Systems Theory, Family Attachments and Processes
of Triangulation: Does the Concept of Triangulation
Offer a Useful Bridge?
Rudi Dallosa
and Arlene Vetereb
This article reviews the usefulness of the systemic concept of triangulation
as a bridge between systemic thinking and practice and attachment
theory. Traditionally attachment theory has theorized and researched
with dyads, parent-child and adult romantic partners, whereas systemic
theory has worked therapeutically with the triangle as the basic human
relationship; that is, when any two people interact, their interactions are
influenced by their respective relationships with the same third person.
Here it is argued, for example, that a child’s attachment representations
are shaped not only by the relationship with each parent but with the
relationship between them. Thus, the process of triangulation in intimate
relationships is seen in this article to link attachment theory with systemic
thinking, as systemic theory needs a theory of love and attachment theory
needs to consider how intimate relationships are nested and interconnected
in an attachment network.
Keywords: interpersonal triangles; processes of triangulation; systems of attachment.
Introduction: triangulation as a systemic concept for practice
The systemic concept of triangulation has been most influential in our
thinking and practice. Possibly it could be said that this concept
distinguishes systemic from other forms of psychotherapy in so far
as practice is focused around triangular relationships (Bowen, 1978;
Dallos and Draper, 2005; Titelman, 2008). Other psychotherapies
have theorized about triangles; for example, within the psychoanalytic
canon under the umbrella of thinking about the Oedipal struggle. A
recent attempt at making use of this thinking in a systemic frame can
be found in Woodcock’s work (2009).
Journal of Family Therapy (2011) ]]]: 1–21
doi: 10.1111/j.1467-6427.2011.00554.x
a
Programme Director, Doctorate in Clinical Psychology, Plymouth University, Plymouth,
UK.
b
Professor of Clinical Psychology, Department of Psychology, University of Surrey,
Guildford, Surrey, GU2 7XH, UK; e-mail: drarlenevetere@hotmail.com.
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice. Published by
Blackwell Publishing, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden,
MA 02148, USA.
(2012) 34: 117–137
Triangulation can be conceptualized as a noun and a verb, in that
we can speak of dynamic triangular relationships between intimates,
and processes of triangulation that draw a third person into consideration
with a dyad. Early systems theory drew on Mills’ sociological
idea that the basic human relationship was a triad, not a dyad,
in that when any two people get together they are influenced by a
third. When two sisters get together, for example, their relationship is
influenced by the relationship each has with their absent sister and so
on.
In therapy, of course this opens up multiple possibilities for
exploring and working positively with powerful sources of influence,
for example, drawing on the spirits of our ancestors, working with
absent or dead relatives as if they were in the room, understanding
the dilemmas of children going through their parents’ divorce, and so
on. The early Milan team’s reliance on circular questioning as a
method of intervention is a good illustration of the power of working
with triangles (Palazzoli et al., 1978). So, in this way of thinking,
human relationships were seen to be a series of interlocking triangles
with the potential for being stable or unstable at any point in time.
However, the notion of triangle does not assume a positive or
negative valence but recognizes that triangles can work in benign and
not-so benign ways to stabilize and destabilize relationships. For
example, a woman married a man 35 years older than herself, and
when he entered his 90th year, she took a younger lover into their
household, with his consent, to help with his care. In another
example, two adolescent sisters, with apparently little in common,
joined forces to persuade their lone parent mother that they both
needed a fifth pair of jeans.
Triangulation and the relationship between systemic theory and
attachment
The concepts of triangular relationships and processes of triangulation
both connect and bridge attachment thinking and systemic
thinking and practice in a number of helpful ways. In focusing our
thinking on the triangle, we draw on ideas about intergenerational
legacies of attachment, corrective scripts, networks of attachment
relationships in households and extended kin groups and, in particular,
the idea of nested relationships, and finally, the power of
systems theory to contextualize socially, politically and economically
the development of attachments in families and communities of
2 Rudi Dallos and Arlene Vetere
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
118
concern, to show how both systems of thought complement and need
each other.
In so doing, we wish to acknowledge the influential work of other
scholar-practitioners such as, Akister and Reibstein (2004), Hill et al.
(2003), Flaskas (1997) and Byng-Hall (1991), who helped pave the
way for exploring the interfaces between these two major systems of
thought: family systems theory and attachment theory.
Modes of triangulation
Essentially triangulation contains the idea that what is happening in a
significant relationship between two people in a family can have a
powerful influence on a third family member, and vice versa, in
mutually reinforcing ways. Early systemic practitioners, such as Weakland
(1976) and Minuchin (1974), Minuchin et al. (1978) observed that
families may engage in a variety of triangular processes that could
lead to stable coalitions or less stable and shifting alliances.
Minuchin outlined three processes by which this could occur. In the
first, the parents experience unacknowledged and unaddressed dissatisfaction
in their relationship that creates anxiety for at least one of
them. Minuchin argued that this anxiety needed an outlet or focus
and if a child behaved in a way that drew both parents’ concerned
attention to her, the child would notice that both parents worked
together and were united in their concern and this reinforced the
child’s behaviour – and thus ‘problem’ behaviour was born. This
premise is built on the developmental idea that children notice and
monitor their parents’ well-being as individuals and crucially, the wellbeing
of their relationship.
Winnicott (1964) also wrote about helping families to untie the
developmental knot caused by a child’s monitoring of their parents’
relationship difficulties. As a possible formulation for some childhood
emotional and behavioural ‘difficulties’ it blames neither the child nor
the parents but seeks to understand how, both consciously and with
non-conscious processing, people get caught up in the ironic consequences
of their own and others’ behaviour and intentions, often in
an effort to make things better.
In another example of the power of ironic consequences, ByngHall
(2008) wrote of children’s role in distance regulation or affect
regulation in a parents’ couple relationship. Two parents are discussing
an issue that arouses anxiety and one or both begin to show signs
of discomfort. The young child, playing in the room, gets up and
Systems Theory, Family Attachments and Processes of Triangulation 3
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
Systems Theory, Family Attachments and Processes of Triangulation 119
toddles over to one parent, who with some relief, bends down, picks
up the child, puts them on their lap and pats them, and so on. And
thus the difficult adult conversation is interrupted, unhelpful arousal
is calmed by the action of attending to and soothing the child and the
child learns, before they even have the use of language, how they can
be helpful to their parents. If this sequence is repeated and positively
reinforced a few times it could become established as a pattern or a set
of expectancies about how to manage unhelpful arousal.
A second triangular process, or cross-generational coalition, was
observed when again, there was distress and dissatisfaction in the
parents’ couple relationship that was not addressed in any constructive
way. Here each parent tried to recruit the child to side with them,
against the other parent. The dilemma for the child in such circumstances
is that to please one parent is to displease the other. This
experience is not uncommon for children going through their
parents’ divorce when hostility arises or is continued through the
legal and psychological process of separation, custody, residence and
contact.
Children’s experience of divided loyalties in these circumstances
has been documented by Dowling and Gorell Barnes (1999) and Blow
and Daniel (2002), who show it has immediate and longer term
adverse social and psychological consequences for some children.
So, when children live in two households it is difficult to manage the
strain of hearing one parent speak ill of the other, whom the child may
love, and for the child to know they cannot speak well of one parent to
the other parent for fear of inciting the latter’s displeasure.
In the third example provided by Minuchin there is an unacknowledged
and unaddressed parental dissatisfaction with the couple
relationship but one parent seeks affection and stimulation outside
the couple relationship in having a sexual affair or perhaps in locating
all recreational activities outside the family (a more disengaged style of
relating). The more residential parent might well turn to one of the
children for emotional and practical support, confiding in them and
relying on them perhaps in a way that is beyond what is emotionally
manageable for the child or in ways that make it harder for children to
take up their developmental place with their peers under the
constraints of societal expectations (a more enmeshed style of relat-
ing).
This focus on triangles can both be an important bridge between
systemic and attachment perspectives and a higher order context
marker for interconnected attachment relationships in a family
4 Rudi Dallos and Arlene Vetere
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
120
system. It allows us to see a child as developing and functioning in
both direct dyadic relationships with each parent and also with the
relationship between them (Pallazoli et al., 1989). In effect, the child
can be seen as developing attachment strategies with each parent and,
at the same time, having those strategies function to meet their
parents’ emotional needs in their relationship, but without the child’s
awareness.
Alan Schore (2009) writes about the role of the relational unconscious
in affect regulation and our capacity to respond to relational
cues without the immediate aid of cognition and language to help us
make sense of what is happening. The situation for the child can
become increasingly complex and confusing when there is no open or
straightforward discussion in the family about what is going on
between them all, what people are feeling and what their intentions
and wishes are. In such situations, children and their parents may
become increasingly confused by, or more unwilling to think about,
the causes of events, what maintains relational patterns and their own
role in the developing problems.
Case illustration of triangulation and interpersonal dilemmas
We can illustrate some of these dilemmas through the experience of a
young woman, Kate, who had been suffering with anorexia. The first
author worked clinically with her and her family several years ago and
also as part of a research study exploring shared attachment themes in
the family (Dallos and Denford, 2008).
Kate, aged 17, was living with her parents and attending school at
the time of the referral. She was the youngest child, with three older
brothers, one of whom had recently moved back home. She had been
suffering with an eating disorder for the past 2 years and had a
number of paediatric ward admissions to assist with weight gain but
had not been admitted as an inpatient at any stage. Both her parents
lived at home and both were in full-time paid employment. Kate’s
mother acknowledged early in the family sessions that she was
unhappy in her marriage and wanted to leave to find a more fulfilling
relationship, whereas Kate’s father wanted to save the marriage.
Both Kate’s parents had troubled attachment histories in their
families of origin, having experienced emotional unavailability from
their respective parents. The grandparents in turn, were reported to
have had troubled lives, including psychosomatic problems and
attempted suicides. The first author worked with Kate and her family
Systems Theory, Family Attachments and Processes of Triangulation 5
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
Systems Theory, Family Attachments and Processes of Triangulation 121
in family therapy for 18 months. The research interviews with the
family used the adult attachment interview (George et al., 1986), both
individually with family members and with parts of the family as a
family interview. This research complemented the therapeutic exploration
of relationships with food, food and comfort, mealtimes and
the family members’ explanations of the eating problems.
The following example from the research interview illustrates how
a child may experience the tensions and conflicts between her parents
as anxiety-provoking, which may eventually result in her expressing
forms of distress. Even more powerfully, a child may feel that she is
being drawn in to take sides in a battle between her parents that she
finds highly distressing and confusing. Below are two quotes from
Kate:
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?’ . . .
The only thing I ever hear them talking about is me and if I didn’t have
this [anorexia] its kind of like, would everything fall apart, at least its
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.
Here we can see how Kate expresses very poignantly what the
experience of being caught up in the conflicts between her parents
feels like. She even offers a very powerful systemic formulation of how
her anorexia can function to stabilize the family system.
Intergenerational consequences of triangulation processes
A wide range of intergenerational issues arise from this research
example, not least that it might seem to be blaming the parents for
having caused this young woman’s condition. There are many reasons
both in the present and in the past why parents, separately and
together, might feel distress, be insecure about each other and engage
in conflict and in hostile interactions, and even why they might involve
a child in such ways. We might even want to contend from a social
constructionist position that this is just Kate’s narrative and represents
her version of events, not a truth about the dynamics in this family.
However, for this family, this was not only Kate’s story but also her
6 Rudi Dallos and Arlene Vetere
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
122
parents’, as they, too, reflected that they were concerned that their
frequent arguments may have helped to cause Kate’s ‘illness’.
Added to this, the parents said that they had been thinking of
splitting up, and her mother especially explained that she was very
unfulfilled in the marriage and wanted to end it. So in effect, this story
represented a form of shared narrative, though with variation
between the three members of this family in how they thought it
related to Kate’s condition. Interestingly, Kate’s older brother held a
view that although their parents argued, this was made worse by Kate
and that her anorexia was a form of selfish and self-centred activity.
His perception here can nicely be seen in terms of Minuchin et al.’s
(1978) formulation that although a child’s problems might originally
be connected to parental conflict, in turn the child’s problems and
behaviour can serve to make it harder for them to reconcile their
difficulties despite their best efforts. More contemporary writers
might describe this as a problem-determined system (Anderson et
al., 1986).
Further reflections on triangles and triangular processes
We have been conducting research as well as clinical work with
families with a range of presenting problems. More research is needed
into the complexities of triangular care and the implications for
systemic practice. In our research we interview family members
individually using the adult attachment interview and a semi-structured
interview about their relationships and their view of the
problems.
We also interview them jointly to talk about their relationships and
attachment experiences and connections in their current family and
across the generations (Dallos, 2006; Dallos and Denford, 2008;
Dallos and Vetere, 2009; Vetere and Dallos, 2008). What we have
discovered from this research and from our clinical experience is that
Kate and her family are somewhat unusual, in so far as they had the
ability to recognize some of these connections between their conflicts
and Kate’s condition. In contrast, in many families, members are able
to describe events and relational dynamics that clearly suggest triadic
processes but do not connect these as being related to the problems
that develop.
Reading this you may be thinking; but surely many parents do
blame themselves for what has gone wrong and feel guilty that they
are failing? Of course, this appears to be true but what may be harder
Systems Theory, Family Attachments and Processes of Triangulation 7
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
Systems Theory, Family Attachments and Processes of Triangulation 123
to acknowledge, and here we speak as parents who both deeply regret
some of the marital conflicts to which we have subjected our own
children, is the fundamentally triadic aspect of this process.
Many parents continue to blame each other and or themselves but
do not, as Kate is starting to see in the example above, understand the
triadic nature of this process. Arguably, this is in part because
European cultures do not seem to hold triadic views of relational
causality, at least in research design and methodology but seem more
focused on either individualistic, or at most, dyadic explanations. In
an implicit way however, Kate seems to be suggesting that she views
both of them as equally responsible but moreover that ‘it’s both of you,
your relationship, that is distressing for me’:
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.
Arguably many parents may flit from thinking that it is their relationship
that is the problem but are likely to drift from this into blaming
the other or possibly even themselves as the primary cause. Also,
many parents are less aware that the triadic arrangement where one
parent has partially withdrawn may still continue to have a powerful
and negative effect on a child who continues to feel the underlying
tension or may, for example, remain very wary of ever saying the
‘wrong thing’. For example, a child may be careful never to say to her
mother that she misses her father for fear that she may upset her
mother, thinking that her mother would feel hurt that she was failing.
This might be a situation where the parents are on the surface
‘amicably’ separated and say little explicitly negative about each other
but the child feels the tension, for example in her mother’s body or a
change in her breath and posture when the child’s father is mentioned.
For such reasons the process of the triangular relationship and
such continuing emotional impacts may not get openly discussed, yet
are remembered and stored in the relational unconscious (Schore,
2009).
One thing a child may therefore be learning is to disguise her true
feelings and develop what Crittenden (2008) called ‘false affect’. Over
many years such learning may result in the child becoming not clear
herself about what she feels at any given time, as pointed out by
Winnicott (1964) in his notion of the ‘false self ’. He theorized that the
8 Rudi Dallos and Arlene Vetere
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
124
false self emerged as a child repeatedly managed the dissonance or
lack of congruence between their caregiver’s ability to meet their
needs and the child’s actual needs for nurturance, reassurance and
comfort.
Systemic research on triangulation
Despite the centrality of triangulation as a key feature of systemic
theory and practice we are hard pushed to find contemporary
research from the systemic field that explores how this process works
and the possible outcomes. There are, however, two strands of
enquiry that we have come across.
The first comes from research on how inter-parental conflict
impacts on children (Amato and Afifi, 2006; Buehler and Welsh,
2009; Buchanan et al., 1991; Cummings et al., 1989; Davies et al.,
2002; Vetere and Cooper, 2005, 2006). The accumulation of this
research supports our clinical experience that conflict between parents,
which may range from domestic violence to, as in Kate’s case, the
constant imagined dread of something bad happening, has a deleterious
effect on children’s mental health (Rutter, 1999).
This body of research has mostly employed a variety of measures of
the level and frequency of parental conflict using inventories from
parents and children and measures of children’s well-being. A consistent
finding is that children display significantly higher levels of
internalizing symptoms, such as anxiety, depression, low self–esteem,
substance misuse and self-harm (Davies et al., 2002). More specifically,
the effects seem to operate on two levels – children find conflicts
between the parents distressing but the effects of this are greater when
the parents pull them in to take sides (as in Kate’s example earlier).
This body of research also suggests that children attempt predominantly
two forms of strategies. One is to try and avoid the conflicts
between their parents by escaping, retreating to their room, staying
out of the home as much as possible and disconnecting emotionally.
The other is to attempt to intervene, for example, to keep the peace,
reason with their parents to stop, and try to be particularly nice and
pleasant. It also seems that some children may intervene with their
own arousal regulation problems by externalizing symptoms, such as
shouting at their parents, becoming hostile and getting into trouble.
These can be even more successful ways of distracting their parents so
the child, not their parents’ relationship, comes to be seen as the
problem, as negative interaction cycles escalate. Finally, again as in
Systems Theory, Family Attachments and Processes of Triangulation 9
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
Systems Theory, Family Attachments and Processes of Triangulation 125
Kate’s example, it may be that the emergence of severe symptoms can
be a very powerful way of distracting the parents from their relationship
issues in a mutually reinforcing cycle of interaction.
We can also see some parallels here with the ideas of the Bateson
group (1972) regarding the double bind. The process of triangulation,
or the invitation from the parents to take sides, makes the
strategy of withdrawing very hard, so that in effect, it is impossible to
leave the relational field. The nub of the double bind – ‘damned if you
do and damned if you don’t’ – lies in not being able to step outside the
interaction and comment on it at a reflective level of understanding.
Somehow it is not safe for the child to run a commentary on what is
happening, even assuming they can think reflexively about these
processes. Thus a third choice or way out of the double bind cannot be
seen, and the child needs the relationship with the parents for
survival. Weakland (1976) had extended double bind theory to see
it as a three-person phenomenon and pointed out how confusing
communications could arise due to each parent knowing when they
were communicating to the child how what they said might be
undermined by the other. Hence, again as Kate describes, she felt
filled by the bad stuff about the other parent rather than feeling that
she was being attended to herself. Similarly, Blow and Daniel (2002)
write about such experiences for children in the context of hostile
parental divorce proceedings.
The second point of connection is with attachment theory. The
above body of research has come to some important general conclusions
that what children experience in these situations of parental
conflict is best described as attachment insecurity. The children display
insecure attachment patterns on attachment measures and,
perhaps surprisingly, instead of being more likely to imitate or learn
their parents’ aggressive behaviour or conflicted styles of relating,
some children are more likely to show indications of anxiety and
depression.
Now attachment theory has essentially been a dyadic theory. Its
central premise is that the child forms an internal working model of
how relationships work, a view of how safe the world is, how valued
they feel as a person and how much they can trust others, from their
dyadic relationship with each parent (or carer) and particularly with
their mother (or the person who spends most time with them and
caring for them). However, it seems that in addition to these dyadic
relationships a child also has a relationship with their parents’ (or
carers’) relationship, actual and remembered. It is this relational
10 Rudi Dallos and Arlene Vetere
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
126
context that contributes greatly to how secure or insecure they come
to feel. Furthermore, we might suggest that it is this higher triadic
context that serves as the key marker for the dyadic relationships.
For example, a new young mother may be helped and supported to
make a relationship with her new baby by her supportive relationship
with her partner or her own mother; or each parent living in a
relational context of couple conflict is likely to be emotionally
distracted, distressed, unavailable or communicating confusingly
with a child, such as trying to put a brave face on things, pretending
that everything is fine when it is not and when they feel sad and
desperate. In our work with family violence we see these patterns
develop over time, as victims of violence comfort themselves with the
idea that the children are protected from knowing what is happening
between the couple (Vetere and Cooper, 2005). This is likely to
exemplify the confusing and contradictory communications that the
Bateson team wrote about many years ago.
Despite Bowlby’s (1969, 1973, 1980) robust interest in family
dynamics and his recognition of the complexity of attachments in
families, he nevertheless suggested that the place to start was to
understand dyadic process and once this was done to build up the
bigger picture. However, it seems that there have been few attempts in
attachment theory to do this. Furthermore, we might question
whether this falls into the systemic non-summativity dilemma, namely
the idea that the whole is greater than the sum of its parts and that we
cannot just add up the dyadic relationships in a family to explain the
bigger triadic and larger picture of nested and interconnected
relationships.
Most attachment research is based either on the use of measures
such as the narrative assessments, for example, the story stems for
children or the adult attachment interview that look at individual
attachment representations or the strange situation paradigm that
looks at the interaction between one parent and a child. There have,
to our knowledge, been very few attachments studies that look at how
one parent and child interact in the presence of another.
This seems a pity since attachment theory in many ways has
contributed to the fascinating development of our understanding of
the connections between the development of children’s inner representational
worlds and relational processes and provides us with a
theory of love across the lifespan. This interest was clearly the domain
of early systemic theory and practice. Many of the early family
therapists were researchers-practitioners who tried to advance aetioSystems
Theory, Family Attachments and Processes of Triangulation 11
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
Systems Theory, Family Attachments and Processes of Triangulation 127
logical models of how personal and interpersonal problems developed
in families as well as addressing the pragmatic questions of what
to do about them.
The second author’s participant observation research, in which she
lived with family groups for a week at a time, developed the notion of
triangulation and triangular processes through the use of family
systems theory and sequential analysis but again, we note, the
tradition of observational research was never developed further in
the systemic field (Vetere and Gale, 1987). We might lament that this
lack of interest in empirical exploration around the development and
maintenance of interpersonal problems has left us as a field guided in
its absence by fashion and pragmatics. However, this may mean that
without a developed theory of why and how problems develop in
different ways in different families we run out of new ideas of how to
promote and support change and may find it harder to fit therapy to
the specific needs of any given family.
The research work of Pat Crittenden
Pat Crittenden (1997, 2006, 2008) has developed a version of attachment
theory that is an exception to this exclusively dyadic emphasis,
in that it recognizes the powerful developmental influence that
triangulated processes can have on children and other family mem-
bers:
Triangulating parents try to protect the children from problems in the
marriage or, in more severe cases, engage the children in protection of
the marriage . . . In triangulated relationships, children perceive themselves
as having a direct relationship with the parent, whereas in
actuality, the parent’s interest is tied to how the children function to
preserve the spousal relationship. . . . Changes in parents’ perceived
threat can have a powerful impact on the parents’ behaviour, often in the
form of emotion based action. When the threat is not tied to, nor visible
to the children, but is acted out with the children as if they had caused
the parents’ behaviour, children become very confused about their
causal contribution to the relationship.
(Crittenden, 2008 p. 182)
In her account Crittenden describes how each parent (as with Kate
earlier) may draw a child in, such that their relationship is continually
influenced by reference to the other parent, including, as in Kate’s
case, possible accusations from each parent that she was too close to
the other parent and should instead side with them against the other.
12 Rudi Dallos and Arlene Vetere
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
128
However, in Crittenden’s description above, it is also possible that the
parent is trying to protect the child from the conflict in the relationship
but in a sense this conflict leaks out in indications of distress,
tension and irritation by the parent, so that the younger child
misattributes this as being her fault. This latter observation is crucial
since there may be many cases where children become concerned and
involve themselves without explicitly being invited to do so by the
parents. However, the parent may respond to the child’s caregiving
behaviour, when the parent’s need for comfort and reassurance is
high, such that a role reversal starts to take place. Clearly, children
learn social competence, develop their capacity for compassion and
learn how to care for themselves and others in family relationships.
Here, though, we are noting what might happen for children if it is
not safe to speak about and process these experiences.
Crittenden goes on to discuss the potential effects that this can have
on the understandings that develop for a child:
The two central issues for the children in triangulated relationships are
(a) forming accurate understanding of self-relevant causation, that is,
knowing what they elicit from their parents versus when their behaviour
is irrelevant to outcomes, and (b) developing a sense of themselves as
important in their own right, that is, developing self-esteem.
(Crittenden, 2008, p. 183)
Central to attachment theory and to Crittenden’s adaptation is that
children develop different attachment models or strategies according
to the nature of the experiences they have gained in their families. In
Crittenden’s version the two insecure patterns consist of A and C: A is
described as dismissive, or a dismissing or emotionally de-activating
strategy consisting of an excessive reliance on rational thought, the
exclusion of feelings – especially negative ones – and self-responsibility
and self-blame. For these children, the expression of negative
feelings, especially anger, is often forbidden and they feel shame and
humiliation when, as is inevitable, they occasionally display intrusions
of negative feelings. In contrast, in the preoccupied insecure C
patterns, in situations of attachment threat there is an excessive
reliance on feelings to guide action, including a tendency to be
overwhelmed by strong feelings and difficulty in managing unhelpful
arousal, at the expense of calm, considered thought with low selfresponsibility
and high levels of blaming of others.
Furthermore, in her model the way children who show insecure
patterns learn these patterns is not seen as being inherently to do with
Systems Theory, Family Attachments and Processes of Triangulation 13
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
Systems Theory, Family Attachments and Processes of Triangulation 129
copying the strategy of emotion regulation of their primary attachment
figure. Rather she postulates that children who show insecure
patterns of attachment might be matching their strategies to their
parents, such that a child may develop a preoccupied, highly emotional
and coercive strategy to get a reaction from a parent who is very
emotionally shut down and unavailable.
This is based on the idea that a child might prefer a negative form of
attention to being ignored. In contrast, a child may develop a dismissive
or emotionally deactivating strategy in the context of living with and
relying upon a very emotionally demanding and volatile, unpredictable
parent. Such a strategy is protective and helps lessen the impact of
unpredictable responding. Children who display secure strategies
appear typically to have parents who also show secure patterns.
However, attachment theory has increasingly needed to take
account of the suggestion that children in the same family may be
involved with different attachment patterns with each of their parents,
which may be rooted in an intergenerational attachment legacy and,
furthermore, that these patterns might be related to triangulation
processes.
For example, a research study by Schindler et al. (2007) suggested
that adolescents who were displaying problems of substance abuse
were predominantly in family configurations where the mother figure
displayed a preoccupied attachment strategy with her child and
emotionally drew him in to an enmeshed relationship that excluded
the father, who displayed a more dismissive pattern:
In families with a triangulated pattern of relationships, the adolescent is
needed to stabilize the family system. They are trapped in a cohesive
mother-child relationship, serving as an emotional substitute for a
dissatisfying marital relationship. The closeness seeking here is on the
part of the mother. Steps towards autonomy cannot be taken in spite of
the developmental stage of the adolescent. Drug use helps to bear this
situation, providing the adolescent with a ‘pseudo-autonomy’ that produces
a feeling of strength and autonomy. At the same time, he remains
dependent on his parents and is becoming dependent on the drug.
(Schindler et al., 2007 p. 114)
Clinical implications
One of the important conceptual and clinical contributions of Crittenden’s
(1997, 2008) dynamic maturational model is that it draws
14 Rudi Dallos and Arlene Vetere
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
130
attention to and embraces such findings by suggesting that children
may develop complex and interacting attachment strategies in difficult
family contexts. A simple classification of insecure patterns will
not be sufficient. Returning to Kate, she showed considerable understanding
of her situation but did not seem to take her parents’
perspective.
Possibly, being inside such a triangle of competing relationships led
her to have difficulty in consistently using either of the two main
attachment strategies in consistent, predictable and helpful ways. It is
important to remember here that insecure strategies are seen in
attachment theory to be functional, to be about survival, safety and
protection and are not necessarily associated with pathology. Dependence
is not seen as a pathological process, rather autonomy and
dependence are seen as two sides of an attachment relationship.
In clinical populations and therapeutic work it is much more likely
that we will see complex or mixed strategies that do not seem to
achieve a consistent sense of relationship satisfaction. For example,
Kate appeared to employ both a preoccupied strategy of insisting, at
all costs, on her own perspective and of fearing loss of protection
altogether if she bent and considered her parents’ perspectives
(Ringer and Crittenden, 2006).
At other times, she showed an attempt to distance herself from
negative feelings, blamed herself and took responsibility describing
herself negatively, ‘I was a very fussy eater’, ‘I knew how to press
Mum’s emotional buttons’ and to act in a concerned and caring
parental role towards her mother. 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, with hindsight, 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.
It appeared in our clinical work with this family that Kate’s mother
needed to be in a close and emotionally involved relationship with her,
perhaps seeking or trying to provide the closeness, caretaking and
intimacy that she did not have from her own mother who had been
depressed and suicidal throughout her own childhood (Byng-Hall,
Systems Theory, Family Attachments and Processes of Triangulation 15
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
Systems Theory, Family Attachments and Processes of Triangulation 131
1991). This could be described as a corrective script that did not fulfil
the mother’s intentions.
On the other hand, Kate’s father displayed a more dismissive
strategy and Kate felt she needed to curtail expression of her
vulnerable emotions and her needs for nurturance with him. Her
father had also experienced a difficult and insecure attachment
history as a child. However, both her parents drew her into taking
sides against each other, which made it hard for her to disengage and
also to show her feelings in a clearer and more direct manner. Neither
of the two attachment strategies, that is, options to avoid the family
conflict or to intervene in it, were functioning for her, especially if she
attempted to intervene by showing any anger, as this upset her
mother, who broke into tears in nearly every family session that we
held.
At these points Kate would eventually offer some comfort and
reassurance but would also express anger and irritation with her
parents’ behaviour. Helping family members understand and walk
around in these unhelpful patterns with a view to de-escalating their
emotional intensity and to softening blame is crucial to strengthening
and developing more mutually satisfying engagement.
Finally, an interesting observation made many years ago by Minuchin
et al. (1978) is that family dynamics can come to be selfmaintaining.
This has latterly been described as a problem-determined
system (Anderson et al., 1986). Theoretically, this is an important
distinction to draw, as Minuchin’s thinking followed in the
tradition of Harry Stack Sullivan’s notion of emotional symptoms as
ironic consequences of people’s attempts to solve troubling relational
dilemmas.
We see this notion of ironic consequences also in the work of the
early strategic family therapists who thought emotional distress
occurred as a result of attempted solutions to interpersonal problems
in that the attempted solution to the problem had become the
problem (Haley, 1987). These formulations of interpersonal difficulties
are thought to be less blaming and direct our attention to the
complexity of people’s intentions, expectations and motivations. The
notion of symptoms as ironic consequences is theoretically distinct
from the idea that symptoms have functional significance and are in
some ways necessary to maintaining family stability and functioning.
The notion of ironic consequences though, also allows us to think
how mutual responding is shaped and reinforced over time to take on
the life of a pattern. So, although Kate felt she could not withdraw
16 Rudi Dallos and Arlene Vetere
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
132
emotionally and risk leaving her parents to cope, they did encourage
her to go to university and become more independent. They also
decided to engage in some couples work to see if they could repair
their marriage. Part of this work involved the parents acknowledging
how important Kate had been in saving their marriage, which had
been very rocky before Kate was born but Kate’s arrival signalled the
most happy period in the couple’s relationship. In a sense she
appeared to have been central not only to each of them but to their
relationship from her first arrival into the family.
For her part, Kate returned from university (some 90 miles away)
almost every weekend, presumably to keep an eye on her parents and
perhaps also because she had become accustomed to her special place
in their relationship. This made it harder for her parents to resolve
their own issues with each other and to develop a more satisfying bond
with each other. Eventually Kate made her own emotionally intimate
relationships and gained some independence from her parents.
Triangles and relationships within family and professional
networks
Finally, we conclude this discussion on the usefulness of the systemic
concept of triangles and triangulation to consider their application in
thinking about and working with relationships in the family and
professional network. We consider an example from two service
settings, that of drugs and alcohol services and that of a domestic
violence project, to provide a conclusion to this article.
In drugs and alcohol work with individuals, couples and families, it
is not uncommon to find that the person who uses substances does so
as an affect-regulation strategy – to help them ‘numb out’, to avoid
painful feelings, such as shame and fear, and to manage anxiety in
social settings. It may be the case that after encountering adversity the
individual has learned not to trust anyone and turns to substances
rather than people for comfort and reassurance. In these circumstances
the ability to make a trusting relationship with an alcohol/
drugs key worker may be the first step or a bridge into further
therapeutic work with a partner or other family members, to address
the impact of substance use on people’s lives and its implications for
relationships and looking after children.
Entering into therapy with significant family members can be felt
as deeply threatening, so the key worker acts as a stable third in
the triangle, bridging and supporting the transition into more
Systems Theory, Family Attachments and Processes of Triangulation 17
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
Systems Theory, Family Attachments and Processes of Triangulation 133
challenging areas of therapeutic work. In our experience it is regular
consultation and liaison, with consent, between the three points of the
triangle, key worker, client/family and systemic therapy team, that
often prepares and ensures a smoother transition into systemic
therapy than might otherwise occur (Vetere and Henley, 2001).
Similarly, when working with family violence, the need to ensure the
safety of all family members is paramount. We developed a safety
methodology for therapeutic practice, including the management and
assessment of risk of violence, that developed the notion of interlocking
triangles (Cooper and Vetere, 2005). When violence in a family is
known or suspected, fear and anxiety about the recurrence of violence
may be high, both in the family system and in the professional system.
These anxieties may be openly expressed or indirectly communicated
in unclear and confusing ways. As part of our approach to safety, we
pay attention to levels of anxiety and how anxiety can appear to slip
around the professional system. We soon learned that if we did not help
the professional system settle down to recognize how we paid attention
to safety and risk, that it would be harder for us to settle into dedicated
therapeutic work with families in a way that was not inadvertently
sabotaged by the unsettled network. So in our safety methodology we
work with triangles as the minimum sufficient network.
The three points of the triangle consist of the family, ourselves and
the stable third, who may be the referrer or someone who is trusted by
the family and the professional system, such as a grandparent, a
health visitor, a faith leader or a community worker. The stable third is
someone who knows the children and has access to the house. They
are in a position to both corroborate what the family tells us about
safety and to help us develop, support and regularly review a safety
plan with the family. The safety plan has to be workable from all
participants’ perspectives. The watching professional network needs
reassurance that we will not maintain confidentiality about issues of
risk or repeat violence and that we will not be falsely reassured that all
is well by simply believing what we are told. Thus, the stable third
helps us manage risk and anxiety in the professional network and
family system and is one aspect of our approach to safety in our work.
Acknowledgement
Our grateful thanks to Brian Cade, editor of Context, and the Context
editorial team, who published a shorter version of this article in
Context 110 in August 2010.
18 Rudi Dallos and Arlene Vetere
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
134
References
Akister, J. and Reibstein, J. (2004) Links between attachment theory and systemic
practice: some proposals. Journal of Family Therapy, 26: 2–16.
Amato, P. R. and Afifi, T. D. (2006) Feeling caught between parents: adult
children’s relationship with parents and subjective well-being. Journal of
Marriage and the Family, 68: 222–235.
Anderson, H., Goolishian, H. and Winderman, I. (1986) Problem-determined
systems: toward transformation in family therapy. Journal of Strategic and Family
Therapy, 4: 1–13.
Bateson, G. (1972) Steps to an Ecology of Mind. New York: Ballantine.
Blow, K. and Daniel, G. (2002) Frozen narratives. Post-divorce processes and
contact disputes. Journal of Family Therapy, 24: 85–103.
Bowen, M. (1978) Family Therapy in Clinical Practice. Lanham, MD: Jason Aronson.
Bowlby, J. ([1969] 1982) Attachment and Loss, Vol. I: Attachment. London: Hogarth
Press.
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.
Buchanan, C. M., Maccoby, E. E. and Dornbusch, S. M. (1991) Caught between
parents: adolescents’ experience in divorced homes. Child Development, 62:
1008–1029.
Buehler, C. and Welsh, D. P. (2009) A process model of adolescents’ triangulation
into parents’ marital conflict: the role of emotional reactivity. Journal of Family
Psychology, 23: 167–180.
Byng-Hall, J. (1991) The application of attachment theory to understanding and
treatment in family therapy. In C. M. Parkes, J. Stevenson-Hinde and P. Marris
(eds) Attachment Across the Life Cycle (pp. 199–215). London: Tavistock and
Routledge.
Byng-Hall, J. (2008) The significance of children fulfilling parental roles:
implications for therapy. Journal of Family Therapy, 30: 147–162.
Cooper, J. and Vetere, A. (2005) Domestic Violence and Family Safety: A Systemic
Approach to Working with Violence in Families. Chichester: Wiley.
Crittenden, P. M. (1997) Truth, error, omission, distortion, and deception:
an application of attachment theory to the assessment and treatment of
psychological disorder. In S. M. Clany Dollinger and L. F. DiLalla (eds)
Assessment and Intervention Issues Across the Life Span (pp. 35–76). London:
Lawrence Erlbaum.
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.
Cummings, E. M., Vogel, D., Cummings, J. S. and El-Sheikh, M. (1989)
Children’s responses to different forms of expression of anger between adults.
Child Development, 60: 1392–1404.
Dallos, R. (2006) Attachment Narrative Therapy: Integrating Attachment, Systemic and
Narrative Therapies. Maidenhead: Open University Press and McGraw Hill.
Dallos, R. and Draper, R. (2005) An Introduction to Family Therapy. Maidenhead:
Open University Press/McGraw Hill.
Systems Theory, Family Attachments and Processes of Triangulation 19
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
Systems Theory, Family Attachments and Processes of Triangulation 135
Dallos, R. and 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. and Vetere, A. (2009) Working Systemically with Attachment Narratives:
Applications in a Range of Clinical Settings. London: Routledge.
Davies, P. T., Harold, G. T., Goeke-Morey, M. and Cummings, E. M. (eds) (2002)
Child emotional security and interparental conflict. Monographs of the Society for
Research in Child Development, 67: vii–viii.
Dowling, E. and Gorell Barnes, G. (1999) Working with Children and Parents
Through Separation and Divorce. Basingstoke: Macmillan.
Flaskas, C. (1997) Re-claiming the idea of truth: some thoughts on theory in
relation to practice. Journal of Family Therapy, 19: 1–20.
George, C., Kaplan, N. and Main, M. ([1986] 1996) Adult attachment interview.
Unpublished manuscript. Berkeley: University of California.
Haley, J. (1987) Problem Solving Therapy. San Francisco, CA: Jossey Bass.
Hill, J., Fonagy, P., Safier, E. and Sargent, J. (2003) The ecology of attachment in
the family. Family Process, 42: 205–221.
Minuchin, S. (1974) Families and Family Therapy. Harvard: Harvard University
Press.
Minuchin, S., Rosman, B. and Baker, L. (1978) Psychosomatic Families: Anorexia
Nervosa in Context. Cambridge, MA: Harvard University Press.
Palazzoli, M. S., Cecchin, G., Prata, G. and Boscolo, L. (1978) Paradox and Counter
Paradox. New York: Jason Aronson.
Ringer, R. and Crittenden, P. M. (2006) Eating disorders and attachment: the
effects of hidden family processes on eating disorders. European Eating
Disorders Review, 14: 1–12.
Rutter, M. (1999) Resilience concepts and findings: implications for family
therapy. Journal of Family Therapy, 21: 119–144.
Schindler, A., Thomasius, R., Sack, P., Gemeinhardt, B. and Kustner, U.
(2007) Insecure family bases and adolescent drug abuse: a new approach to
family patterns and attachment. Attachment and Human Development., 9: 11–126.
Schore, A. (2009) The paradigm shift; the right brain and the relational
unconscious. Plenary Address, Annual conference of the American Psychological
Association, 8 August, Toronto, Canada.
Titelman, P. (ed.) (2008) Triangles: Bowen Family Systems Theory Perspectives. New
York: Haworth Press.
Vetere, A. and Cooper, J. (2005) Children who witness violence at home. In A.
Vetere and E. Dowling (eds) Narrative Therapies with Children and Their Families
(pp 75–89). London: Routledge.
Vetere, A. and Cooper, J. (2006) The effects of domestic violence on children:
trauma, resilience and breaking the cycle of violence. Journal of Critical
Psychology, Counselling and Psychotherapy, 1: 26–38.
Vetere, A. and Dallos, R. (2008) Systemic therapy and attachment narratives.
Journal of Family Therapy, 30: 374–385.
Vetere, A. and Gale, A. (1987) Ecological Studies of Family Life. Chichester: Wiley.
Vetere, A. and Henley, M. (2001) Integrating couple and family therapy into a
community alcohol service – a pan theoretical approach. Journal of Family
Therapy, 23: 85–101.
20 Rudi Dallos and Arlene Vetere
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
136
Weakland, J. (1976) Toward a theory of schizophrenia. In C. E. Sluzkiand and D.
D. Ransom (eds) Double Bind: The Foundation of the Communicational Approach to
the Family. New York: Grune and Stratton.
Winnicott, D. W. (1964) The Child, the Family and the Outside World. Cambridge, MA:
Perseus.
Woodcock, J. (2009) Love, hate and the Oedipal myth. In D. Pocock and C.
Flaskas (eds) Systems and Psychoanalysis (pp 129–146). London: Karnac.
Systems Theory, Family Attachments and Processes of Triangulation 21
r 2011 The Author(s)
Journal of Family Therapy r 2011 The Association for Family Therapy and Systemic Practice
Systems Theory, Family Attachments and Processes of Triangulation 137