First, they lead a person to be emotionally detached
from a partner and to form superficial, cool relationships that lack the
vitality and bonding power of affection and intimacy. Second, the tendency
to avoid confronting relational problems may leave conflicts unresolved and
increase a partner’s irritation and resentment. Third, deactivating strategies
may lead partners to feel frustrated and dissatisfied due to the constant
rejection of their bids for intimacy and affection as well as the person’s
dismissal of, and nonresponsiveness to, their signs of distress. As a result,
relationship satisfaction can decrease and the likelihood of relationship
dissolution can increase.
Deactivating attachment strategies maintain inhibitory links with other
behavioral systems. They block cognitive exploration and the incorporation
of new evidence and unusual thoughts, because such thoughts may challenge
prior knowledge and lead to uncertainty, tension, and confusion. In fact,
relaxed exploration and ‘‘loosening’’ of cognitive operations can open the
door to threats and dangers, which activate the attachment system. Thus,
deactivating strategies favor cognitive closure and rigidity, and block the full
activation of the sexual and affiliation systems, because mutual exploration
of sexual pleasures with a lover and having flexible, unscripted fun with
friends can lead to the sort of intimacy and closeness that avoidant persons
are not comfortable dealing with. These strategies also block the activation
of the caregiving system, because empathic responsiveness to others’ needs
entails emotional involvement, acknowledgment of others’ distress, and
dealing with the closeness that this empathic reaction implies. The demands
of caregiving work against the goals of deactivating strategies—to distance a
person from all sources of suffering and all kinds of closeness to others.
D. SUMMARY
Our model outlines the cognitive operations, response strategies, and
dynamics of the attachment system in adulthood. It also describes the goals
of each attachment strategy and their psychological manifestations and
THE ATTACHMENT BEHAVIORAL SYSTEM 87
consequences. Whereas the goals of security-based strategies are to form
intimate relationships, to build a person’s psychological resources, and to
broaden his or her perspectives and capacities, the goal of secondary
attachment strategies is to manage attachment-system activation and reduce
or eliminate the pain caused by frustrated proximity-seeking attempts.
Hyperactivating strategies keep the person focused on the search for love
and security, and constantly on the alert for threats, separations, and
betrayals. Deactivating strategies keep the attachment system in check, with
serious consequences for cognitive and emotional openness. This framework
serves as our ‘‘working model’’ for understanding the activation and
functioning of the attachment system in adulthood. It also provides a
framework for reviewing our research findings, which is the mission of the
next section.
III. Empirical Assessments of the Theory
In this section, we review our research program on adult attachment style.
Our studies have examined the three different modules of our model and
have empirically assessed our ideas concerning (1) attachment-system
activation, (2) attachment-figure availability, (3) the interpersonal and
intrapersonal manifestations of attachment strategies, and (4) imp