THEORY

Bowen Systems Theory

and Mediation
'Waym Re^na
The field of mediation is largely technique- and process-driven. In an effort
to develop a more comprehensive theoretical system for mediation, this article applies Bowen Systems Theory to mediation. The article be^ns with a
brief overview of the development of Bowen Systems Theory; it proceeds to
define its central concepts, including differentiation, triangulation, and chronic
anxiety: firmlly, the article applies these concepts to mediation and discusses
the relationship between theory and technique, using caucusing and comediation as examples. The article concludes that Bowen Systems Theory
naturally extends to mediation, providing thisfieldwith a comprehensive theoretical system.

Much has been written about training effective mediators and the stages of
mediation itself (Moore, 1996; Schwartz, 1994; Gray, 1989; Lax and Sebenius,
1986). There is also a great deal of literature about mediation as a successful
example of alternative dispute resolution (Avruch, 1998; Constantino and
Merchant, 1996). The results of research clearly show that mediation is an
effective tool in creative conflict resolution, empowering others to make
healthy choices in their lives and promoting the causes of peace, environmental justice, and social justice (Maser, 1996; Roberts and Lundy, 1995; Suskind,
Babbit, and Segal, 1993; Stamato andjaffe, 1991; Ury, Brett, and Goldberg,
1988). Thefieldof mediation, though, is largely technique- and process-driven
and lacks comprehensive theoretical perspectives, upon which effective academic disciplines are based.
This article is the first part of a larger work that proposes one such theoretical system. It examines the interplay of Bowen Systems Theory and mediation and views mediation as a system that can be understood by using Bowen
Systems Theory. 1 vwU demonstrate that, like all other examples of human systems, effective mediation is a product of the interplay between various Bowen
Systems Theory concepts, including differentiation, triangulation, and chronic
anxiety, as well as the mediation process itself.
MEDIATION QUARTERLY, vol. 18, no. 2, Winter 2000 O Jossey-Bass, a Wiley company

111

112

Regina

A brief review of relevant Bowen Systems Theory concepts is presented;

then the theory is applied to mediation. This article cannot possibly do justice
to the complexity and extensiveness of Bowen Systems Theory. As a theoretical system, Bowen Systems Theory is over forty years old, and to treat it fairly,
1 would have to write an entire book explicating the theory. It is elaborate, elegant, complex, and simple all at once, and overarching as well. Other writers
have more fully articulated the theory, so I will not attempt to provide a comprehensive overview here. Rather, I refer the reader to those who have done
justice to the theory, including Bowen (1978), himself, as well as Friedman
(1991), Papero (1990), Ken (1981), and Kerr and Bowen (1988). Here, I will
provide a brief summary of the most applicable components of the theory as
it relates to mediation, and I trust that the reader will investigate the theory
further with the depth that it so richly deserves.
In the 1950s and 1960s, psychoanalysis, behaviorism, and the so-called
third-force humanistic perspectives dominated psychology and psychiatry
(Garfield, 1981; Rogers, 1951, 1961; Frankl, 1956; Maslow, 1954; Skinner,
1938). All three of these theoretical systems explicated the primacy of the individual in understanding human behavior and in treating psychological disorders. It was in this environment in the 1950s that Murray Bowen began his
professional career as a psychiatrist, except that he, like other radical psychiatrists and psychologists of the time, began rejecting the dogma of theory and
treatment based on individual perspectives (Framo, 1996; Wylie, 1991;
Bowen, 1971; Bowen, 1959; Kerr and Bowen 1988). These pioneers began
incorporating new approaches to treatment and formulating new theories for
health and pathology. Out of this creative firmament, family systems theory
was born in its many manifestations. While others developed object relations
family therapy (Dicks, 1967; Framo, 1965), communications theory (Jackson,
1961; Wynne, Ryckoff, Day, and Hirsch, 1958; Bateson, Jackson, Haley, and
Weakland, 1956), experiential family therapy (Whittaker, 1958; Whittaker
and Malone, 1953), and strategic theory and therapy (Haley, 1963), Bowen
was bold enough to articulate a comprehensive theory of human functioning
based on natural systems (Bowen, 1966; Kerr and Bowen, 1988).
According to Bowen, his Family Systems Theory was only one example of
the application of natural systems (Kerr and Bowen, 1988). In stark contrast
to the prevailing paradigms of the day, Bowen postulated that humans were
more similar to than different from other forms of life. And like "lower" forms
of life, Bowen said that much of our behavior is automatic, based on all that
was "written" before. He termed this depth of automatic functioning and reactivity "emotional" and said that the moth drawn to the flame, the ant storing
food for the colony, the fish swimming in tight school formations, and the
woman yelling at the driver who cut her off are all examples of emotional functioning, that is, behavior that is automatic, reactive, and based in a multiplicity of factors (Friedman, 1991). In humans, these factors include instinct.

Bowen Systems Theory and Mediation

113

multigenerational family history, experience, and biology. While saying that

humans are deeply connected to and thus similar to all other life forms, Bowen
theorized that humans are also unique and different from all other life forms
through the evolution of both a highly developed feeling system (as contrasted
with the "deeper" emotional system) and a thinking system (Kerr and Bowen,
1988). In fact, Bowen said that it is the thinking systemand its most evolutionarily advanced features in the prefrontal lobe, where higher-order thinking takes place (planning, foresight, and so forth)that allows us the
opportunity to respond differently from our animal cousins.
One of the hallmarks of Bowen Systems Theory, then, is the notion that
humans are social animals who have the capacity to move beyond their systems "programming" and respond autonomously rather than simply react automatically (Friedman, 1991). While we do have this innate capacity for
independent response, Bowen said that it must be cultivated or else we are
likely to emotionally react rather than respond creatively and individually.

Bowen Systems Theory

Over the years, Bowen refined his theory of human systems. As a systems
theory, it has been applied to leadership and organizational systems (Regina,
2000; Ferrera, 1996; Friedman, 1996; Gilbert, 1996; McCullough, 1996;
Bowen, 1974), educational systems (Dillow, 1996), congregational systems
(Friedman, 1985), human society (Friedman, 1991; Papero, 1990), nonhuman
systems (Kerr, 1997; Muraskin, 1994), and family systems (Schnarch, 1997;
Regina and LeBoy, 1991; Carter and McGoldrick, 1989; Kerr and Bowen,
1988). The same principles that apply to other human systems make Bowen
Systems Theory a natural fit in examining mediation.
This article reviews several important tenets of Bowen Systems Theory and
demonstrates its application to the field of mediation. Included in this discussion is the role and function of differentiation in the mediation system (basic
and functional), triangles and triangulationboth within the mediation
environment and in assessing and de-triangulating situations so that agreements can be reached, the role of the mediator in monitoring and reducing
anxiety in the mediation systemincluding his or her own chronic and situational anxiety, and theory and technique as viewed through caucusing and
comediation.
. One of the challenges in presenting an overview of Bowen Systems Theory is that the concepts are highly interdependent and related. As such, it is
difficult to explain one concept without using other concepts that have not yet
been fully articulated. Nonetheless, the overview contained in this article is an
attempt to explain the concepts within the context of mediation. Where related
concepts are introduced early, the reader's patience is requested, as the interrelationship of the concepts will eventually be made clear.

Regina

Differentiation
The concept of differentiation is complicated. Many authors focus on differing
aspects (Friedman, 1991; Bowen, 1978; Kerr and Bowen, 1988). For this article, 1 will focus on differentiation as the interplay and expression between the
individuality and togetherness life forces (Kerr and Bowen, 1988). Differentiation, then, is the capacity for self-definition and self-regulation, that is, being
one's ovm person (the individuality life force) while remaining authentically
connected to others (the togetherness life force). Differentiation is not static
interpersonally or intrapersonally; rather, it is individualistic (one individual
is more differentiated than some and less differentiated than others) and
dynamic within individuals (depending on the level of strain within which an
individual finds himselQ. While one's basic level of differentiation is a function
of many factors, including that which is emotionally "inherited" from one's
family, it tends to remain fairly static throughout life unless one actively works
to increase it. A person's functional level of differentiation changes as a result of
strains he or she is going through. That is, a person in a relatively calm and
stable environment may look very healthy and differentiated. If the strains
and pressures increase, however, that person looks less adaptive and more
reactive. The higher the basic level of differentiation, the less the change in
functional differentiation over time and across situations.
In sum, then, a more highly differentiated person is more adaptive. She is
less reactive, her responses are more consistent, autonomous, and thoughtful
in a variety of stressful and nonstressful situations, she is more emotionally free
and capable of responding to different situations with different responses, and
she is spontaneous, has more resources, and is expressive of her feelings
and thoughtful in her responses vwthin a given context. In essence, a more
highly differentiated person is a more mature person, v^alling "to take responsibility for [her] emotional being and destiny" (Friedman, 1996, p. 16).
In contrast, a less differentiated person has fewer resources and is less
adaptive. Like a rubber band stretched too often, he lacks the resiliency to
bounce back and reacts more automatically. He is less emotionally free, and his
reactivity can take many forms, including reactive thinking and reactive feeling. He is less focused on his ovwi internal compass and is more regulated by
those in his present and past. He will use cutoff and triangulation to manage
conflict.

Differentiation and Mediation

The basic level of differentiation of the parties in mediation is important. Parties with moderate to high levels of differentiation are likely to mediate a solution that will endure over time. They realize that mediation is more likely to
get them what they want after exploring their BATNA (best alternative to a
negotiated agreement). They realize that mediation can create a win for them.

Bowen Systems Theory and Mediation

115

they are not as attached to positions, and they are more likely to respond to
interest-based negotiations. In fact, it may be that the more highly differentiated the person, the more she can undertake interest-based negotiations, as
she is less reactive, more "objective," and can separate positions from interests.
In contrast, the less differentiated a person, the more likely she will be emotionally reactive in her thinking and feeling, the more attached she will be to
a positional outcome, as the position is viewed as a fused extension of the self,
and the less likely it is that a mediation agreement v^dll survive the test of time.
Other factors, of course, influence the mediation process and outcome as well,
especially the level of current stressors on the parties and their systems (see my
discussion of a two-factor theory of functioning, coming up later).
The basic level of differentiation of the mediator is also important in securing a successful mediation. More highly differentiated mediators are more capable of responding to a variety of situations and conflicts. They are not attached
to certain outcomes in the mediation. Rather, they can function admirably in
highly charged situations and can tolerate elevated levels of ambiguity.
Although not attached to outcomes, these mediators are very involved in the
process, staying connected with all the parties. They clearly outline the structure and roles of all involved, are self-defined so as to control the mediation
process without controlling the outcome, and use appropriate humor to deescalate the seriousness that highly invested people bring to the negotiating
table. More highly differentiated mediators may find themselves more attracted
to one of the parties or to one of the proposed solutions over another, but they
are capable of remaining fairly neutral and more objective. Like coaches for a
team, they help bring out the best in the "players" without getting on the field
and playing themselves. They neither overfunction nor underfunction out of
anxiety. That is, they create an environment where all of the parties can
fully and appropriately participate in their respective roles.
In contrast, less differentiated mediators see themselves as directors and
players rather than as coaches. They are often attached to outcomes, becoming
triangulated in the conflict by siding with one individual or group over the
other. They are more emotionally anxious and react from subjective positions
that are usually not in the best interests of the mediation process. Less differentiated mediators do not tolerate highly charged situations well. They seek certainty and control and are uncomfortable in ambiguous environments. Rather
than having fun with the process, they experience anxiety, anger, frustration,
and other feelings that often emotionally drive the mediation. Anxiety either
makes them overfunction and so they are too activeinfluencing the parties to
underfunction and not take full responsibility for the mediationor it makes
therh underfunction, thereby not providing an appropriate vessel within which
the parties can safely and fully explore the issues. Agreements may merely be
attempts by the parties to end the mediation and therefore may not stand the
test of time. More often than not, mediation breaks down as one of the parties
becomes aware of the nonneutrality and reactivity of the mediator.

116

Regina

When applying Bowen Systetns Theory, it is important to not dichotomize

the concept of differentiation. Differentiation is not an either-or phenomenon
(one is not either differentiated or not differentiated). Rather, differentiation
exists on a continuum in functioning. Individuals in every human (and nonhuman) living system span the range of differentiation, be they in families,
organizations, or school systems. In addition, as humans, we are all capable of
modifying some of our behaviors and choosing the timing of many of our
actionsthat is, responding more objectively.
For mediators, there are certain guidelines that can be helpful in functioning more effectively as we are all responsible for creating an environment
that promotes basic and functional differentiation. For example, mediators can
self-monitor and self-regulate their own level of strain. If at all possible, mediators can schedule mediations at times when they are least stressed or at times
of the day when they function best. In addition, mediators must be aware of
their emotional triggers. All mediators experience more difficulty with certain
individuals or situations than with others. For example, if mediating a child
custody case in which child abuse has occurred and the mediator has strong
negative reactions to perpetrators of abuse, he can pass the case on to another
mediator or work with a comediator who may not have that same sensitivity.
Also, mediators can monitor the states of one or both of the parties. Mediation
is more likely to succeed when the parties are calm than when they are anxious and agitated. While parties in mediation are by definition in conflict, the
environment the mediator helps create can either help reduce the anxiety in
the room, thereby increasing functional differentiation and increasing the
chances of mediation success (defined as the parties becoming clear about
what they want, which may or may not mean continuing mediation to find a
mutually acceptable solution), or it can increase anxiety, thereby decreasing
functional differentiation and decreasing the chances of mediation success.
Creating an emotionally safe space for mediation does not mean that the
mediator avoids escalating tension and conflict when that intensification is
needed. In fact, escalating tension to create disequilibrium is often very helpful
in moving people from positional to interest-based negotiations. Disequilibrium that is created through escalating conflict and confrontation, however,
must be managed through the mediation triangle (see my discussion of triangles and triangulation, coming up later) in order to maximize the potential for
success.
In addition, supervisors and trainers must monitor the people who are
being trained. Not everyone functions well in highly stressful, highly ambiguous environments. Assisting trainees and beginning mediators vwth techniques
may be of some benefit. There are times, however, when trainers and supervisors should recommend a different occupation. Much of the profession is selfand peer-regulated. As such, it is in the best interest of everyonetrainees,
supervisors, the public, and members of the professionthat we not shy away
from helping certain individuals self-select out of the profession. This is done in

Bowen Systems Theory and Mediation

117

every self- and peer-regulated profession. Mediators should be vigilant about not
fostering those without the appropriate differentiation levels to practice in a new
and growing occupation or impose themselves on an unsuspecting public.
Making a long-term commitment to increase one's own differentiation, of
course, is always the best opportunity to decrease reactivity and increase one's
own functioning (Friedman, 1991). Individuals throughout history have faced
enormous obstacles and overcome them through personal growth and transformation. Many avenues have been available, including meditation, psychotherapy, martial arts, and spiritual practices. Unfortunately, many growth
disciplines tend to foster dependency, and while they may temporarily increase
functional differentiation, they do not necessarily promote increases in basic differentiation. Coaches and supervisors trained in Bowen Systems Theory can,
over time, assist individuals and groups to achieve modest increases in basic
differentiation.
A Two-Factor Theory of Functioning Applied
to Mediating Parties
Friedman (1991) conceptualized functioning by using a two-factor graph with
a differentiation level from 0 to 100 on the y-axis and current strainthat is,
life stressors, from 0 to 100, on the x-axis. This four-quadrant model looks at
response/reaction and readaptation as a function of one's position on the graph.
Persons with low to moderate strain (0-50) and moderate to high levels of
differentiation (50-100) are in Quadrant 1. They manage their current life challenges well. Triangulation is minimal; that is, they deal with conflict and stress
directly, their resources are plentiful, and their adaptation and readaptation
are high. These individuals do the best in mediation as participants and as mediators. They separate positions from interests, are highly adaptive, are clear
thinkers, and are capable of more objectivity in their thinking and behavior.
Persons with low to moderate strain (0-50) and low to moderate levels of
differentiation (0-50) are in Quadrant II. They find it more formidable meeting their current life challenges. Even though their current stressors are not
overwhelming and may in fact be about the same as they are for those in
Quadrant 1, these people do not have the resources that those in Quadrant 1
possess, as they have more inherent chronic anxiety As such, their adaptation
ability is compromised. They triangulate more; that is, they channel their anxiety through alcohol, drugs, blaming others, affairs, and so on. They do not
have as many available resources and often find themselves in the legal or
social service systems that force solutions on them. If stress on the x-axis can
be reduced, these individuals can often attain a higher level of functional
differentiation and often do fine in society. It is recommended that people
in Quadrant II work to decrease their x-axis stressors to the degree possible
before undertaking mediation. Of course, this is not always possible and, in
fact, the conflict that brings them into mediation may be exacerbating their

118

Regina

maladaptive functioning. These individuals need more "connection" from the

mediator. The mediator must be careful to stay in contact with these people
without siding with them and thereby creating a triangle with the other party
on the outside position.
Individuals with moderate to high strain (50-100) and moderate to high
levels of differentiation (50-100) are in Quadrant III. They may appear multisymptomatic and poorly functioning for a time. Some are experiencing moderately high stress that is temporarily overwhelming. Others on the higher end
of the X-axis have experienced a catastrophic loss or incident and so appear to
be less healthy than they actually are. That is, their functional level of differentiation is lower than their basic level of differentiation. High-end x-axis situations include such stressors as the untimely loss of a family member; being
a victim of war, terrorism, or natural disaster; and experiencing a serious illness, trauma, or accident. These individuals respond well to care and treatment and, despite their current life stressors, they have the internal and
external resources necessary to help in their recovery. Prognosis is good. The
events themselves may become life lessons and these people often eventually
retum to their lives scarred but functional. Many are in fact transformed by
their experiences and dedicate their lives to serving others.
Even though these people are generally high in their functioning, their
current life circumstances mitigate against mediation at this difficult time. If
possible, mediators should delay mediation with these individuals until they
have achieved some semblance of equilibrium, after which they are likely to
do well in mediation. For example, in a victim-offender mediation, a victim in
a highly charged emotional state is likely to be retraumatized in mediation and
must reestablish emotional equilibrium before confronting an offender.
Individuals with moderate to high strain (50-100) and low to moderate
levels of differentiation (050) are in Quadrant IV These are people vnlh difficult or catastrophic life situations and they are the least capable of managing
them. They greatly consume society's resources and are ever present in the system, be it the courts, prisons, or social services. Often, their lives are in constant turmoil and crisis, and they are on a downward spiral. Sets of poor
decisions or unfortunate events create other stressors. These individuals are
the poorest candidates for mediation. They tend to see themselves as victims,
blaming others for their problems. It is often preferable for them to take a legal
recourse, as this provides the structure they need, and the solution provided
by the courts is at least right or wrong in their eyes. Managing ambiguity and
interest-based negotiations is not usually possible for them.
In every quadrant, reductions in the x-axisthat is, reductions in
stressorswall lead to increased functioning. Reducing chronic anxiety and
increasing differentiation is harder and takes much longer. Even when using a
two-factor system for understanding functioning and adaptation, assessing
a person's differentiation level at any moment in time is difficult. For accuracy,
differentiation must be assessed over a period of time and a variety of situations.

Bowen Systems Theory and Mediation

119

To the degree possible, then, mediators are wise to observe their clients functioning over time. A person who comes in with impediments in tracking the
process may be in a temporarily difficult situation but may do much better as
he or she begins the process of adapting and learning. Mediation itself is a
stressful situation. People are sometimes court referred. As such, they may be
present only because they have to be. Union leaders cany the burden of representing a large number of people; business leaders must answer to corporate
officials. In multiparty mediations, representatives from various stakeholder
groups must balance the needs of the stakeholders with the needs of creating
an agreement that may be the best they can obtain, given the complexities
of the situation and the realities of the courts. Parents in conflict over a marriage that is ending are sometimes predisposed to triangulate the children and
use them as weapons in a battle to hurt each other again.
As mediators, we are not fortunate enough to work exclusively with individuals in Quadrant I. In addition, except for marital dissolution and child custody cases, it is more likely that the differentiation levels of the mediating
parties will be mixed on the quadrants. Bowen Systems Theory hypothesizes
that people marry at the same level of differentiation, even though one may
appear more differentiated at a given time because of the "borrowing and trading of self that occurs in intimate relationships (Friedman, 1985). Understanding the two-factor approach to functioning, however, will help mediators
assess, both initially and over the course of the mediation, whether mediation
is or remains a viable option.
Understanding how differentiation levels interact with current life stress
levels also helps mediators avoid prejudging the outcome of a mediation based
on one or two sessions and helps mediators avoid being triangulated into the
system the parties in conflict have created.
Triangles, Triangulation, and De-triangulation. Bowen (1978) said that
for humans the basic emotional building block is the triangle. In essence, he
hypothesized that a twosome is inherently unstable and will, under stress,
automatically seek to restabilize itself by bringing in a third. This restabilization attempt is called triangulation. Usually, the third point of the triangle is a
person. The twosome may, however, triangulate an object (for example,
a house or other possession, money, alcohol, or drugs) or an idea (for example, religion, political conviction, sex). Recognition of the universahty of the
triangle, of course, is not new. Many cultures throughout time have beheved
in archetypal symbols, which, in addition to the triangle, have included the
cross, the circle, the square, and other two- and three-dimensional objects and
symbols, such as the sphere, cube, and pyramid. For Bowen, though, the triangle was the most important symbol of human functioning. In Bowen
Systems Theory, the sides of the triangles are seen as conduits by which emotional energy or anxiety moves. They designate which two are to be on the
inside positions of the triangle and which one is to be on the outside position
of the triangle. Those on the inside positions usually keep their "closeness" by

120

Regina

keeping the third out; those in conflict on the inside seek to move outside to
reduce conflict, to move to a new inside position with the third, forcing the
other to the outside position, or to create a new, interlocking triangle (Kerr and
Bowen, 1988).
Triangles and triangulation are dynamic and complex processes. A central
triangle will form interlocking triangles when further stressed and when anxiety needs additional places to flow. This emotional energy flow is called chronic
arvdety. Understanding chronic anxiety is essential to understanding triangles.
In Bowen Systems Theory, anxiety is not a surface state equated with the
feeling system. Rather, anxiety is chronica reaction more commensurate with
the deeper (older) emotional system, and it is inversely related to a person's
degree of differentiation. In this model, then, anxiety is a reptilian response. That
is, it is deeply rooted in the organism and is automatically activated in people
to greater or lesser degrees, based on two factors: the level of differentiation of
a person (which is inversely and directly correlated to the level of chronic
anxiety) and the level of strain on the system (current life stressors) (see
my previous discussion of a two-factor theory of functioning). Anxiety is
usually managed automatically through the formation of triangles and the
resultant process of triangulation (Bowen, 1978). Triangulation, then, is a selfpropelling process by which anxiety moves through a system. It is an automatic way that two people in conflict bring in a third to lower the anxiety of
the system, attain temporary relief from the suffering that conflict brings, and
try to achieve greater temporary balance.
By definition, the two parties in mediation are triangulated by the conflict.
The emotional flow of anxiety is bound up in the issue, person, or thing being
mediated. The issue, person, or thing, then, becomes the anxiety binder for the
parties. Anxiety binders are third points of triangles (Kerr and Bowen, 1988).
Anxiety binders can be people, places, things, ideas, businesses, and so forth.
All anxiety binders help create inside and outside positions. All anxiety binders
are emotional attachments and are, as such, fused extensions of the undifferentiated self.
In mediation, each party is usually vying for the inside position of the triangle with the anxiety binder. Two business partners trying to dissolve their
corporation, for example, may get stuck over distribution of resources, current
employees for the new company they each want to form, or the current company name. All can become anxiety binders whereby emotional attachments
create positional stances and detract from interest-based negotiations. The
stronger the degree of fusion or attachment with the anxiety binder, the more
the person's self is wrapped up in and invested with the anxiety binder^which
makes it more difficult to reach a mediated solution. Positions become hardened and impasses are likely.
In these situations, which are more or less present in all mediations, the
mediator must attempt to form what 1 term a mediating triangle. A mediation
triangle is an attempt by the mediator to insert herself into the system by

Bowen Systems Theory and Mediation

121

connecting equally with each party (similar to the equidistant sides of an equilateral triangle). That connection must be as differentiated as possible. That is,
the mediator must have a strong, nonanxious presence, clear boundaries, and
a firm capacity for self-definition and self-regulation. In addition, she must connect with the parties through empathy, understanding, and clarity about each
person's role in the mediation system. Through that differentiated connection,
it is possible to lessen the fusion that attaches the parties to the anxiety binder
by temporarily increasing the functional differentiation of the mediation system. In a best case scenario, the mediating triangle can become a differentiating
triangle. I have developed the concept of the differentiating triangle to present
a counterbalance to the concept of triangulation. That is, in triangulation, two
people in conflict deal with anxiety by bringing in a third to bind it for them,
temporarily creating stability and reduction in short-term anxiety, while simultaneously raising long-term anxiety and decreasing both functioning and differentiation. An example of this is a couple that has triangulated alcohol in their
relationship. The husband using alcohol may create a teniporary decrease in
anxiety widiin the system by drinking to reduce his craving, but the more he
drinks, the worse the alcoholism in the system becomes. In contrast, a differentiating triangle is one in which three people or two people and a third "other"
are connecting in such a way that great creativity, problem solving, or other
higher-level functioning and adaptation is possible. For example, two parents
can leam and grow from their child in ways that help each become more differentiated,'which in tum strengthens the family and promotes differentiation
in the child. Two composers can be in a differentiating triangle such that
together they write an exquisite piece of music.
It is theoretically possible that a mediator can bring such a highly differentiated self to the mediation so as to shift the emotional field of the mediation system to a "higher" level of differentiation. 1 believe that the most skilled
mediators in the profession are capable of such emotional field shifts. In these
situations, the parties will say that the mediator brought out the best in them.
The parties may walk away not only with a solid agreement that holds over
time but also with a true understanding of and respect for the other party In
this scenario, mediation itself can be a transformative experience.
The key to managing triangles and anxiety is to help people tolerate shortterm increases in anxiety for the sake of long-term reductions (Kerr and
Bowen, 1988). That is, by keeping the parties in contact with a mediator who
is minimally reactive, moderately to highly differentiated, and calm, the system can seek interest-based altematives to the current conflict without resorting to habitual, positional patterns of responding. A mediating triangle and a
differentiating triangle, in effect, can help loosen the chokehold of fusion with
the anxiety binder and the positions, freeing individuals to make more
autonomous, interest-based choices rather than reactive, automatic ones.
This strategy is called detriangulation (Kerr and Bowen, 1988; Friedman,
1985). To be effective in detriangulating, the mediator must first recognize that

122

Regina

triangulation is occurring. This requires that the mediator label the central and
interlocking triangles of the system. If, for example, two business partners are
trying to dissolve their business relationship, the partners and the business
are likely to be the central triangle. Other interlocking triangles may, however,
be significantly interfering with their capacity to manage this triangle. In this
example, a spouse of the partner may be overinvolved, thus complicating the
negotiations between the business partners. Knowing the central and interlocking triangles will help the mediator focus on detriangulating the central
triangle and free the pariies from some of the anxiety bound in the interlocking triangles.
Second, the mediator must remember that triangulation is automatic and
that everyone does it; it is a natural function of all human systems. In mediation, the parties are usually trying to triangulate the mediator into the system
by creating an interlocking triangle with the mediator and the conflict on the
inside position with one party and the other party on the outside position.
(Two or more persons, places, or things can occupy a point on a triangle.) The
parties are looking for relief from their suffering; they seek an ally to use as a
weapon against each other. Mediators must be very careful not to automatically
become part of the triangled system. Being aware of this automatic process will
make it easier to avoid. The triangle is there. Whether or not the mediator is
triangled into the conflict is a function of the mediator's ability to remain relatively objective, calm, and self-defined in the situation and connected to the
parties. That is, remaining detriangled is a function of his or her ability to
remain relatively differentiated in the system and to not absorb the anxiety generated by the parties in mediation.
J positions reduce anxiety built up around conflict and they discourage
channeling that anxiety through positions in the triangle (Kerr and Bowen,
1988; Friedman, 1985). The hallmarks of detriangulation, I positions are effective in any multi-person system that is under strain. They are self-definitional
statements designed to strengthen the self and stay connected to the mediating parties. For example, take the statement "I am here to help you both find
solutions that are mutually acceptable. 1 am not here to judge or take sides."
A statement such as this is a standard part of the mediator's introduction. As
an 1 position, the intention of the mediator is less to communicate the mediator's role to the parties than to define one's position while staying in contact
with others, the twin markers of differentiation. In the previous example, the
mediator is self-defining his position while keeping in contact with the parties,
thus avoiding being caught in the system.
As previously stated, the most important concern for the mediator is to
not absorb the anxiety of the mediating parties. If the mediator absorbs the
anxiety for the system, she becomes another anxiety binder for the parties.
Each of the parties is likely to vie to bring the mediator to the inside position
of the triangle and keep the other party on the outside. The job of the mediator is to recognize the triangle and refuse the invitation to participate (that is.

Bowen Systems Theory and Mediation

123

avoid being triangulated). Here, too, 1 positions are the key to remaining calm
and detriangulated. The mediator must be empathetic and state clearly that
she understands how upsetting this situation or event is and that her role is
not to take sides but to help the parties arrive at mutually agreed upon solutions. By remaining empathetic and detriangulated, the mediator can be
supportive, can encourage (as opposed to direct) a more healthy response to
the conflict, and can avoid absorbing another's anxiety. Responses such as these
promote role clarity, healthy functioning, and appropriate differentiation in the
system; that is, they promote the system's abihty to manage and reduce anxiety on its own and find appropriate solutions.
In sum, in all systems, including the mediation system, it is essential to
recognize that triangulation is a natural response to anxiety. Seeing triangles is
essential for creating clarity, decreasing anxiety, and perpetuating emotional
calmness, like any skill and growth-promoting activity that seeks to increase
differentiation, seeing triangles, stating I positions, and detriangulation all take
constant work and rehearsal. This effort becomes more successful the more it
is practiced. The success of any mediation rests on the differentiation of the
mediator and his ability to form a mediating triangle, thus avoiding being triangulated by the emotionally charged system.
Technique and Theory. In North American models of mediation, training
strongly relies on technique (Moore, 1996). Although a technique-based
approach to training mediators and conducting mediation is certainly pragmatic, as an academic discipline, the field of mediation will benefit from a
stronger and more comprehensive theoretical foundation. Technique devoid
of theory will only advance the profession of mediation to a certain point. A
systemic understanding of the relative roles of the mediator and the mediating
parties, as well as the mediation environment itself, will assist the profession
to evolve and become even more successful. Bowen Systems Theory, as an
example of the application of natural systems in general and human systems
in particular, seems uniquely suited to provide a comprehensive theoretical
model for understanding ourselves as mediators and the mediation process. In
addition, having a stronger theoretical position will promote more success
in implementing techniques, for the techniques become ways of promoting
differentiation of the parties, detriangulation in the system, more effective functioning of the mediators, and, ultimately, more durable agreements that can
withstand the test of time.
Two examples of using theory to increase the effectiveness of techniques
are discussed. The first example concerns caucusing and the second example
focuses on comediation. In both instances, I will demonstrate how viewing and
using techniques from the theoretical lens of Bowen Systems Theory can
increase the understanding and effectiveness of the techniques themselves.
Caucusing from a Bowen Systems Theory Perspective. The technique of
caucusing is a standard tool in the mediator's repertoire (Moore, 1987; Bethel,
1986). Mediators know the benefits and dangers of using caucusing. Some use

124

Regina

it routinely; others refuse to use it at all. When conceptualized through a BST

perspective, caucusing is a technique in which the mediator is extremely vulnerable to triangulation. By definition, meeting in caucus creates inside and
outside positions. The emotional anxiety that is a part of being on the outside
position for one of the parties is what makes caucusing such a high-risk technique. Nonetheless, caucusing can be a powerful tool for breaking impasses
(Moore, 1996). It allows mediators the opportunity to discover important
information, build stronger relationships vnth one or both of the parties, do
reality testing, check for safety issues, and in other ways determine the causes
of impasses. On the other hand, caucusing can also create a triangled situation
that makes breaking impasses more difficult. As a technique, mediators are
trained to meet with both parties separately during the caucus. At a technical
level, mediators know that caucusing can create the illusion of partisanism.
The emotional consequence of that partisianism, real or imagined, is triangulation. Understanding triangulation, detriangulation, and the mechanisms by
which anxiety is likely to flow in mediation will assist mediators in using caucusing wisely.
To revisit the example of two business partners terminating their business:
during a caucus with party A, the mediator finds out that party A is unwilling
to tell party B about his wife's insistence that any agreement he approved by
her. His wife is angry and fearful over the impending corporate dissolution and
is reluctant to agree to any terms that may quicken the end of the business
partnership. Party A asks the mediator not to reveal this to party B, and the
mediator had agreed at the beginning of the caucus that information obtained
in the caucus would be kept confidential. The mediator has been put in an
inside position of the triangle with party A regarding this information. The
degree to which the mediator is triangled depends on her response both to
party A and to the mediation when it resumes, ln one scenario, sharing this
information in the caucus relieves anxiety for party A while the mediator
absorbs anxiety Party A feels better, the mediator is triangulated and becomes
more reactive, and the mediation is more likely to stall.
If the mediator refuses to be triangled, however, party A is left vwth managing his anxiety on his own. This detriangulation can take many forms. In
this instance, for example, this infonnation may be essential to the mediation
process. The mediator understands the interlocking triangle the wife's involvement creates. The mediator can say, "1 respect your wish to keep this information private. However, I do not believe that you will get a mediation agreement
acceptable to both panies unless this issue is resolved. How do you suggest we
proceed?" This I position keeps the anxiety where it belongswith the parties, and it frees the mediator to respond from a nonattached positiona position of strength rather than weakness.
As such, rather than deciding when or if to use caucusing as a mediator,
in using a Bowen Systems Theory perspective, the focus becomes how to

Bowen Systems Theory and Mediation

125

use caucusing as a way of promoting differentiation of the system. If that is not

possible because of the low differentiation level of one or both of the parties,
the highly reactive nature of the conflict, and so forth, the mediator should
not use caucusing. Viewed from this perspective, caucusing is not a technique at all but an approach to implementing Bowen Systems Theory in
mediation.
The second example is comediation. Many mediators struggle with
whether or not and in what circumstances a comediator should be used. As a
strategy and technique, comediation has many advantages and some disadvantages (CDR Associates, 1996; Moore, 1996), Advantages include having a
second mediator to help sort through the complexities of the case, having the
opportunity to lead and follow so as not to get too caught up with the parties,
and making the mediation more fun and productive by having a second person involved. Disadvantages include working with a mediator who views the
conflict so differently that it is difficult, to work together, having a partner who
takes over the mediation rather than shares it, having the parties skillfully
manipulate one mediator against the other, and the financial constraints of a
second mediator's involvement.
As an application of Bowen Systems Theory, the decision to use or not use
a comediator rests on different assumptions. The mediator must determine
whether the comediator has the capacity to remain detriangulated, whether
the comediator has the skill and confidence to lead and follow without undue
anxiety, and whether the comediator can tolerate the additional ambiguity that
working with another mediator brings. In short, using this model, the mediator must assess whether the comediator is sufficiently differentiated to be an
asset to the mediation or whether the mediator's relative lack of differentiation
will impede the mediation process.
Making this determination is sometimes difficult. For example, in the Justice Court where I work, mediators are often paired together from an available
list. This experience has given me the opportunity to work with a variety of
comediators. As a result, I now recognize with whom I am effective and with
whom I cannot mediate. Through the lens of Bowen Systems Theory, I can
observe the emotional field that each mediator brings to mediation. Comediation, then, becomes an opportunity to work on my ovwi differentiation in an
environment that also promotes differentiation in both the comediator and the
mediating parties.
Mediation is one form of nonviolent, altemative conflict resolution. It can
successfully assist individuals to take control of their ov^m lives and empower
them to make decisions rather than having those decisions be made for them.
Mediation can bring people together in ways that increase respect for differences. It can be a source of positive change through creative conflict resolution
in a world that too often relies on an adversarial approach to conflict. Mediation is also a relatively new academic field. Like all new disciplines in the

126

Regina

infancy stage, it is an exciting time that has encouraged creativity, innovation,

and experimentation. The field of mediation has been largely driven by techniques and strategies and has enjoyed success and a growing acceptance. This
article is an attempt to ground mediation in a comprehensive, systems theory
in order to increase its effectiveness and credibility. Of course, grounding mediation in Bowen Systems Theory or any other comprehensive theory will not
make every mediation successful. No theory can accurately predict human
behavior and attain unqualified success. Rather, by viewing mediation through
the theoretical lens of Bowen Systems Theory, I have demonstrated that mediation is a human endeavor governed by the same theoretical concepts and
principles as other human enterprises. As a theory of human behavior
and functioning, Bowen Systems Theory has much to offer the field of mediation. The successful application of Bowen Systems Theory to mediation is a
testament to the resiliency of the theory and it supports the belief that the academic discipline of mediation can only benefit from being grounded in a comprehensive theory.
References
Avruch, K. Culture and Conflict Resolution. Washington, D.C: United States Institute of Peace Press,
1998.
Bateson, G., Jackson, D. D., Haley, J., and Weakland, J. H. "Toward a Theory of Schizophrenia."
Behavioral Sdence, 1956,1, 251-264.
Bethel, C. "The Use of Separate Sessions in Family Mediation." Mediation Quarterly, 1986, 2 (3),
257-271.
Bowen, M. "Family Relationships in Schizophrenia." In A. Auerbach (ed.). Schizophrenia: An Integrated Approach. Somerset, N.J.: Ronald Press, 1959.
Bowen, M. "The Use of Family Therapy in Clinical Practice." Comprehensive Psychiatry, 1966, 7,
345-374.
Bowen, M. "Family Therapy and Family Group Therapy" In H. Kaplan and B. Sadock (eds.).
Comprehensive Group Psychotherapy. Baltimore: Williams and Wilkins, 1971.
Bowen, M. "Toward a Differentiation of Self in Administrative Systems." In L. McClenathan and
L. McClenathan (eds.), Georgetown Family Symposia, Volume I. Washington, D.C: Georgetown
University Medical Center, 1974.
Bowen, M. Family Therapy in Clinical Practice. Northvale, N.J.: Aronson, 1978.
Carter, E., and McGoldrick, M. "Overview: The Changing Family life Cycle^A Framework for
Family Therapy" In B. Caner and M. McGoldrick (eds.). The Changing Family Li/e C^cle;
A Framework for Family Therapy. (2nd ed.) Needham Heights, Mass.: Xilyn & Bacon, 1989.
CDR Associates. Mediating Environmental and Public Policy Disputes. Boulder, Colo.: CDR Associates, 1996.
Constantino, C. A., and Merchant, C S. Designing Conflict Management Systems. San Francisco:
Jossey-Bass, 1996.
Dicks, H. W Marital Tensions: Clinical Studies Toward a Psychological Theory of Interaction.
New York: Basic Books, 1967.
Dillow, R. "The School as an Emotional System." In P A. Comella and others (eds.). The Emotiona]
Side of Organizations. Washington, D.C: Georgetown Family Center, 1996.
Ferrera, S. "Lessons from Nature on Leadership." In P A. Comella, and others (eds.). The Emotional Side of Organizations. Washington, D.C: Georgetown Family Center, 1996.

Bowen Systems Theory and Mediation

127

Framo, J, L, "Rationale and Techniques of Intensive Family Therapy," In 1, Boszormenyi-Nagy

andJ. L, Framo (eds,), Intensive Family Therapy. New York: HarperCollins, 1965,
Framo, J, L, "A Personal Retrospective on the Family Therapy Field: Then and Now." Jounwl of
Marital and Family Therapy, 1996,22 (3), 289-316,
Frankl, V E. "From Psychotherapy to Logotherapy" Pastoral Psychology, 1956, 7, 56-60,
Friedman, E, H, Generation to Generation. New York: Guilford Press, 1985,
Friedman, E, H, "Bowen Theory and Therapy." In A. S. Gurman and D, P Kniskem (eds,). Handbook oJ Family Therapy, Volume 2, New York: Brunner/Mazel, 1991.
Friedman, E, H. Reinventing Leadership: Change in an Age qf Anxiety. New York: Guilford Press, 1996,
Garfield, S, L, "Psychology: A 40-Year Appraisal," American Psychologist, 1981, 36, 174-183.
Gilbert, R. "A Natural Systems View of Hierarchy" In P A. Comella and others (eds,). The Emotional Side of Organizations. Washington, D.C,: Georgetown Family Center, 1996.
Gray, B. Collaborating. San Francisco: Jossey-Bass, 1989.
Haley, J. Strategies of Psychotherapy. Philadelphia: Grune & Stratton, 1963,
Jackson, D. D. "Interactional Psychotherapy" In M. T. Stein (ed.). Contemporary Psychotherapies.
NewYork: Free Press, 1961.
Kerr, M. E. "Family Systems Theory and Therapy." In A. S. Gurman and D. P. Kniskem (eds.).
Handbook of Family Therapy. New York: Brunner/Mazel, 1981.
Kerr, M. E. "Bowen Theory and Evolutionary Theory." Family Systems, 1997, 4 (2), 119-179.
Kerr, M., and Bowen, M. E. Family Evaluation. NewYork: Norton, 1988.
Lax, D. A., and Sebenius, J. K. The Manager as Negotiator. NewYork: Free Press, 1986.
Maser, C. Resolving Environmental Conflict. Delray Beach, Fla.: St. Lucie Press, 1996.
Maslow, A. H. Motivation and Personality. New York: HarperCollins, 1954.
McCullough, P G. "Leadership from the Viewpoint of Bowen Theory" In P A. Comella and
others (eds.). The Emotional Side of Organizations. Washington, D . C : Georgetown Family
Center, 1996.
Moore, C. W "The Caucus: Private Meetings That Promote Settlement." Mediation Quarterly, 1987,
16, 87-101.
Moore, C. W The Mediation Process. (2nd ed.) San Francisco: Jossey-Bass, 1996.
Muraskin, M. R. "Nonreproductive Sexual Behavior in Wild Apes." Family SysteTns, 1994, i (2),
142-163.
Papero, D. V Bowen Family Systems Theory. Needham Heights, Mass.: Allyn & Bacon, 1990.
Regina, W "Bowen Systems Theory and the Nature of Organizational Leadership and Followership." Unpublished manuscript, 2000.
Regina, W, and LeBoy, S. "Incest Families: Integrating Theory and Practice." Family Dynamics of
Addiction Quarterly, 1991, i (3), 21-30.
Roberts, K., and Lundy, C. "The ADA and NLRA: Resolving Accommodation Disputes in Unionized Workplaces." Negotiation Journal, 1995, i i (1), 29-41.
Rogers, C. R. Client-Centered Therapy. Boston: Houghton Mifflin, 1951.
Rogers, C. R. On Becoming a Person: A Therapist's View of Psychotherapy. Boston: Houghton
Mifflin, 1961.
Schnarch, D. Passionate Marriage: Sex, Love, and Intimacy in Emotionally Committed Relationships.
New York: Norton, 1997.
Schwartz, R. T?ie S/eilled Facilitator San Francisco: Jossey-Bass, 1994.
Skinner, B, F The Behavior of Organisms. Englewood Cliffs, N.J.: Appleton-Century-Crofts,
1938.
Stamato, L., and Jaffe, S. "Mediation and Public Policy: Variations on a Consensus Theme." Mediation Quarterly, 1991, 9 (2), 165-178.
Suskind, L., Babbit, E., and Segal, P "When ADR Becomes the Law: A Review of Federal Practice." Negotiation Journal, 1993, 9 (1), 59-75.
Ury, W L., Brett, J. M., and Goldberg, S. B. Getting Disputes Resolved: Designing Systems to Cut the
Cost of Conflicts. San Francisco: Jossey-Bass, 1988.

128

Regina

whittaker, C, A, "Psychotherapy with Couples," American Journal of Psychotherapy, 1958, 12,

18-23,
Whittaker, C, A., and Malone, T. P. The Roots of Psychotherapy. New York: Blakiston, 1953,
Wylie, M. S, "Family Therapy's Neglected Prophet," Family Therapy Networker, 1991, 15 (2),
24-37, 77,
e, L, C, Ryckoff, 1, M,, Day, J,, and Hirsch, S, I, "Pseudomutuality in the Family Relations
of Schizophrenia," Psychiatry, 1958, 21, 203-220,

Wayne Regina is director ofSkyview School, a K-8 charter school; professor of peace studies
at Prescott College; and a licensed psychologist, licensed marriage and family therapist, and
certified mediator.