DIALOGICAL ENCOUNTERS: CONTEMPORARY

PERSPECTIVES ON “CHAIRWORK” IN PSYCHOTHERAPY

SCOTT KELLOGG

The Rockefeller University

This article looks at the use of

“chairwork” (2-chair and “empty”

chair) dialogues through the lens of 5

psychotherapies: Gestalt,

process-experiential therapy, redecision

they involve the replaying of difficult or

traumatic scenes from the past; or (d)

corrective, which means that the

emphasis is on replacing maladaptive

cognitions or schemas with ones that

are healthier. A potential foundation

for Gestalt and cognitive–behavioral

integration is proposed on the basis of

the idea that gestalts are schemas (I. G.

Fodor, 1996) and that chairwork is

outline contemporary visions of how chair dia-

logues can be used as seen through the lenses of

several psychotherapeutic schools. The patterns

and structures that emerge from this exploration

can help guide therapists in their use of this ap-

proach while laying the foundation for future

developments.

Chairwork, which originally developed as a

technique used in psychodrama (Carstenson,

1955; Fowler, 1992; Lippitt, 1958; Perls, 1973),

Gestalt therapy has spread throughout the planet

with Gestalt institutes in major cities (Greenberg

& Brownell, 1997), a recent Delphi study on the

future of psychotherapy (Norcross, Hedges, &

Prochaska, 2003) found that experts in the field

believed that the influence of Gestalt therapy

would continue to decline over the next 10 years.

In comments that were germane to this article, the

panel of experts believed that cognitive–

behavioral and integrative therapies would con-

through the lens of five psychotherapies: Gestalt

therapy (Baumgardner, 1975; Perls, 1969a, 1973,

Davis, & Slatick, 1998; Greenberg, 1979; Green-

berg, Rice, & Elliot, 1993; Greenberg, Watson, &

Goldman, 1998; Wolfe & Sigl, 1998), which is a

combination of Gestalt, experiential, and client-

centered therapies; redecision therapy (Goulding,

Psychotherapy: Theory, Research, Practice, Training Copyright 2004 by the Educational Publishing Foundation

2004, Vol. 41, No. 3, 310–320 0033-3204/04/$12.00 DOI 10.1037/0033-3204.41.3.310

1997), which represents a uniting of transactional

analysis and Gestalt therapy; cognitive–behav-

ioral therapy (Goldfried, 1988, 2003; Samoilov &

Goldfried, 2000); and schema therapy (Young,

Klosko, & Weishaar, 2003), which incorporates

aspects of cognitive–behavioral, psychodynamic,

and Gestalt/experiential therapies. A full descrip-

Each of the therapies has shared and unique

goal or outcome, and the role of the therapist.

These differences can be clearly seen in the ways

that chairwork is undertaken and understood. In a

Greenberg, Safran, and Rice (1989) explored the

differences between experiential and cognitive–

behavioral therapies. They described the experi-

ential therapies as facilitating; the goal was to

inner world.

An examination of the five therapies reveals a

further delineation of the modifying group. While

the Gestalt and process-experiential therapies are

facilitating, among the modifying approaches, re-

decision therapy is a conflict therapy, Goldfried’s

(1988) rational restructuring approach is a cor-

rective therapy, and schema therapy is both a

conflict and a corrective therapy. The conflict

therapies, typically through the use of chairs and

junction. This is conveyed through the words and

actions of this figure. The child, as a means of

survival, makes a decision to accept the patho-

logical injunction. This sets into play a dysfunc-

tional script or a lifelong pattern of problematic

behavior. Using chairs and imagery, the patient is

brought back to a scene that connects to the origi-

nal injunction and acceptance decision. The pa-

tient now confronts the parent or pathogenic fig-

ure or figures and tells them that he or she will no

longer accept the injunction and the patient will

early maladaptive schemas. Schemas are the-

matic structures “comprised of memories, emo-

tions, cognitions, and bodily sensations” (Young

et al., 2003, p. 7) that serve as a blueprint for the

child’s world. Again, while they may have had a

survival value for the child in a dysfunctional

situation, they typically impair later functioning.

They are also seen as being a core component of

the Axis II disorders as well as many Axis I dis-

orders—especially those that are recurring. Pa-

Schema therapy also uses chairs to dispute the

validity of the schemas, and, in this regard, it is

also a corrective therapy.

What these therapies share in common is the

belief that events from the past continue to play a

detrimental role in present-day functioning. Illus-

trating this, Tobin (1976) wrote the following:

External dialogues frequently consist of what

Greenberg, 1995) has referred to as “unfinished

business.” This typically occurs when an indi-

his father, it was no longer true for himself. He

then went on to restructure his life in such a way

(Goulding & Goulding, 1997).

In these dialogues, sometimes the parent figure

will change and support the patients’ new deci-

sion and sometime they will not. If not, Goulding

Gestalt community (i.e., From, 1984), who could more clearly

be defined as taking a facilitating approach.

make the decision in defiance of the parent. They

then ask the patient to find other supportive fig-

ures (such as other family members or teachers)

who will support the patient in the change pro-

cess. These individuals are then put in the chair,

and they express their support for what the pa-

tient is doing. Another variant is to have patients

become a new, affirming father or mother to

new father.... Be the kind of father you want

now....andtell Abe what you enjoy about him”

(p. 77). In this way, she was trying to create more

positive introjects. In the language of transac-

tional analysis, she was strengthening the nurtur-

ing parent at the expense of the critical parent;in

schema therapy language (Young et al., 2003),

she was developing the healthy parent and dimin-

ishing the power of the punitive parent.

Not infrequently, these scenarios involve trau-

time, however, they are given a weapon so that

they can defend themselves directly or, if that

feels too difficult, their adult selves or the thera-

pist steps in and defends the child self.

As noted above, Goulding and Goulding

(1997) have some clear goals behind their inter-

ventions. They want to “turn the scene from [a]

tragedy to a drama that ends well” (p. 168). It is

interesting to note that they emphasize the crucial

importance of the patient making a decision to

sexual and physical abuse. The structure that they

use takes this form:

1. The patient describes an abuse scene from

2. The patient and the therapist then discuss

3. An empty chair is then brought in for the

4. The next step is to have the child relive the

cause my fun didn’t cause you to rape me! It was

your perversity!” (Goulding & Goulding, 1997,

p. 248).

Not surprisingly, these scenes may need to be

revisited a number of times before this kind of

resolution can take place. Goulding and Goulding

(1997) emphasized that no matter what happened

or what the children did, all guilt lies with the

perpetrator. If the patient has difficulty with feel-

ings of guilt, the therapist will organize a two-

sages that are being given to them. They are en-

couraged to be self-affirming. Unlike in cases of

physical and sexual abuse, sometimes the figures

are put in the chair so that the patient can better

understand what drove them to behave that way

and how they may have been projecting their own

issues onto the patient.

Using combinations of imagery and chairwork,

Young et al. (2003) also has patients challenge

parental figures, other figures from their past, and

people in their current life situation. This is done

to help break the strength of the schema, which

A specific form of unfinished business is “say-

patient’s growth and prevent further develop-

ment. “The individual is still carrying around

frustrations, hurts, guilts, and even unexpressed

love and appreciation” (Tobin, 1976, p. 375). Not

only may people need to say goodbye to those

who have died or those whose relationships have

ended through such events as divorce or matura-

tion, they may also need to release their connec-

tion to people who they do not know, such as

one. Patients are then asked to invite the indi-

vidual into the empty chair. They are asked what

they are experiencing as they imagine that per-

those feelings to them. Patients then switch chairs

and respond from the perspective of the deceased

or missing person. Keeping with the Gestalt em-

phasis on balance, it is of great importance that

patients ultimately express both the resentments

and the appreciations that they have for this per-

not; if not, the reasons for not doing so are ex-

plored and respected. While saying goodbye may

ultimately be a better solution, Tobin will allow

patients to defer making this kind of resolution,

but he does want them to take responsibility for

making that decision. These scenarios may need

to be repeated over a number of sessions before

all of the issues can be worked through (Fodor,

1987).

Goulding and Goulding (1997) added to this

some specific procedures that they use when the

patient is saying goodbye to a deceased person.

to bring up the image of the person as dead and

Internal dialogues are seen as useful when pa-

tients experience conflicts between different parts

of themselves, when they are of “two minds,” or

when they are “at war with themselves.” The dis-

tinction between internal and external is not hard

and fast in actual practice because some of the

disturbing internal voices are actual introjects of

parental figures (Perls, 1973). Dialogues that be-

dresses the issue of inner conflict. They, as do

others in the Gestalt tradition (Fagan et al., 1976),

call these “splits.” Splits or conflicts often in-

volve issues of desire and criticism or of desire,

fear, and criticism.

In many cases, these kinds of situations in-

volve a harsh and critical voice (also known as

the “inner critic”). In Greenberg’s (R. Elliott &

Greenberg, 1997; R. Elliott et al., 1998; Green-

berg et al., 1993, 1998) model, one chair embod-

other things, depression and procrastination. In

the first series of chair dialogues, it became clear

that as the inner critic made its demands, she

retreated and avoided. In this way, her procrasti-

nation is a way of coping with these harsh inter-

nal voices. Later in treatment, she was able to

have dialogue between the critic and the creative

side of herself. Here it emerged that the critic was

actually frightened of the creative side; she was

afraid that it would be overwhelming. The cre-

would destroy her. When she was finally allowed

ative side out “in moderate doses” and end the

developed a three-way dialogue between his de-

veloping assertive self; a critical, repressive, and

moralistic voice that he connected with his father;

and a fearful, anxious voice that he connected

with his mother. The resolution in a situation like

this could involve expressing one’s desires and

creating plans to act on them while also clarifying

one’s moral code and being cognizant of realistic

dangers that might exist.

Inner critic issues relate to Perls’ (1973) dis-

value, importance, or use, and lets the rest go.

This is an integrative process, not an all-or-

nothing process. These dialogues enable this as-

similation process to occur.

A second type of internal conflict is what Perls

(1973) called a retroflection and what Greenberg

(1979; Greenberg et al., 1993) called a “self-

interruptive split.” In this situation, one part of

the person does something to another part of the

person. Adapted examples from Greenberg

in the other.

One may also be of more than “two minds”

about something. Young et al. (2003) developed

mode therapy, a variant of schema therapy, to

address the problems experienced by individuals

with severe character disturbance—particularly

borderline and narcissistic personality disorders.

They see the inner world of these patients as be-

ing populated by a number of inner figures in-

cluding the vulnerable child, the angry child, the

detached protector, the punitive parent, and the

healthy parent. While Young et al.’s model in-

Another kind of inner dilemma can be found

different projections about one’s future. Deci-

sions to stay in a relationship or leave, to take a

new job or stay in the current one, to allow a child

to take a year off after high school or insist that he

or she attend college immediately may not have a

clear right or wrong answer. Having each chair

represent a side of the argument and having the

patient speak from that perspective (“I want to

stay in my current job,”“I want to take the new

job”) can help him or her get a clearer sense of

nected to schemas or modes, and they frequently

gave names to the different perspectives or the

different selves that have emerged in the work.

This kind of approach is also applicable to pro-

crastination. Goulding and Goulding (1997) re-

ported a case in which a patient was procrastinat-

ing in the completion of her dissertation. As she

worked with “I want to write” and then “I won’t

write,” she realized that her anger at her parents

was playing a role in her lack of productivity. The

cussion on abuse above, the adult patient spoke to

himself when he was a child. In blocked decision-

making situations, it can be helpful to speak

about the decision from a future time perspective;

that is, patients can speak about how their life is

1, 5, 10, or more years from now, given that they

had made a specific decision. Patients can be

prompted to explore the impact of the decision on

specific areas of their lives. “You decided to take

that job and it is now 5 years later. How are you

marriage? How is your health? Your sexuality?

individual and various body parts or diseases. In

certainly appear to be an approach worth explor-

ing. Young et al. (2003), again using a combina-

tion of imagery and dialogue, described a case of

a physician who had been in therapy for 20 years

in an attempt to address his concern that he had a

“migrating tumor” (p. 83). The patient was asked

to imagine the tumor and then have a dialogue

with it. The tumor said that the patient “has not

been doing his best work and is very bad. The

tumor is in his body to punish him. Paul [the

fronted by his extremely demanding father. The

therapist concluded that “like the tumor, the fa-

schema” (Young et al., 2003, p. 83).

Cummings (1999) wrote about the value of

Greenberg’s (Greenberg et al., 1993) process-

experiential therapy in the treatment of patients

with genital herpes. First, she noted that

the self engendered by the disease: e.g., being a good versus

bad person, feeling out of control versus gaining self-control,

self-blame versus other blame. (p. 147)

This could include using the empty chair to speak

to the person who transmitted the disease to them.

It could also be used to practice telling a new

partner that they have herpes. In her case ex-

ample, Cummings encouraged the patient to put

her herpes in the other chair and gave her the

needed to say. At the beginning of treatment, the

structure is somewhat different. This use of chair-

work is centered in the cognitive–behavioral ap-

proaches and is also found in schema therapy. It

grows out of the disputation tradition and in-

volves the patient first expressing the dysfunc-

tional thought or schema in one chair and then

countering it in the other. In a sense, this is the

most directed use of chairs, in that the therapist

may purposefully work with patients to create a

dialogue that counters the dysfunctional one.

of life. The Elliotts believe that most psychopa-

thology and psychic anguish comes from the “in-

ner critic,” a punishing, judgmental inner voice

that seeks to control the individual. The core

technique in this work is the anthetic dialogue. In

a recent formulation of this approach (by Elliott

& Elliott), the patient describes his or her prob-

lem and the view of the inner critic is elicited.

This view is typically filled with “shoulds” that

the individual must follow or that lead to an ex-

manner based on his or her values and beliefs,

rather than out of fear.

Working from a cognitive–behavioral perspec-

tive, Goldfried (1988, 2003; Samoilov & Gold-

fried, 2000) made the case that both clinical prac-

tice and neuroscience are pointing to the impor-

tance of “hot” or emotionally laden cognitions in

the change process. He viewed the incorporation

of chairwork into cognitive–behavior therapy as a

way to more effectively change patients’ cogni-

tive structures. Patients are invited to engage in a

dialogue between the “realistic” and “unrealistic”

parts of themselves. They are also told that this is

a way of “taking what is internal and implicit and

making it external and explicit” (Goldfried, 1988,

p. 65). Again, the purpose is to enable the patient

to experience emotional arousal so that his or her

cognitive structure is more amenable to change.

Young et al. (2003) built on this tradition by

dence. Schema therapists will often encourage

enact both sides of the dialogue.

In a case example, a patient named Daniel was

presented. His background included alcoholism

in his father and sexual, physical, and emotional

abuse at the hands of his mother. His primary

schemas were Mistrust/Abuse and Defectiveness.

In short, he had doubts about his worth and he

was extremely mistrustful of others. He had a

goal of developing a long-term relationship with

a woman, but his schemas were interfering. In

schema side was encouraged to speak, and then

the healthy side took a turn. For example, the

patient, sitting in the schema chair, said, “Women

can’t be trusted, and they’re very unreasonable

and erratic, and it will be very difficult to figure

out just what to do. And I don’t think you can do

it.” He then responded in the healthy chair by

saying, “Women are people just like men are, and

they can be very reasonable, and they’re very

nice to be with” (p. 103). Young et al. empha-

important because the patient may first accept the

new perspective intellectually but not emotion-

ally; the goal is to have the patient eventually

accept it on an emotional level.

In a case with a similar structure, a patient

named Ivy was presented. She had a Self-

Sacrifice schema, which meant that she put the

needs of others before her own. This was done to

such an extreme that it was causing her to feel

angry and depressed. She was specifically angry

at her friend Adam because she felt that she lis-

tened to all of his problems while he did not show

needs were met as well. As part of her change

care of her mother. She took further steps to let

go of the schema by telling her mother, “It cost

me too much to take care of you. It cost me my

sense of self” (p. 148).

Again, this approach contrasts with the “un-

folding” perspective of the Gestalt or process-

experiential therapists. Young et al. (2003) be-

lieve that the more troubled the patient is, the less

available are the healthy schema and mode

voices; in a sense, that is a core aspect of their

Dreams

Perls (1969a, 1973, 1975) strongly believed in

the importance of working with dreams as a way

to transformation. In his writings and in the tran-

A central aspect of this, which is discussed

below, is the eliciting of polarities or opposites

from the image within the dream. This is quite

clear from an account by Miller (1992), who de-

scribed a dreamwork session by Perls in 1966:

needs by herself” (Zahm & Gold, 2002, p. 869).