http://ccp.sagepub.com/
Psychiatry
Clinical Child Psychology and
http://ccp.sagepub.com/content/14/3/389
The online version of this article can be found at:
DOI: 10.1177/1359104509104048
2009 14: 389Clin Child Psychol Psychiatry
Patricia McKinsey Crittenden and Rudi Dallos
All in the Family: Integrating Attachment and Family Systems Theories
Published by:
http://www.sagepublications.com
can be found at:Clinical Child Psychology and PsychiatryAdditional services and information for
http://ccp.sagepub.com/cgi/alertsEmail Alerts:
http://ccp.sagepub.com/subscriptionsSubscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.com/journalsPermissions.navPermissions:
http://ccp.sagepub.com/content/14/3/389.refs.htmlCitations:
What is This?
- Jun 10, 2009Version of Record>>
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
All in the Family: Integrating Attachment and
Family Systems Theories
PATRICIA MCKINSEY CRITTENDEN
Family Relations Institute, USA
RUDI DALLOS
University of Plymouth, UK
A B S T R AC T
This article brings together ideas from attachment and systemic family therapy.
There is both growing interest among systemic practitioners in the conceptual and
empirical base of attachment theory and also the need for attachment theory to
expand dyadic patterning to include its context in family functioning. We propose
the Dynamic-Maturational Model (DMM) as being the most compatible and
useful variant of attachment theory. With its emphasis on the functional nature of
behavior, a dynamic view of development and change, and a focus on multiple
attachments and representational systems, the DMM fits systemic concepts well.
We propose that many apparent discrepancies between the theories will disappear
if careful distinctions are made between observed behavior, functional explanations,
and attributions. We conclude with theory-based recommendations for
selecting treatment strategies. Several case examples that are theory based,
counterintuitive, and tied to disorders that are difficult to treat are offered to give
substance to our ideas.
K E Y WO R D S
attachment, dynamic maturational model, family systems, family therapy
AT TAC H M E N T T H E O RY A N D FA M I LY S YS T E M S theory each offer the other crucial
areas of knowledge and expertise that, when combined, may yield a more comprehensive
and effective basis for treatment. Melding the theories is logical because the theories
share substantial common ground. Both conceptualize human functioning in systemic
terms, avoiding symptom-based and pathology- or disease-oriented definition of
problems, understand distress in functional terms, and describe patterns of dyadic and
family functioning that are compatible. Attachment theory brings to this common
ground (a) a clear focus on the functions underlying presenting problems, (b) extensive
empirical work on developmental processes, including the development of intellect,
affect, representation, and identity, and (c) systematic methodologies that translate
Clinical Child Psychology and Psychiatry Copyright © The Author(s), 2009.
Reprints and permissions: http://www.sagepub.co.uk/journalsPermissions.nav
Vol 14(3): 389–409. DOI: 10.1177/1359104509104048 http://ccp.sagepub.com
389
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
naturalistic observation into empirically validated assessment. Family systems theory
(FST) brings a focus on complex family structures and patterns of behavior, description
of family dynamics in troubled families, and experience in treatment strategies and
change processes. One might say that attachment theory addresses the developmental
and diagnostic ‘front end’ of dysfunction more thoroughly whereas family systems
theory addresses the complexity of development gone awry and treatment processes.
Together attachment and family systems theories could yield an approach to family
problems that is (1) both focused and flexible, (2) developmentally sound, (3) respectful
of both individual perspectives and also those of dyads and larger family units, (4)
amenable to assessment without pathologizing family experience, (5) relevant to prevention,
and (6) structured to promote healing and adaptation.
In the following sections, we address attachment theory, then family systems theory
and therapy, and finally ways in which we think the theories can be integrated to improve
treatment. We begin with attachment theory because its strength has been early development
and dyadic relationships. We proceed to review contributions from family systems
theory, thus expanding from attachment theory’s emphasis on children to FST’s focus
on families with children and from attachment theory’s focus on dyadic processes to
family systems theory’s focus on more complex family units. This leads to the application
of the full set of ideas from both theories to treatment.
It should be noted that,like other maturing theories,both attachment theory and family
systems theory have variants. The variants within attachment theory tend all to agree on
a set of ideas drawn from the work of Bowlby and Ainsworth, but their further development
heads in incompatible directions. Thus, to discuss current attachment theory, certain
fundamental choices must be made. On the other hand, the variants within family systems
theory tend to emphasize different aspects of family functioning (e.g. structural, strategic,
solution focused, narrative) without these being inherently incompatible with one
another. This facilitates exploration of how the ideas generated by FST fit together and
can be brought forward into a broader integration with attachment theory.
Contributions from attachment theory
The roots of attachment theory
Attachment theory was developed by John Bowlby (1969/1982; 1973, 1980) out of
concerns for children, particularly juvenile delinquents and those separated from their
CLINICAL CHILD PSYCHOLOGY AND PSYCHIATRY 14(3)
390
AC K N OW L E D G E M E N T S : This article combines and expands the papers presented by the
authors in the symposium ‘Attached to Our Childhoods: Integrating Attachment and
Systemic Theories in Practice’ (Arlene Vetere and Rudi Dallos (Chairs)), at the 5th Congress
of the European Family Therapy Association, Berlin, Germany, 30 September 2004.
PAT R I C I A M . C R I T T E N D E N is an Independent Researcher, Trainer and Consultant in
developmental psychopathology and innovator of the DMM model of attachment.
C O N TAC T : Patricia M. Crittenden, 9481 SW 147 St., Miami, FL 33176, USA.
[E-mail: pmcrittenden@att.net]
RU D I DA L L O S is Programme Director and Professor of Clinical Psychology in the Department
of Clinical Psychology, University of Plymouth. He is also a practising Clinical Psychologist
and Family Therapist working with children and adolescents and their families in
the NHS and social care.
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
parents during World War II. To understand these diverse conditions, Bowlby expanded
his training in psychoanalytic theory to include ethological theory, general system theory,
and emerging theory on information processing. He proposed that the roots of psychological
disorder lay in the unavailability of protective attachment figures and that adults’
recall of trauma referred to real events, albeit possibly recalled with distortion. His
concern for children set attachment theory apart from other theories of psychopathology
by placing development ahead of dysfunction, rather than trying to explain retrospectively
how adults’ dysfunction might have occurred.
Mary Ainsworth applied Bowlby’s theory in two ground-breaking studies. The first
used naturalistic observation of mother–infant dyads in Uganda and helped to establish
both the importance of individual differences in maternal behavior and also the applicability
of attachment theory across diverse cultures (Ainsworth, 1967). The second used
naturalistic home observation of American mothers and babies over the first year of life,
followed by systematic observation at 11 months in a procedure that would become the
‘gold standard’ of attachment, the Strange Situation (Ainsworth, Blehar,Waters, & Wall,
1978). Ainsworth’s work underscored the importance of detailed description of behavior
and, in addition, provided attachment theory with (a) a firm developmental perspective,
(2) a tradition of empirical work preceding applications, (c) an assessment methodology
(the Strange Situation), and (d) the ABC patterns of attachment.
Later work demonstrated that differences in mothers’ sensitive responsiveness at
home were related to differences in infants’ attachment to their mothers at 1 year of age
as well as to differences in children’s later development. Specifically, emotionally
rejecting mothers, that is, those who were displeased by infants’ negative affect, tended
to have infants who turned away from them when reunited after a brief separation.
These infants were described as using a Type A strategy of inhibiting the display of
negative affect. Mothers whose response was unpredictable tended to have infants who
were emotionally labile and mixed in their feelings about closeness to the mother
(both desiring it and not). These infants were described as using a Type C strategy of
exaggerated display of mixed negative affects. The third group of mothers was both
warm and predictably protective of their children. After a brief separation, their infants
sought close bodily contact and then explored near the mothers. These infants used a
Type B strategy of showing affect without inhibition or distortion.
A number of researchers have modified attachment theory to fit the behavior of populations
at high risk, for example, of neglect, abuse, suicide, extreme emotional distress
and older children and adults. The most widely used approach, Main’s ABCD model
(Main & Solomon, 1986), adds disorganization (Type D) as a fourth category. Empirical
studies show that Type D is associated with risk, but it does not differentiate risk from
nonrisk well and does not differentiate among types of risk (Crittenden, Claussen, &
Kozlowska, 2007; Solomon & George, 1999, 2008; Spieker & Crittenden, in press).
One of us (PMC) took the opposite approach of looking for additional organizations
that helped children to protect themselves from severe threat. In my dissertation, carried
out under Ainsworth’s guidance, the threats were child abuse and neglect and the new
organization was an alternation of A and C strategies (A/C; Crittenden, 1983). Applying
this approach to older children and adults led to the Dynamic Maturational Model of
attachment and adaptation (DMM; Crittenden, 1995, 2006, 2008). The DMM is based
not only on attachment theory, but also on my training as a behavioral and family
systems therapist. We propose that the DMM is more clinically applicable than the
ABCD model and also more attuned to the systemic structure of FST. For these reasons,
we use it as the basis for integration of attachment with family systems theory and
therapy.
CRITTENDEN & DALLOS: INTEGRATING ATTACHMENT
391
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
The Dynamic-Maturational Model of attachment and adaptation
The DMM applies attachment theory to the experience of endangered people. Bowlby’s
thinking about separation and loss is expanded to include all kinds of danger, including
physical, emotional and sexual harm, and Ainsworth’s ABC patterns of attachment
are seen as self-protective strategies learned in the context of particular attachment
relationships.
The DMM and danger Protection of self and progeny is a biological imperative. It is
seen in the DMM as the central goal of psychological and behavioral functioning. That
is, if a sufficient number of individuals do not live, reproduce, and protect their progeny
to reproductive maturity, not only will the individuals die, but in addition, the species
will become extinct. These functions reintroduce sexuality into the notion of adaptation
and place adaptation in an interpersonal context.
Attachment as interactive and dynamic It is ironic that Bowlby’s discovery of systemic
theory and Ainsworth’s focus on dyads’ reciprocal organization over time has devolved
into a broad understanding of attachment as a stable personal characteristic. In the
DMM, it is quite the opposite. ‘Attachment’ is always a relationship term, it always
implies reciprocal processes,and it is ever-changing (dynamic!). Indeed,when it becomes
too personal (i.e. not adapting to new people) and too stable (i.e. not adapting to new
situations), it is almost always maladaptive and less functional in terms of protection and
reproduction.
Expanding the array of self-protective strategies The DMM gives neurological and physical
maturation a central role in children’s ability to construct strategies to cope with threat.
Thus, the DMM offers a developmentally increasing array of possible strategies. These
begin with Ainsworth’s strategies for infants, adding increasingly complex strategies as
children mature. These strategies retain the Ainsworth’s ABC nomenclature and carry
the labels used by Bowlby whenever he discussed the pattern first (e.g. compulsive caregiving,
compulsive promiscuity, compulsive self-reliance; see Figure 1). The DMM adds
a connecting theoretical structure that (a) suggests a greater array of patterns and (b)
explains how these are tied to information processing.
Information processing and strategies Information processing underlies the strategies.
Findings from cognitive and trauma research inform understanding of individual differences
in self-protection and adaptation. In the DMM, physiology, cognition and affect
are three ways of representing experience, each suggesting a response that might protect
the self or one’s progeny. Somatic information includes genetic, epigenetic, cellular, and
organic functioning. Cognition is based on repeated temporal sequences. It operates on
principles described by learning theory. Affect is based on intensity of stimulation with
arousal-elicited feelings becoming associated with dangerous and safe contexts. When
the three representations are congruent, action proceeds unimpeded. When they are
different, the individual faces a dilemma: act quickly on one or delay long enough to
resolve the discrepancy. Quick action creates risk of error; careful thinking creates risk
of danger.
Transformation and representation Physiology, cognition and affect can be omitted from
processing (physiology: ‘No, I didn’t feel anything’; cognition: ‘I can’t recall’, ‘I don’t
know’; affect: ‘It hurt my mother to see me bleeding’). All three can be distorted by
exaggeration (always, never, very, very, etc.) or minimization (a bit, not really, etc.), and
CLINICAL CHILD PSYCHOLOGY AND PSYCHIATRY 14(3)
392
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
all can be falsified and used deceptively. False somatic states, false cognition and false
positive affect can leave others unprepared if one acts on the hidden information. These
transformations are used to organize the various behavioral strategies as shown in
Figure 1.
Physiological states (somatic states), cognition and affect can be represented in many
ways (see Figure 2). Somatic information can be represented (a) implicitly in organic
states, that is, a knotted stomach or skin eruptions or (b) explicitly and verbally in ‘body
talk’ about the physical state of the self; somatic representation is sometimes the only
sign of trouble in an otherwise ‘adapted’ person, one with psychosomatic distress.
Cognitive information can be represented (a) implicitly and nonverbally in action,that
is, it can be enacted preconsciously (procedural memory), or (b) explicitly and verbally
in generalized semantic statements. Procedures are routines that one carries out repeatedly
without conscious awareness, that is, they are ‘enactments’ in FST terminology.
Semantic memory consists of contingencies and expectations drawn from procedural
regularities. The basic semantic form is a when/then or if/then statement such as, ‘when
my mother has a headache, she becomes angry. If I pester her, then she will scream at
me’. Often, however, the descriptive quality of semantic memory is distorted to a
prescription, what one ought to, should or must do. Too great a focus on prescriptive
CRITTENDEN & DALLOS: INTEGRATING ATTACHMENT
393
Figure 1. A Dynamic-Maturational model of self-protective strategies.
B3
A/CA/CA/C
B4–5B1–2
A1–2
A3–4
A5–6
A7–8
C1–2
C5–6
C3–4
C7–8
AC
Psychopathy
Menacing/
Paranoid
Punitive/
Seductive
Aggressive/
Feigned Helpless
Threatening/
Disarming
Comfortable
False Positive Affect
Integrated False Information
(Tpe AC) False Cognition
True Cognition
IntegratedTrue Information
(Type B) True Negative Affect
DistortedAffect&OmittedCognition
DistortedCognition&OmittedAffect
ReactiveReserved
Socially Facile/
Inhibited
Compulsively
Caregiving/
Compliant
Compulsively
Promiscuous/
Self-Reliant
Delusional
Idealization/
Externally
Assembled
Self
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
semantic statements can lead to discrepancies between what one says and what one does.
This can lead to confusion for others to be able to understand and to predict the person’s
actions and also difficulty for the person in self-awareness and in managing relationships
and decision-making processes.
Affect can be represented (a) implicitly in sensory images (visual, auditory, gustatory,
tactile, and olfactory images) of one’s context, or (b) explicitly and verbally in connotative
language. For example, certain smells (your mother’s perfume, alcohol on your
father’s breath) can trigger powerful feelings of comfort or threat. Use of evocative
words (e.g. detest, pighead) can both convey intense feelings and also elicit feelings in
others.
These six forms of information can be integrated to create narrative episodes. In
addition, all can be made conscious and reflected upon to generate new and more
accurate information. Reflective integration enables us to resolve discrepancies, reduce
distortions of information, and adapt our response to our current context. However,
because verbal processing and reflective integration proceed slowly, threatened people
less often use verbal representations and may resist reflecting.
Development and transformations Only the simplest transformation, omission, can be
made in infancy whereas distortion and falsification take greater maturation. Verbal
representations cannot be generated until after the third year of life. Reflective functioning
first becomes possible at about 7 years of age, but is not fully mature until the
mid-30s. Consequently, children’s representations are necessarily constrained as
compared to those of adults. Management of complex or ambiguous situations requires
the assistance of attachment figures. This occurs through open communication in which
CLINICAL CHILD PSYCHOLOGY AND PSYCHIATRY 14(3)
394
Sensory Stimulation
Physiology Temporal Order Intensity
Soma Cognition Affect
Organic States Procedural Memory Imaged Memory
Verbalized ‘Body Talk’ Semantic Memory Connotative Language
Episodic Memory
Reflective Integration
Figure 2. Representational systems as a function of somatic, cognitive and affective information
organized in both developmental order and also order of psychological complexity.
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
positive and negative feelings can be processed and also through scaffolding by the
parents to assist children to understand increasingly complex situations. To do this,
parents need to be able to see things from the child’s perspective, communicate that they
understand the child and offer alternative perspectives and elaborated explanations
attuned to the child’s readiness to develop complex understandings.
Information processing and strategies Variations in information processing combined with
Ainsworth’s ABC strategies generate a psychological model of the behavioral strategies.
Type A strategies are more ‘cognitive’ in that they rely more on experienced temporal
contingencies than on negative feeling states to predict outcomes and organize behavior.
Type C, on the other hand, involves organizing strategic behavior on the basis of current
feelings states. Type B involves flexibility in being able to choose between cognition or
affect to organize behavior. This implies taking the context into account when considering
the potential appropriateness of strategies. When somatic information is not integrated
into the strategy (A, B, or C), somatic expressions are often observed.
An important attribute of this model is the transformation of a categorical model into
a dimensional model (with the horizontal dimension being extent of integration of
physiology, cognition and affect and the vertical dimension being the degree of transformation).1
With a dimensional model, there can be varied presentations of the strategies
as well as intermediate strategies. For example, between A1–2 and A4 (compulsive
compliance), we can describe compulsive performance (A4, fearing only loss of love and
approval and not punitive punishment). This gives the DMM greater flexibility to
describe and explain variations in human behavior.
In the DMM, no strategy is necessarily the best strategy for a particular problem. It
is the interaction of strategy with circumstance that defines the use of a strategy as
adaptive or maladaptive. On the other hand, because life circumstances invariably
change, flexibility of strategy and appropriateness of strategy selection become crucial;
these are the advantages of the mature Type B strategy.
Danger, information processing and strategies Danger heightens the intensity of learning.
However, when learning takes place under dangerous circumstances, the learning may
be limited and selective, that is to say, at risk of distortion. That which is learned is given
priority and will be relinquished very reluctantly, especially when threat is perceived.
Consequently, when children are not sufficiently protected, they learn about danger, but
they learn within the limitations of their maturation. In other words, they learn, but in a
distorted manner. When conditions are constant, this ‘immature’ strategy usually
functions adaptively. However, when conditions change, for example, through maturation,
the strategy may become maladaptive. Major change periods include turning
2 years old, beginning school, puberty, the transition to adulthood, and becoming a
parent. Each is marked by an increase in referrals to psychotherapy as well as a change
in the sorts of dilemmas that children face and consequently the types of problems that
are presented.
Assessment One strength of attachment theory is that its central constructs have been
operationalized in assessments (e.g. the Strange Situation, the Adult Attachment
Interview). The assessments (1) are developmental (with different assessments for
infancy, toddlerhood, the preschool years, adolescence and adulthood); (2) are drawn
from careful observation (as opposed to self-report); (3) use regulated mild danger to
elicit the strategy; and (4) are always being modified as discrepancies are identified, for
example in cases that do not fit the expected patterns. These assessment have been
CRITTENDEN & DALLOS: INTEGRATING ATTACHMENT
395
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
used to validate theory, provide guidance to therapists and courts, and generate new
information.
Taking the DMM to the family level Families are the primary means by which individuals
accomplish the functions of protecting the self, reproducing, and protecting the next
generation. When these functions are jeopardized, strategic functioning is activated.
Knowing in detail how this occurs could help clinicians to cut through the screen of
symptoms, problems, and complaints that family members bring to treatment to find
their roots in threats to these essential functions.
Some early attempts were made at connecting constructs employed to describe
family configurations with the dyadic strategies described by Ainsworth and colleagues
(Hillburn-Cobb, 1998). However, such attempts have been limited in number and so
far have produced mixed findings. One reason may be that the families are highly
complex and an initial starting point should be to limit the focus to processes in triads,
which some systemic family therapists regard as the fundamental building block of
family life. That said, the DMM notion would be that all levels of organization (from
genetic through interapersonal, dyadic, triadic, familial, etc., to political, cultural and
societal) are crucial, with none being privileged over the others. Instead, our choice of
focus might best be explained by the availability of ‘technology’ to study each type
of organization.
A crucial aspect of moving to the family level is to recognize that families are seen in
terms of relationships which operate in terms of mutual influence. They are composed
of individuals in an array of dyadic and triadic relationships, with each of these consisting
of patterns of interaction that are continuing, recursive, mutually influencing patterns
(reciprocal causality). Each of these relationships can be conceptualized both in terms
of the individual experiences of each member and also in terms of the relationship
between them.
A contribution of systemic family therapy has been that it is possible to describe
families not only in terms of the sum of their individual experiences, but also in terms
of the patterning of their relationships. For example, Watzlawick, Beavin, and Jackson
(1967) and Watzlawick, Weakland, and Fisch (1974) described how dyads could show
patterns of complementary and symmetrical escalations. They also suggested that any
dyad was potentially unstable and needed the input from a third person to achieve
stability. The implications of this for single parents with one child could be powerful,
informative, and unsettling for current family policy.
The shift in level of organization from individuals to dyads to families means that we
cannot apply the same intrapsychic constructs, such as thinking, feeling and reflecting, to
families. For example, when a parent speaks to all the children at once, each child hears
the parent separately and assigns meaning to the communication based on their unique
development and history with the parent. Thus, one communication may affect many
relationships and result in several different responses. Nevertheless,we can see how their
level of disagreement, polarization of their explanations, intensity of their emotional
reactions and so on may escalate and increase. A similar situation can be seen in terms
of a family therapist whose behavior is perceived differently by each family member.
However, what can also be seen is that there may be an increasing polarization, or
alternatively increasing agreement about how the therapist’s communications are
viewed by the family. An important attachment feature is that the level of trust in the
therapist by family members may show an overall increase as the therapeutic relationship
– a secure base – develops, or it may decrease if the array of individual and dyadic
reactions is not adequately accommodated by the therapist.
CLINICAL CHILD PSYCHOLOGY AND PSYCHIATRY 14(3)
396
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
As an example of utilizing individual, dyadic and family lenses, we consider the case
of some professional women who experience postnatal depression. The women may
experience a destabilizing change of role, including anxiety that they have to depend on
partners or family when they had previously striven to be independent and self-sufficient.
Women in these circumstances are often given medication or CBT, either of which might
relieve their symptoms. Our observations of dyadic interactions suggest that sometimes,
after treatment, the mother’s sad mood will have become false positive affect and her
semantic statements will have changed from expectations of futility to false prescriptive
statements of enjoying mothering. The proof of the success of the treatment is in the
baby’s behavior and in the relationship between the mother and the infant. We have
found in many such cases that before treatment the babies looked sad and avoided eye
contact with their distant and affectless mothers. After treatment, the mothers looked
bright, but the babies were unchanged and there were no mutual, reciprocal and positive
exchanges between the babies and their mothers. Moreover, the mothers’ positive
semantic statements were often inaccurate. For example,one mother commented brightly
on how her silent, sad baby was ‘so chatty today’. This mother performed like a good and
happy mother, but she did so without reference to her baby. Some of these mothers used
a strategy of compulsive performance, trying very hard to please and excel in what was
expected of them. After childbirth, they felt threatened by their inability to excel in all
their roles concurrently and this underlay their depression.
Individual treatment may relieve the symptoms of depression without resolving the
underlying problem. Likewise, mother–infant treatment may only teach the mother to
act falsely bright in ways that make her feel like a good and competent mother. This
outcome can look, if one focuses on the mother’s behavior, as if the interactional process
has changed between infant and mother, but we may be deceived by false prescriptive
words and false positive smiles. Success at ameliorating the mother’s depression and
baby’s sad isolation depends upon mutuality, shared pleasure, and context-adapted
change in the dyadic process. We need to look for subtle cues in both individuals and in
how these fit – or do not fit – together at the dyadic level. If the mother’s change is only
a change in what behaviors she displays and not an interpersonal change, then someone
else must adapt to make the mother’s appearance of being happy sustainable.
Who? A child! Maybe the baby, maybe an older sibling, maybe both. How? By
becoming compulsive caregivers to the mother. That is, by themselves assuming a false
positive strategy, children (even infants) can accommodate their mothers’ limitations in
ways that permit the mothers to slide through this major life change without changing.
That is, without lowering their standards of perfect performance, such mothers can
perform ‘perfectly’ in old and new roles – if their children accommodate by asking for
little and giving care to their mothers instead.
Alternatively, the father might increase his caregiving of the mother and take over
some of her duties, thus allowing her to both retain her compulsive strategy and also
perform well. If, however, the father used a Type C strategy of expecting to receive caregiving
and resenting the mother’s focus on the baby, he might also increase his own
demands on her. If, in addition, the baby was too distressed to become a caregiver, then
possibly no one will change to accommodate the mother’s strategy and someone will
become symptomatic.
This example of a triangulating process suggests that the development of each individual
and the strategies used within each dyad may need to be addressed to resolve the
symptoms of one individual in a way that does not harm other family members. In
addition the dyads need to be seen as interwoven within a wider family pattern. Hence,
postnatal depression may be both a personal problem and also a family problem
CRITTENDEN & DALLOS: INTEGRATING ATTACHMENT
397
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
requiring formulation and therapeutic input at several different levels that exceed
individual treatment of the women.
Compulsive performance is not a ‘bad’ strategy and we need to understand the
pressures and conflicting demands that many women experience from a combination of
their personal history, familial and cultural expectations. Indeed, compulsive performance
generates many achievements that society needs and values. We reward such
performance well. But, as the example above indicates, it is not the best strategy for all
circumstances. Every strategy is adaptive in some context and maladaptive in others.
Instead of attempting to eliminate ‘bad’ strategies, we might want to make their useful
application explicit so that their use can be restricted to those situations. For example,
enabling new mothers who experience depression to discover that ‘good enough’ is
really enough when one has several important and competing priorities might enable
them to employ a more balanced strategy. However, when extreme self-protective
strategies have been employed to protect the self from childhood threats, they tend to
be applied too widely later. It is the inappropriate application of strategies that is
maladaptive and that leads to the need for psychotherapy.
These ideas from the DMM suggest that therapists’ attention should be directed
toward (a) resolving actual threats to survival and reproduction and (b) correcting
misperceived threats when these are the presenting problems. That is, DMM treatment
would be ‘solution focused’, but the nature of the actual problem might not be understood
exactly as the family members present it.
Family systems theory and therapy
Background
Both attachment and family systems theories were conceived in the postwar years when
limitations of psychoanalytic and learning theory were being articulated and when there
was an interest in systemic processes and the biological roots of behavior. Nevertheless
the theories have developed separately, because, although integrations have been
proposed (Byng-Hall, 1991; Erdman & Caffery, 2003; Johnson, 1996; Kozlowska &
Hanney, 2002), the two approaches have never quite merged. Instead, they seem to have
met occasionally, talked, noted some similarities, but also differences, and moved on to
develop in their own ways.
Systemic family therapy
Family therapy initially developed out of a variety of strands one of which was attachment
theory. In fact Bowlby has been hailed as one of the originators of family therapy
because he was one of the first clinicians to suggest the idea of meeting with all the
members of a family to help resolve their problems. In the 1950s, concepts from the
emerging discipline of cybernetics and systems theory were employed by a number of
therapists and researchers. Don Jackson (1957) offered a convincing analogy of how
troubled families could be seen as self-maintaining systems which had evolved to a point
of ‘homeostasis’ or ‘stuckness’ in their interactions, with problematic behavior functioning
to maintain the problem. This counter-intuitive formulation indicated that, despite
the family members’ protestations that ‘all would be better if they did not have the
problem’, they could be seen to behave in many ways that served to maintain the
problem.
The incorporation of ideas from systems theory into clinical work with families
came to be known as systemic family therapy (SFT). It offered a radical move to an
interpersonal, relational view of clinical problems and away from the intrapsychic and
CLINICAL CHILD PSYCHOLOGY AND PSYCHIATRY 14(3)
398
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
individual orientation of the preceding psychodynamic and behavioral models. Rather
than viewing maladaptation as resulting from individual deficit or pathology, problems
are seen as resulting from relationship processes (Dallos & Draper, 2005). One unifying
idea was ‘nonsummativity’, that the whole is greater than the sum of its parts, a view that
the functioning of families could not be reduced to the sum of their individual members.
Consequently, concepts such as family hierarchy, boundaries, processes, symmetrical and
complementary escalations and shared belief systems emerged as valid conceptualizations
to embrace the idea that the family processes were related to, but at a different
and distinct level of explanation than individual processes.
Attempts to find links between individuals’ internal states and displayed problems to
patterns of family dynamics were central to early systemic family therapy. Much effort
was devoted to finding such links, for example the double-bind theory (Bateson,1972),
Bowen’s (1971) notion of psychosis as resulting from trans-generational patterns shaping
the development of ‘a fragile, undifferentiated ego’, Minunchin, Rosman, and Baker’s
(1978) model of psychosomatic families and Palazzoli, Cecchin, Prata, and Boscolo’s
(1978) ideas of covert and incongruous communicational processes in families with
members with a diagnosis of anorexia and psychosis. These interests also led to the
development of various research-based models such as the circumplex model of family
life (Olson, 1989) which described families as being organized around two dimensions:
Cohesion and flexibility. These dimensions also embraced some early structural SFT
concepts, such as enmeshment and disengagement (Minuchin et al.,1978) and attempted
to identify which family patterns linked to which problems. It has been suggested
that these dimensions map unto findings from attachment research, for example
Hillburn-Cobb (1998) offers evidence of a correspondence between disengagement and
enmeshment and Types A and C, respectively.
Many of the pioneers of the SFT movement (e.g. Jackson, 1957; Minuchin et al., 1978;
Palazzoli, Cecchin et al., 1978) came from psychoanalytic backgrounds and maintained
a strong interest in linking individual experience, unconscious dynamics and ‘problems’
with family processes. A connecting thread was the idea of the ‘function’ of a symptom
for the family which was also linked to homeostasis in families. This reflected the psychodynamic
concepts of the ‘unconscious function’ of sympoms in that the symptoms shown
by one member of a family were seen as protecting the family from other greater
dangers. For example Palazzoli, Boscolo et al. (1978) described the destructive behavior
of a young teenage boy as an attempt to preserve the privileged emotional relationship
(attachment) that he had with his single-parent mother. This was seen as being threatened
by the arrival of a new man in her life. Ferreira (1963) proposed the concept of the
‘family myth’, a shared distortion of the situation in terms of a story about the need to
cure the symptom. Others, such as Minuchin et al. (1978) and Haley (1976), saw
problems as emerging attempts to solve family dilemmas which instead made matters
worse. For example, a couple’s solution of detouring their marital conflict through a child
could hold a solution of not forcing a showdown about their marriage with the risk of
potential separation, but had the negative effect of the child becoming psychosomatic
(which can be addressed as a somatic representation that is both nonverbal and also
unintegrated with outer representations; see Figure 2).
These approaches emphasized family structure, social roles and processes to make
sense of the development of particular problems. However, there was also an emphasis
on the role of cognition and meaning making in families. Watzlawick et al. (1964, 1974)
suggested that people’s different individual punctuations, their explanations of events in
a relationship, could lock together to produce mutually self-maintaining patterns. In his
classic example of a cycle of pursuing and distancing in a couple, one partner may
CRITTENDEN & DALLOS: INTEGRATING ATTACHMENT
399
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
perceive the other in terms such as ‘nagging’, emotionally suffocating and consequently
see themselves as having to ‘respond’ to their partner’s behavior by ‘withdrawing’,
seeking distance and taking time away from the other. In turn, their partner may see the
other as continually ‘withdrawing’ and organizes their behavior to seek contact, intimacy
and ‘pursuing’ them. Each partner can be seen to be making a personal choice, but it is
based on a partial understanding of their interactional cycle. Each omits the wider
systemic recursive picture of how their own actions trigger their partner’s actions, thus
capturing them both in a self-maintaining cycle. It is also important to note that the
description of such interlocking cycles of actions and beliefs tended, in the early systemic
literature (with the exception of Minuchin’s approach and some with psycho-dynamic
links), to overlook how emotions fuel such patterns. As examples, the couple mentioned
may fear the risk of losing each other as attachment figures, thus using cycles of pursuit
and withdrawal to embody core patterns of dismissing vs preoccupied attachment
strategies for each partner acquired from their childhoods.
These efforts at linking patterns of family dynamics with particular types of disturbance
were eventually critiqued as being reductionist, modernist and potentially family
blaming (Hoffman, 1985; Sluzki, 1983). Further, there was a distinct lack of empirical
evidence to support them. More widely, the focus on identifying possible connections
between family dynamic and distress led to such an outcry that legal injunctions were
taken out in the USA preventing research on family dynamics, for example, regarding
the potential links between family processes and schizophrenia. This is an extremely
sensitive and complex issue, but it is also salutary to note a raft of recent research that
has found powerful evidence of an extremely high rate of undetected physical, sexual
and emotional abuse in people treated for psychotic problems and,more broadly,of links
to a variety of severe psychological disturbances and trauma in families (Read, 2005;
Read, Agar, Argyle, & Alderhold, 2003). Perhaps family therapists are in a difficult
position here: we wish to work alongside families in nonblaming and neutral manners,
but at the same time we may need to acknowledge that family members sometimes also
act with each other in ways that result in unintended distress and trauma. Thus, a crucial
issue may be differentiating intentions from outcomes (Crittenden, 2008).
Looking at triangles
One of the most enduring ideas to emerge from SFT practice and research was the
notion of triangulation. Minuchin (1974), and Minuchin et al. (1978) argued that families
may engage in a variety of triangular processes: when parents were involved in conflict,
one common process is to detour their conflict by directing their attention onto a child,
perhaps focusing on and possibly amplifying his/her problems. It was argued that a child
could sense the distress and tension between the parents, for example, in raised voices,
and the resulting distress shown by the child would be a point of focus and serve to
distract the parents from their conflict. It could eventually become established as a way
of avoiding their difficulties. A second triangular process, cross-generational coalition,
was seen as when each parent tries to recruit a child unto their side against the other
parent, or alternatively where one parent shows excessive concern and involvement
(enmeshment) with a child while the other parent becomes emotionally distant (disengaged).
The focus on triangles can be an important bridge between systemic and
attachment perspectives. It allows us to see a child as functioning in both direct dyadic
relationships with each parent and also the relationship between them (Pallazoli, Cirillo,
Selvini, & Sorrentino, 1989). In effect, the child can be seen as having an attachment
strategy with each parent, but having those strategies function to meet parents’ needs in
their relationship without the child’s awareness. The situation for the child becomes
CLINICAL CHILD PSYCHOLOGY AND PSYCHIATRY 14(3)
400
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
increasingly complex and confusing, especially when there is no open discussion in the
family about what is going on, what people are feeling, what their intentions are and so
on. In such situations, children and their parents may become increasingly confused
about the causes of events and their own role in the problems.
We can illustrate some of the dilemmas through the experience of a young woman
who had been suffering with anorexia. RD worked clinically with her and her family and
also as part of a research study exploring shared attachment themes in the family (Dallos
& Denford, in press). Carla describes poignantly her experience of a cross-generational
coalition regarding her parents relationship and a sense of how this related to the
development of her eating problems:
The only thing I ever hear them talking about is me and if I didn’t have this
[anorexia] it’s kind of like, would everything fall apart, at least it’s keeping them
talking. And they won’t argue while I’ve got this because it might make me worse.
So um . . . that’s kind of bought, sort of like, I’m not in control as such but I’ve got
more control over the situation that way. (Dallos, 2006, p. 189)
Carla’s account illustrates how she felt herself to be tangled in the relationship and
conflict between her parents. Seeing her experience in terms of a triangular relationship
provides a powerful bridge between attachment and systemic perspectives. Attachment
theory emphasizes that children need and seek protection from danger. In Carla’s case,
each parent could offer a supportive relationship with her, nevertheless the relationship
between her parents negated this, providing a context of fear:
They used to hate each other so much I always used to be so scared that one of
them would do something stupid and I would come home and, I used to hate
coming home just in case something happened. And they’ve both got the worst
tempers, even dad . . . dad’s is rarely seen but it is really bad.
Her parents’ relationship left her feeling anxious and in danger such that her anxiety
came to pervade her dyadic interactions with each of them. Moreover, she frequently
felt drawn into taking sides:
They used to really hurt me because they used to play each other off . . . And they
would be like ‘Go on tell me all the bad stuff about the other one’. And I used to
sit there and think to myself I am made up of half of each of these people and they
hate each other and do they hate me? That used to play on my mind for ages when
I was really young and that was the limit of my thought, I didn’t analyse it further.
Moreover, despite showing considerable understanding of her situation, Carla does not
take her parents’ perspective. Possibly, being inside such a triangle, led to Carla’s organizing
a C5–6 strategy of insisting, at all costs, on her own perspective and fearing loss of
protection altogether if she bent and considered her parents’ perspectives (cf Ringer &
Crittenden, 2007). At other times, she showed an attempt to distance herself from
negative feelings, blaming herself and taking responsibility (‘I was a very fussy eater’, ‘I
knew how to press mum’s emotional buttons’) and acting in a caring parental role towards
her mother (A3–4 strategies). She appeared not to be able to employ either strategy
consistently and possibly displayed a mixed pattern (A/C) to cope with the conflicting
demands. She describes how this led to her feeling that the only solution was her anorexia:
I worked out that crying doesn’t work. No matter how hard I cried, it never
worked. Nothing ever changed and I became very good at just crying on the spot,
but it didn’t do anything so it (anorexia) is just another way of crying.
CRITTENDEN & DALLOS: INTEGRATING ATTACHMENT
401
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
Transformations and distortions of meanings
SFT has moved to a focus on families as meaning making systems. Families members are
seen as holding unique perceptions of their experiences and as attempting to make the
best sense that they can of their experiences. This has been variously described as a move
to a constructivist and subsequently a social constructionist perspective. These have
emphasized the centrality of language in the construction of family’s experiences. For
example, ideas of what it is to be a ‘normal’ family, to be a mother or father, what counts
as disturbance or ‘mental illness’ are all culturally constructed ideas that are absorbed
by families and come to shape their relationships with each other. A number of systemic
therapists and researchers have also explored the idea of shared actions and beliefs in
families: Reiss (1980) developed the idea of families as having shared paradigms, essentially
a set of shared beliefs that regulate members’ actions and feelings. Dallos (1991)
and Procter (1981) drew on George Kelly’s (1955) personal construct theory to develop
the idea of families as organized on the basis of a shared construct system. Likewise,
Eron and Lund (1993) have developed an approach in which they conceptualize families
in terms of the interplay between their stories or ‘viewings’ and their actions. Byng-Hall
(1991) developed the concept of family scripts which again contained the idea of
experience as both shared and unique. He also proposed the idea of ‘corrective’ and
‘replicative’ scripts which suggested that family members attempt in their current
families to perpetuate what was experienced as good or remedy what was experienced
as ‘bad’ in their families of origin. These approaches make attempts to look at how
beliefs develop in families and, more implicitly, they also reveal that emotional processes
in families can interact with these, for example, Eron and Lund (1993), and Dallos (1991)
indicate how beliefs in families may gradually become more extreme and polarized.
These approaches can be complemented by a deeper developmental perspective
offered by attachment theory which can help to track the style of narratives in terms of
the transformations of information, extent of verbalization of representations, and selfprotective
strategies that children develop in their early relationships. An important
related point is that SFT, especially in its contemporary versions, is shy of using concepts,
such as distortion or falsification since these concepts are seen as having negative and
blaming overtones. We suggest that what is needed is a model which helps us to see how
the transformations and ‘distortions’ of available information in families occur and how
this relates to emotional states of the members involved. Returning to Watzawick’s
classic example earlier, it is clear that couples caught in such a cycle become increasingly
emotionally aroused and, in fact, extreme feelings can be triggered very rapidly between
them (Gottman, 1982). Furthermore, they can be seen to bring with them to the relationship
a propensity to see others, and hence also their partner, as avoiding, rejecting, suffocating
and so on which is related to their prior attachment histories in their own families
(as well as previous romantic relationships and friendships). From a DMM perspective,
they would be seen as having distorted probabilities regarding the danger the other
implied to themselves and a misunderstanding, sometimes even to the point of a falsely
inverted prediction, of the effects of their own behavior on the other.
Transgenerational perspectives
Both SFT and attachment theory share an interest in patterns across the generations.
Again for early SFT, this was a major area of interest. For example Bowen (1971),
drawing on psychodynamic ideas,offered a powerful model of how problems evolve over
several generations. Attachment theory has developed some ingenious assessment tools
such as the Strange Situation (Ainsworth et al., 1978), School-age Assessment of Attachment
(Crittenden, 1997–2005), and Adult Attachment Interview (George, Kaplan, &
CLINICAL CHILD PSYCHOLOGY AND PSYCHIATRY 14(3)
402
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
Main, 1986/1996; Crittenden, 1999) and the Parents’ Interview (Crittenden, 1982) that
can be used to explore transgenerational patterns in detail. Focusing on what is passed
across the generations allows us to use concepts, such as distortions and falsification,
without ‘blaming’ parents since we are able to see the struggles and difficulties that they
faced themselves and the aspects of these that they have brought to their current family.
What is learnt are self-protective strategies that were needed for their emotional survival
and, if we can understand and communicate this understanding, we can be authentically
respectful and sympathetic to their actions.
Carla’s parents, for example, both described emotionally cold childhoods. Carla’s
maternal grandmother had been depressed, suicidal and not available to Carla’s mother.
Likewise, Carla’s father described his own father as angry and intimidating and his
mother as permanently ill, overweight and emotionally unavailable. As if this wasn’t
enough, Carla’s mother had desperately wanted a daughter to repair the sad experiences
she had with her own mother. Instead, she had given birth to three boys and had
abandoned hope of a daughter when finally Carla came along.This had a powerful impact
on the parents’ marriage which had been in crisis, and her father described the
2 years after Carla’s birth as the happiest period in their marriage. From her birth, it
seemed she was meeting not only her mother’s needs, but also had been saving their
marriage!As has been seen so often inAAIs read by one of us (PMC),strategies intended
to repair the old problem often reverse the problem, like the proverbial pendulum,
leaving well-meaning parents totally unprepared for the problems they create.
Applying the integration of family systems therapy and the
DMM to psychotherapy
Attachment theory is not a theory of treatment. The DMM, however, is a fully articulated
developmental theory that can (1) contribute to formulating an understanding of
dysfunction; (2) assess individual psychological organization, dyadic strategies and
multigenerational relationship strategies; and (3) provide a rationale for selecting among
treatment modalities. Further, attachment theory has generated a wealth of data on
normative children and parents; this could reduce pathologizing of normative processes
and life transitions. Moreover, with input from all the major theories of treatment, the
DMM is being developed into a comprehensive theory of treatment inclusive of ideas
from all major theories of treatment (Crittenden, 2008).
Family systems therapy can contribute a family perspective, a number of patterns of
family organization, and extensive experience with treatment from which to derive
hypotheses regarding the effects of specific treatment strategies on individuals and
families who differ in psychological organization and self-protective strategy.
Understanding dysfunction and maladaptation
Assessment and formulation The DMM has generated a developmentally attuned array
of assessment tools from infancy to adulthood. FST could adapt these to the family
context to facilitate description of various types of problems displayed in families,
including eating disorders, depression, self-harm, conflict and violence. Further, it is clear
that in some families, especially troubled families, family members use different strategies.
Furthermore, they are often not able to accept each other’s different ways of
coping with emotional distress and become locked in cycles of blame and accusations
with, for example, one partner seeing the other as distant and unemotional (i.e. Type A)
and the other, in turn, seeing their partner as overemotional or neurotic (Type C). The
CRITTENDEN & DALLOS: INTEGRATING ATTACHMENT
403
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
DMM assessments can help to clarify the different strategies that family members use
and consider the impacts they have on dyadic as well as triadic patterns in the family.
Both FST and the DMM benefit from offering a conceptualization of all families (both
well and poorly adapted) rather than seeing a separation between pathological and
normal families. For the DMM, these constructs include transformations of information
and behavioral strategies. That is, what actually happened, what one thinks and feels
about it, and how one acts, are conceptualized as being related in predictable, but not
necessarily identical, ways. When there is a wide disparity within and among family
members, the likelihood of dangerous or endangering behavior is increased.
Case study 1: Rosie
In a case of factitious illness by proxy, parents’ excessive zeal in seeking unnecessary
surgical treatment for their daughter Rosie was attributed by professionals to intentional
and deceptive harm. Using the AAI to offer both an assessment of the parents’ attachment
strategies as well as the rich qualitative information available from the interview
about their childhoods and current concerns, it became clear that both parents had imaged
representations of the early and traumatizing deaths of their fathers; both deaths occurred
quickly, without forewarning, and with disastrous consequences for their families. A
compelling formulation from their interviews was that, in the context of Rosie’s transient
illness at birth, these nonverbal, imaged representations elicited feelings of fear that, in
turn, motivated excessive effort to protect her life. However, the parents were not
explicitly aware of how Rosie’s illness may have triggered these unresolved feelings and
so their effects on behavior could not be regulated consciously by them. It appeared that
the mother’s compulsive caregiving strategy (A3) pushed her into overdrive, protecting
her daughter from illusory dangers while her husband’s dismissing strategy (A1) caused
him to distance himself and defer to his wife. Both parents did what they believed would
protect their child. Professionals, on the other hand, saw only the absence of illness in the
child and the parents’ irrational attempts to obtain ever more intrusive and dangerous
medical intervention (Kozlowska, Foley, & Crittenden, 2006).
This example highlights the importance of learning that occurs in the context of danger
to guide later protective behavior in ways that are usually adaptive but sometimes dysfunctional.
Further, it demonstrates how multigenerational information can be extracted
from attachment assessments and applied to family treatment. It also highlights how
implicit information can result in discrepant representations that, when not perceived and
reconciled, can lead to maladaptive behavior. In this case, two discrepancies are important:
that among the parents’ representations and that between the parents’ intentions and the
professionals’ attributions regarding the parents’ intentions.
Description, function, and attribution Possibly discrepancy among perspectives can be
resolved if description, function, and attribution of meaning are differentiated. Behavior
can only be understood if it is described accurately and fully.
Case study 2:Albert
At 22 years old, Albert, a White male, had been hospitalized three times with a diagnosis
of schizophrenia. Neither supportive therapy nor medication had helped and medication
sometimes worsened his condition. Albert lived alone with his mother who appeared to
be caring, with no obvious dysfunction. He had suspended his studies and had no friends
or girlfriends.
Because his condition was becoming chronic, a consultation was arranged with one of
us (PMC). Albert’s AAI showed no feelings whatever. Further, he described several
childhood events that indicated occasional physical abuse, chronic neglect, and ongoing
CLINICAL CHILD PSYCHOLOGY AND PSYCHIATRY 14(3)
404
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
sexual involvement and spousification (Minuchin, 1974). Albert himself, however,
made no such accusations. Instead, he idealized his mother to the point of delusion and
yet it appeared that she might be depressed and that he was her entire life. With the interviewer,
he was meek and obedient. His AAI was classified as (a) Type A (compulsive
spousification, i.e. needing to attend to his mother’s emotional needs as if he were her
husband, compliance, and delusional idealization, i.e. A3, 4, and 7) with (b) multiple unacknowledged
and unresolved traumas and (c) a pervasive sense of futility, that is,
depression.
When the AAI was considered in the context of Albert’s history, his behavior during
the interview, and his mother’s situation, several hypotheses were proposed. First, it
appeared that the function of the hospitalizations was to give respite to Albert from his
mother without causing interpersonal distress. That is, if Albert had wanted to live alone
or with a partner, his mother would have felt abandoned. But if he was ill with a serious
and inexplicable disorder, then, of course, she wanted him to have treatment. Second,
Albert was in the ‘transition to adulthood’ when, developmentally, the dominant cultural
expectation was that he would be leaving home, but given his mother’s needs he could not,
nor could this be articulated by either of them. Instead, his behavior expressed both his
mounting sense of urgency to be able to move forward in his life and his sense of futility
about this. Third, his diagnosis both explained (to everyone) why his life was on hold and
also relieved both his mother and himself from any responsibility for change. Fourth, if
Albert had access to negative affect, it would destroy his relationship with his mother;
without it, he could not be motivated to seek sexual satisfaction or independence. Finally,
a developmental perspective suggested that Albert’s problems had not come to light
earlier because, at younger ages, Albert had needed his mother more than he needed
independence and, therefore, the conflict was not yet irreconcilable.
When the unspeakable conflicts became too much, Albert’s symptoms became acute
and he went to hospital. Attributing his behavior to a disease condition protected Albert
and his mother from change at the point in life when change was mandatory if Albert was
to move forward (Crittenden, 2008).
The AAI and Albert’s behavior in the psychiatric interview revealed an unexpected
family process – one that was quite different from the idea of biological anomaly currently
used to explain schizophrenia. The new story, however, fits the understanding of early
family therapists quite well; it differs in having new methods of data gathering and
analysis. The DMM emphasizes the functional nature of all attachment strategies,
especially highly distorted strategies, while recognizing that what may have been functional
in one context may be dysfunctional or even dangerous in others.
Implications for treatment
DMM theory of attachment and adaptation suggests that there are three related topics
to address in psychotherapy: (1) The presenting problem, (2) the underlying threat to
survival, reproduction, or survival of one’s children, and (3) the information processing
that transforms threat to maladaptive behavior. It is proposed that it is the underlying
threat that makes the presenting problem stand out against the background of life’s
many problems. If it is not addressed, the solution may be temporary and another
problem, with the same roots in threat, may replace it. Further, it is information processing
that brought solutions from the past to the current context where they do not
function adaptively. If information processing is not addressed, it may be applied in an
unchanged manner to other problems in the future, leading to new misunderstandings
and possible maladaptive behavior. Finally, each family member will frame the presenting
problem differently, process information differently, and have a different history of
threat, development and problem-solving skills. Therein lies the complexity. Addressing
CRITTENDEN & DALLOS: INTEGRATING ATTACHMENT
405
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
these unique developmental issues can make each family member’s behavior comprehensible
to both themselves and others. Compassion and forgiveness may be most easily
achieved when behavior is understood from each person’s perspective and, without
these, no solution to a family problem can be lasting. With these, enduring change
becomes possible.
Systemic family therapy has tools and experience in the processes of change. These
already reflect an integration of psychoanalytic principles with systemic theory and
knowledge of family functioning and communication patterns. Attachment shares these
conceptual roots, but without the experience in therapeutic technique. In addition,
however, attachment brings principles and ideas from behavioral and cognitive therapy.
These can expand further the conceptual and applied base of systemic family therapy.
The idea is that theories of treatment need not compete, but instead can be integrated
around their compatible elements with, in addition, each theory contributing what it
addresses best. For SFT, that is working with complex family units to benefit the functioning
of all members.
With whom should we work? Specific individuals or dyads within the family might
need different therapeutic approaches because they differ in the dangers they face, their
developmental maturity, or their preferred strategy. Should we work with the individual,
the family, the neighborhood, or the cultural context? In safe communities, individuals
living in unsafe families may need psychological treatment whereas families living in
unsafe communities may need community-level intervention more than psychological
treatment. In countries that are unsafe for everyone or for specific subgroups (for
example, for women or religious minorities), intervention at the cultural/political level
may be essential.
Presuming that we decide to work with individuals, which individuals should it be?
FST reminds us both to focus on larger units that include the patient and also that the
effects of treatment can be transferred through systems such that the designated patient
is not always the one who should receive the treatment. In what circumstances should
we offer parent training to the parents of a disturbed infant? Couples therapy to the
parents of an acting-out or withdrawn child? Individual work that will affect family
functioning? Family work? Some combination of these? Historically, FST has had the
flexibility to convene different subsystems of a family as was relevant and necessary
(Minuchin et al.,1978). From the perspective of the DMM, this reflects everyday family
experience in which individuals and subsystems are in constant flux while the secure
attachment relationships among them make trust in privacy and goodwill possible.
With the advantage of the structured assessment and formulation that attachment
brings, together with a refined developmental perspective, systemic family therapy is
positioned to make the leap to the next level of competence. Bringing all of its perspectives
and skills, from across the range of approaches developed by family therapists, and
adding ideas funneled through the DMM could yield a powerful array of therapeutic
tools. Were all therapists to become competent in (a) formulating problems in developmental
and adaptive terms and (b) selecting and applying tools from the full range of
possibilities, SFT might become more effective and also better able to document its
efficacy.
A research agenda
Together, FST and the DMM can offer an elegant and cogent way to conceptualize
many forms of dysfunction. Application of these ideas to treatment, however, requires
accumulating precise knowledge about the effects on different individuals of various
psychotherapeutic techniques.
CLINICAL CHILD PSYCHOLOGY AND PSYCHIATRY 14(3)
406
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
Current research compares randomized treatment groups that are defined by
symptom-based diagnoses. However, symptoms can serve different functions. If so, a
single form of treatment could be good for those who use the symptom in one manner
and harmful to those who use it in another. In other words, the studies suggesting that
psychotherapy has only limited effects may average, and thus obscure, substantial effects
for one set of patients and, possibly, detrimental outcomes for another.
A new research agenda may be needed. Such research would compare groups defined
by patterns of information processing and self-protective strategies. Both positive and
negative outcomes would be tested and reported. In particular, the effects on Types A
and C individuals would be tested. If carried out systematically, a body of information
(‘a psychotherapists’ desk guide to treatment techniques’) could be generated. This
would assist therapists to select treatment strategies that were suitable and safe for
specific individuals and families.
Posting the banns: A plan for joining attachment and
family systems
Family systems theory describes the struggles that occur within families and the symptoms
of distress that this generates whereas the Dynamic-Maturational Model of attachment
and adaptation provides an understanding of the self-protective, reproductive, and
progeny-protective functions that underpin these conflicts. Clarifying and prioritizing the
issues underlying distress will promote more effective and efficient treatment with
reduced risk of harm. We think the time is ripe for attachment and family systems theory
to draw closer together. With so much in common and so much to gain from each other’s
strengths,union might bring us nearer to a universal approach to psychological treatment.
If so, the true winners will be the families who come to us with their hope.
Note
1. Transformation refers to representations of the probability of specified conditions pertaining
in the future. For example, an extremely dangerous occurrence can produce a distorting
transformation in which the event is never expected (omission), always expected
(distortion), expected when an irrelevant aspect of the occurrence is present (error), or
expected when opposite signals are present (falsification).
References
Ainsworth, M. (1967). Infancy in Uganda. Baltimore, MD: Johns Hopkins.
Ainsworth, M.D., Blehar, M.C.,Waters, E., & Wall, S. (1978). Patterns of attachment: A
psychological study of the strange situation. Hillside, NJ: Lawrence Erlbaum.
Bateson, G. (1972). Steps to an ecology of mind. New York: Ballantine.
Bowen, M. (1971). The use of family therapy in clinical practice. In J. Haley (Ed.), Changing
families: A family therapy reader (pp. 159–192). London: Grune and Stratton.
Bowlby, J. (1973). Attachment and loss,Vol. II: Separation, anxiety and anger. London:
Hogarth Press.
Bowlby, J. (1980). Attachment and loss,Vol. III: Loss. New York: Basic Books.
Bowlby, J. (1982). Attachment and loss,Vol. I: Attachment. London: Hogarth Press.
[Originally published in 1969]
Byng-Hall, J. (1991). The application of attachment theory to understanding and treatment
in family therapy. In C.M. Parkes, J. Stevenson-Hinde, & P. Marris (Eds.), Attachment
across the life cycle (pp. 199–215). London: Tavistock/Routledge.
CRITTENDEN & DALLOS: INTEGRATING ATTACHMENT
407
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
Crittenden, P.M. (1982). Parents’ interview. University of Virginia, Charlottesville,VA,
available from the author.
Crittenden, P.M. (1983). Mother and infant patterns of interaction: Developmental
relationships. Dissertation Abstracts International, 45(8-B), 2710.
Crittenden, P.M. (1995). Attachment and psychopathology. In S. Goldberg, R. Muir, &
J. Kerr, (Eds.), John Bowlby’s attachment theory: Historical, clinical, and social significance
(pp. 367–406). New York: The Analytic Press.
Crittenden, P.M. (1997–2005). School-age assessment of attachment coding manual.
Unpublished manuscript, Miami, FL, available from the author.
Crittenden, P.M. (1999). Attaccamento in etá adulta: L’approccio dinamico-maturativo alla
Adult Attachment Interview. Milano: Cortina.
Crittenden, P.M. (2006). A dynamic-maturational model of attachment. Australian and
New Zealand Journal of Family Therapy, 27, 105–115.
Crittenden, P.M. (2008). Raising parents: Attachment, parenting, and child safety.
Cullompton: Willan.
Crittenden, P.M., Claussen,A.H., & Kozlowska, K. (2007). Choosing a valid assessment of
attachment for clinical use: A comparative study. Australia New Zealand Journal of
Family Therapy, 28, 78–87.
Dallos, R. (1991). Family belief systems, therapy and change. Milton Keynes: Open
University Press.
Dallos, R. (2006). Attachment narrative therapy. Milton Keynes: McGraw-Hill.
Dallos, R., & Denford, S. (2008). A qualitative exploration of relationship and attachment
themes in families with an eating disorder. Clinical Child Psychology and Psychiatry, 13,
305–322.
Dallos, R., & Draper, R. (2005). An introduction to family therapy. Maidenhead: Open
University Press/McGraw-Hill.
Erdman, P., & Caffery,T. (Eds.). (2003). Attachment and family systems: Conceptual,
empirical, and therapeutic relatedness. New York: Brunner-Routledge.
Eron, J.B., & Lund,T.W. (1993). How problems evolve and dissolve: Integrating narrative
and strategic concepts. Family Process, 32, 291–309.
Ferreira,A.J. (1963). Family myths and homeostasis. Archives of General Psychiatry, 9,
457–463.
George, C., Kaplan, N., & Main, M. (1986/1996). Adult attachment interview. Unpublished
manuscript. University of California, Berkeley.
Gottman, J.M. (1982). Emotional responsiveness in marital conversation. Journal of
Marriage and the Family, 32, 108–120.
Haley, J. (1976). Problem solving therapy. San Francisco, CA: Jossey Bass.
Hillburn-Cobb, C. (1998). Adolescent–parent attachments and family problem-solving
styles. Family Process, 35, 57–82.
Hoffman, L. (1985). Beyond power and control: Toward a ‘second order’ family systems
therapy. Family Systems Medicine, 3(4), 381–396.
Jackson, D. (1957). The question of family homeostasis. Psychiatry Quarterly Supplement, 31,
79–99.
Johnson, S. ( 1996). The practice of emotionally focused marital therapy: Creating
connections. New York: Brunner/Mazel.
Kelly, G. (1955). The psychology of personal constructs. New York: Norton.
Kozlowska, K., Foley, S., & Crittenden, P.M. (2006). Factitious illness by proxy: Understanding
underlying psychological processes and motivations. Australia and New Zealand Journal of
Family Therapy, 27, 92–104.
Kozlowska, K., & Hanney, L. (2002). The network perspective: An integration of attachment
and family and systems theories. Family Process, 41, 285–312.
CLINICAL CHILD PSYCHOLOGY AND PSYCHIATRY 14(3)
408
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from
Main, M., & Solomon, J. (1986). Discovery of an insecure disorganized/disoriented
attachment pattern: Procedures, findings and implications for classification of behavior. In
M.W. Yogman & T.B. Brazelton (Eds.), Affective development in infancy (pp. 95–124).
Norwood, NJ: Ablex.
Minuchin, S. (1974). Families and family Therapy. Cambridge, MA: Harvard University
Press.
Minuchin, S., Rosman, B., & Baker, L. (1978). Psychosomatic families: Anorexia nervosa in
context. Cambridge, MA: Harvard University Press.
Olson, D.H. (1989). Circumplex model of family systems VIII: Family assessment and
intervention. In. D.H. Olson, C.S. Russel, & D.H. Sprenkle (Eds.), Circumplex model:
Systematic assessment and treatment of families (pp. 7–50). New York: Haworth Press.
Palazzoli, M.S., Boscolo, L., Cecchin, G., & Prata, G. (1978). Hypothesising-circularityneutrality:
Three guidelines for the conductor of the session. Family Process, 19, 3–12.
Palazzoli, M.S, Cecchin, G., Prata, G., & Boscolo, L.(1978). Paradox and counter paradox.
New York: Jason Aronson.
Palazzoli, M.S., Cirillo, S., Selvini, M., & Sorrentino,A.M. (1989). Family games: General
models of psychotic processes in the family. London: Karnac.
Procter, H. (1981). Family construct psychology. In S. Walrond-Skinner (Ed.), Family therapy
and approaches (pp. 350–366). London: RKP.
Read, J. (2005). Child sexual abuse and schizophrenia. British Journal of Psychiatry, 186, 76.
Read, J.,Agar, K.,Argyle, N., & Aderhold,V. (2003). Sexual and physical assault during
childhood and adulthood as predictors of hallucinations, delusions and thought disorder.
Psychology and Psychotherapy: Theory, Research and Practice, 76, 1–22.
Reiss, D. (1980). The family construction of reality. London: RKP.
Ringer, R., & Crittenden, P.M. (2007). Eating disorders and attachment: The effects of
hidden family processes on eating disorders. European Eating Disorders Review, 14, 1–12.
Sluzki, C. (1983). Process, structure and world views: Towards an integrated view of systemic
models in family therapy. Family Process, 22, 469–476.
Solomon, J., & George, C. (1999). The measurement of attachment security in infancy and
early childhood. In J. Cassidy & P.R. Shaver (Eds.), Handbook of attachment
(pp. 287–316). New York: Guilford.
Solomon, J., & George, C. (2008). The measurement of attachment security and related
constructs in infancy and early childhood. In J. Cassidy & P.R. Shaver (Eds.), Handbook
of attachment (pp. 383–416). New York: Guilford.
Spieker, S., & Crittenden, P.M. (in press). Comparing the validity of two approaches to
attachment theory: Disorganization versus danger-informed organization in the preschool
years. Clinical Child Psychology & Psychiatry.
Watzlawick, P., Beavin, J., & Jackson, D. (1967). Pragmatics of human communication.
New York: Norton
Watzlawick, P.,Weakland, J., & Fisch, R. (1974). Change: Principles of problem formation
and problem resolution. New York: Norton.
CRITTENDEN & DALLOS: INTEGRATING ATTACHMENT
409
at Masarykova Univerzita on December 2, 2011ccp.sagepub.comDownloaded from