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

2004, Vol. 41, No. 2, 180–189 0033-3204/04/$12.00 DOI 10.1037/0033-3204.41.2.180

GESTALT THERAPY: PAST, PRESENT, THEORY,

AND RESEARCH

LAURA E. WAGNER-MOORE
University of Massachusetts at Boston
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Orthodox gestalt therapy suffered a Classical Gestalt Theory and

This document is copyrighted by the American Psychological Association or one of its allied publishers.

rather unfortunate fate; gestalt theory Perls’s Eccentricities

has been poorly articulated, and gestalt Fritz Perls has been simultaneously praised for
his creative exuberance and criticized for aspects
techniques have received minimal of his style that simply defy the term scientifi-
empirical validation. These weaknesses cally derived. The “Perlsian” form of gestalt
are, in part, a consequence of F. therapy primarily embodies the history and per-
Perls’s biographical history, which led sonality of Perls himself, rather than a scientific,
to an integration of disparate structured, empirically derived or theoretically
consistent model of psychotherapy. Gestalt
theoretical models that were theory is an intellectually fascinating, philosoph-
exacerbated by F. Perls’s haphazard, ically complex set of diverse but poorly articu-
idiosyncratic personal style. However, lated and poorly substantiated beliefs.
recent empirical research suggests that Conversely, it is interesting that the actual
the 2-chair technique is superior to techniques used by gestalt therapists have been
other therapeutic interventions for clearly delineated and have received some em-
pirical validation for their effectiveness. This dis-
conflict splits, decisional conflict, junction between theory and research makes the
marital conflict, and unfinished study of gestalt theory and technique decidedly
business and that the 2-chair technique Byzantine and perplexing. This review and cri-
is as effective as Rogerian and tique presents the most coherent aspects of gestalt
cognitive–behavioral therapies. therapy that are based on principles of Gestalt
psychology and provides a review of recent em-
Although F. Perls’s techniques may pirical work on gestalt techniques.
have been generated largely from his Historical context dictated a nomadic life for
idiosyncratic personality Perls, with moves from Berlin (where he was
characteristics, these techniques have heavily influenced by psychoanalysis) to Johan-
some validity for very specific nesburg in 1933, to New York in 1946, and later
psychological dilemmas. to California. Geographic location seemed, in
part, to influence Perls’s concepts and practice of
psychotherapy, on the basis of his teacher “du
jour.” As a result, Perls’s approach to gestalt
Laura E. Wagner-Moore, Department of Psychology, Uni- theory and therapy was, at best, eclectic. He bor-
versity of Massachusetts at Boston. rowed some ideas from his analysts, like Reich
This article represents a modified paper written for a gradu- and Horney (Miller, 1974). As Miller (1974)
ate school requirement at the University of Massachusetts at
noted, Perls integrated Horney’s notion that “neu-
Boston. Assistance was provided by Joan Liem, Ester Sha-
piro, and Don Kalick, University of Massachusetts at Boston.
rotic behavior is based on manipulation, designed
Correspondence regarding this article should be addressed to win love” (p. 5-24) and was attracted to “Sar-
to Laura E. Wagner-Moore, PhD, who is now at the Center for trean” existentialism’s idea of individual respon-
Child and Family Health, Duke University Medical Center, sibility and choice (Miller, 1974). In the wake of
3518 Westgate Drive, Suite 100, Durham, NC 27707. E-mail: the zeitgeist of World War II, which was so heav-
[email protected] ily permeated by phenomenological and existen-

180
 Gestalt Therapy

tial thought, Perls rejected much of his analytic 1997). This process is best encapsulated in the
training in favor of the new zeitgeist (Yontef & notion of figure–background gestalt formation
Simkin, 1989). From Jan Smuts, the prime min- and destruction. Gestalt psychology suggests that
ister of South Africa, Perls borrowed the idea of a mass of unstructured individual data in the en-
holism. Ironically, he initially had minimal expo- vironment (i.e., parts) are subjectively structured
sure to or understanding of Gestalt psychology by the perceiver into wholes that have both form
itself and has been criticized for this on several and structure (Perls, Hefferline, & Goodman,
accounts (Wheeler, 1991). Wheeler (1991) noted 1951). The person’s actual experience is deter-
that Perls’s early work, titled Ego, Hunger and mined by the gestalt, rather than the raw pieces of
Aggression, was simply revised and renamed The data. The way in which multiple data are shaped
Beginnings of Gestalt Therapy (renamed at his is based on the individual’s needs, appetites, and
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

wife’s suggestion) with little description of what impulses.

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Perls meant by Gestalt psychology or therapy. These concepts can be applied to the realm of
Perls himself called the book “sketchy,” and psychological needs as well. In theory, a need
Wheeler convincingly argued that Perls’s original arises and becomes foreground; if it is satisfied, it
text is full of “vague philosophical musings and becomes background as the gestalt is completed.
self-aggrandizement in the Freudian manner” Pathology arises when this process is disrupted.
(Wheeler, 1991, p. 43). When gestalt formation is
The fate of classical gestalt therapy is a sad blocked or rigidified at any stage, when needs are not recog-
one, which finds itself without a clear set of bind- nized or expressed, the flexible harmony and flow of the
ing theoretical principles and without a prolific organism/environment field is disturbed [and] unmet needs
body of literature dedicated to the critical analy- form incomplete gestalten that clamor for attention . . . and
sis and dialogue that could further its develop- interfere with the formation of new gestalten. (Yontef, 1969,
as cited in Simkin, 1976, pp. 223–234)
ment. Miller, a practicing gestalt therapist and
student of Perls, argued that gestalt has “slipped A need may be blocked by an unclear sensation
into a middle-aged decline” and is plagued by or a lack of awareness of one’s needs (Greenberg
“a persistent intellectual thinness” (Miller, 1974, & Rice, 1997). Therapeutic work focuses on in-
p. 21). Perls’s striking “anti-intellectual bias” creasing awareness to bring about change, so that
(Miller, 1974) has been transmitted over time. the emerging need may be identified, satisfied,
His famous “Lose your mind and come to your and enabled to retreat into the background
senses” best embodies his preference for acquir- (Simkin, 1976). Awareness of one’s experience
ing knowledge through experiencing and feeling and needs is considered the “royal road to the
rather than through empirically validated or ra- cure” (Greenberg & Rice, 1997).
tional, logical thought processes. Perls’s similar
preference for the idiographic over the nomothet- Experience/Contact Cycle
ic makes the systematic study of gestalt quite The processes underlying gestalt formation
challenging. Despite these difficulties, neo- and destruction were obtusely described by Perls
gestaltists have successfully described Perlsian as the experience/contact, or metabolism, cycle.
notions of gestalt therapy linked to classical ge- The cycle consists of four main phases, including
stalt psychological theory. The two concepts ex- awareness, excitement, action, and contact. In
plained most completely involve (a) figure– gestalt theory, the term contact does not equate
background gestalt formation and destruction and with the popular definition meaning closeness.
(b) the contact/experience cycle, as they related Gestalt theory uses the word contact as an ab-
to the etiology of psychopathology in gestalt stract, formal concept that refers to the exchange
theory. between an individual person and the surround-
ing environment (Miller, 1994). If the boundary
Figure–Background Gestalt Formation between the self and the environment (or other)
Perls, in keeping with other humanistic ap- becomes unclear or lost, then there is a distur-
proaches, believed in the self-actualizing poten- bance of contact and awareness (Yontef &
tial of the individual, which assumes that an or- Simkin, 1989). When the cycle is functioning
ganism ultimately knows what is best for its self- smoothly, awareness of internal or external
regulation and actualization (Greenberg & Rice, stimuli leads to excitement, which potentiates an

181
 Wagner-Moore

action tendency; the action tendency leads to ure, against a background, rather than focusing
need satisfaction (optimally) and contact (Green- on the entire field. Second, the model is criticized
berg & Rice, 1997). Dysfunction is considered for assuming cycle disturbances can always be
the interruption of the cycle at any stage (Green- traced back to a problem in awareness itself. The
berg & Rice, 1997), and resistances to contact model suggests that if an individual has aware-
account for these disruptions. ness of a goal and attempts to act on that impulse
Perls retained the traditional psychoanalytic but ultimately fails in that action, the failure is
notion of defenses to explain additional disrup- due to misunderstanding the need or not empow-
tions in the contact cycle including retroflection, ering the need from the very beginning. Wheeler
introjection, projection, and deflection. For in- (1991) has noted that failing to meet a goal may
stance, a break in the cycle between excitement not solely be attributed to a problem with aware-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

energy and action could be accounted for by ret- ness; instead, the individual may have misper-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

roflection, which occurs when a split within the ceived the original problem or the individual may
self and resistance of aspects of the self take have had clear awareness but difficulty following
place, wherein the self either (a) does to oneself through with actions that would lead to a success-
what one wants to do to someone else or (b) does ful outcome.
for the self what one wants someone else to do For instance, survivors of posttraumatic stress
for the self (e.g., being overly self-sufficient, re- may have developed a relatively clear awareness
sulting in isolation). In both cases, the self has a of individual needs and desires while simulta-
need with energy, but the energy is diverted away neously having cognitive distortions or inaccu-
from its natural object and turned back against rate (or accurate) perceptions of environmental
itself (Wheeler, 1991). Increasing awareness is a threat that impede recovery; in other words, con-
primary psychotherapeutic tool because the cycle textual factors are quite relevant in decreasing
can be interrupted at the first stage if a need is personal distress. Perls’s conceptualization
blocked by dull sensation or poor awareness makes the individual unduly responsible for
(Greenberg & Rice, 1997). meeting his or her own needs and simultaneously
Implicit in the model, however, is the assump- fails to account for peculiarities in the environ-
tion that all disturbances in the cycle can be ul- ment that counteract or conflict with the individu-
timately traced back to a problem with awareness al’s need.
itself (Wheeler, 1991; Yontef & Simkin, 1989). It Perls’s quirky beliefs, peppered throughout ge-
follows, then, that gestalt psychotherapy uses stalt theory, may help uncover the mystery be-
here-and-now (present-centered) experiments in hind several significant sources of weakness and
directed awareness to increase the individual’s incongruence in the match between gestalt theory
awareness as well as the individual’s awareness and therapeutic practice. There are multiple dan-
of the process of awareness (Yontef & Simkin, gers inherent in considering individual impulses
1989). Contrary to traditional psychoanalytic and needs of primary importance. Miller argued
methods of excavating the past, gestalt therapy that Perls’s “aggressive self-expression borders
focuses on awareness and contact in the present on what Sartre characterized as ‘that diligent and
moment, using methods that ultimately serve to almost sadistic violence I call the full employ-
clarify present experiencing. ment of oneself ” ’ (Miller, 1974, p. 19). One of
Perls’s experience-cycle model has been criti- Perls’s great strengths, however, was his focus on
cized by philosophers of Gestalt psychology for individual potential, although he failed to realize
its insistence on focusing on individual impulses how destructive this position could be in rela-
or desires in isolation; significant contextual is- tional contexts.
sues are either minimized or ignored. This is a Perls believed that freeing oneself from com-
prime example of how gestalt theory often em- mitment and dependence on others was essential
bodies a particular penchant of Perls’s personal- (Perls, 1968). Perls embodied this ideal when,
ity structure rather than a predictable concept after becoming established in New York, he left
driven by theoretical constructs. his wife and children and drifted westward to
The cycle, criticized by Goldstein (as cited in pursue his own goals (Miller, 1974). Impulses
Wheeler, 1991), has been labeled “figure-bound.” and drives that are not moderated by reason, re-
This suggests that the model rests largely on the straint, and consideration of the “other” can end
immediate need or impulse that has become fig- up in frenzied, relatively autistic, and chaotic re-

182
 Gestalt Therapy

lationships that lack reciprocity and empathy. tic style led to gimmicky techniques that flour-
Perls’s focus on separateness and self-reliance ished. He was known to “sprinkle his audiences
are reflected in the first few lines of his mantra and trainees with slogans” as he made up new
from Gestalt Therapy Verbatim: techniques “on the fly,” which he presented as
“the latest essence of Gestalt therapy” (Miller,
I do my thing, and you do your thing. I am not in this world
to live up to your expectations. And you are not in this world 1994). Theory-driven, empirically validated tech-
to live up to mine. You are you and I am I, and if by chance nique was progressively abandoned and subse-
we find each other it is beautiful. If not, it can’t be helped. quently replaced by the methods generated by
(Perls, 1968, p. 4) Perls’s own dramatic, off-the-cuff flare.
Unfortunately, although classical gestalt therapy Modern-day gestalt therapy has retained
is more than capable of promoting self-reliance Perls’s applied phenomenological approach and
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

and the drive toward individuation, it fails to ac- creative techniques. Contrary to Perls’s style,
This document is copyrighted by the American Psychological Association or one of its allied publishers.

knowledge the benefits inherent in relational and modern gestaltists consider the relationship be-
intersubjective approaches to psychotherapy. The tween the therapist and client one of the most
similarities between the greatest weakness in ge- important aspects of psychotherapy and use less
stalt theory and Perls’s own style are striking and stereotypic techniques (Yontef & Simkin, 1989).
provide a way of understanding the peculiarities Yontef argued that traditional gestalt techniques
of gestalt theory. stressing skillful frustration, client manipulation,
and self-sufficiency served to provoke shameful
The Fate of Classical Gestalt Therapy and reactions in patients (Yontef & Simkin, 1989).
the Rise of Modern Gestalt Therapy Yontef and others argued that modern gestalt
therapy is less harsh (Aleksandrov, 1997; Yontef
Almost 30 years have passed since Perls’s & Simkin, 1989) and has turned its focus to the
death, and gestalt therapy has certainly changed, genuine contact between patient and therapist
moving from the original Perlsian emphasis on (Greenberg & Rice, 1997; Yontef & Simkin,
skillful frustration and self-reliance to a gentler, 1989). Although there are over 60 gestalt therapy
“Rogerian-ized” version of gestalt therapy. The institutes throughout the world, no national orga-
1960s version of gestalt included psychodrama nization or standards have been established
techniques that contained a philosophy that em- as criteria for empirically validated gestalt
bodied existential principles of freedom and re- treatments.
sponsibility, analytic notions of defenses, and ge- Modern gestalt has changed in several respects
stalt psychological principals of gestalt formation but still embodies the majority of Perls’s original
and destruction. Modern gestalt has retained ideas and therapeutic techniques. Despite Perls’s
many of Perls’s original ideas but has also soft- lack of rigorous intellectual explanation of gestalt
ened in many respects. theory, his unique and creative style has given
As Perls’s practice of gestalt therapy pro- modern gestalt therapy several fascinating and
gressed, he wrote less about theory; his seminal effective therapeutic techniques. Even though
work, Gestalt Therapy: Excitement and Growth Perls attempted to connect the puzzle between
in the Human Personality (Perls, Hefferline, & gestalt theory and his own technique, empirical
Goodman, 1951), was considered one of the only validation of Perls’s techniques remained largely
comprehensive texts written about gestalt theory. unsupported until recently.
This work, however, is a rather arcane and un-
successful attempt to illuminate gestalt prin- Gestalt Psychotherapy Techniques
ciples. Jerry Kogan, a well-respected and admir-
ing student of Perls said that he began reading Despite the loose connection between gestalt
Gestalt Therapy and “thought it was terrible,” theory and practice, the techniques of gestalt
although he found Perls himself “a model of a therapy and Perls’s application of them are cre-
brilliant teacher and therapist” (Kogan, 1976, ative and artistic and embody an unusual charis-
p. 255). Perls rarely referred to this text after its matic and authentic approach to treatment. Re-
original printing and instead preferred to print cent research by Leslie Greenberg on the two-
transcripts of his work (therapy seminars) rather chair dialogue and the empty-chair dialogue for
than expositions explaining it (Perls, 1968). As conflict splits and unfinished business has helped
Perls’s focus on theory diminished, his narcissis- explain the two-chair method and has brought a

183
 Wagner-Moore

new understanding to the effectiveness of Perls’s aspects of the self ” (Clarke & Greenberg, 1988,
work. p. 5-19). Greenberg has identified three types of
Gestalt therapists often create experiments that splits: conflict split, subject–object split, and at-
help clients increase awareness by uncovering as- tribution splits (Greenberg, Elliot, & Lietaer,
pects of their experience; these therapists may 1994). An example of a conflict split occurs when
share hunches about what is occurring or may an individual wishes for a desired goal, such as to
teach clients ways in which they are interrupting be married, but simultaneously feels that he or
or avoiding their own experience (Greenberg & she should remain single to preserve his or her
Rice, 1997). A core belief is that clients will more independence (Greenberg, Rice, & Elliott, 1993).
fully understand their own emotions and needs In this split, there are two “I’s” that oppose each
through a process of discovery, rather than other, resulting in a sense of struggle. Greenberg
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

through insight or interpretation. In many in- has found that this type of split usually involves
This document is copyrighted by the American Psychological Association or one of its allied publishers.

stances, the client may discover a conflict be- a conflict between an individual’s principles and
tween aspects of experience or conflicts within fundamental emotional needs and wants (Green-
the self (Greenberg & Rice, 1997). The confron- berg et al., 1993). During two-chair work, the
tation between these conflicting aspects of expe- part of the self that primarily embodies needs,
rience can be facilitated by techniques such as the wants, and gut-level emotions is called the “ex-
two-chair or empty-chair dialogue. periencing self ”; the part of the self that embod-
ies either “shoulds” (e.g., “I should be able to be
Two-Chair Dialogue: Theory and happy”), negative evaluations of the self (e.g.,
Empirical Research “I’m just worthless”), or societal standards or val-
ues is called the internal critic (Clarke & Green-
The two-chair technique is one of the most
berg, 1988). The goal for two-chair work is to
powerful and widely used of the gestalt tech-
bring the experiencing self and the internal critic
niques. Orthodox gestalt therapists may feel frus-
into contact with each other, for the client to at-
trated with traditional gestalt therapy being bro-
tend to both sides, for “covert internal dialogue to
ken down into parts (e.g., using gestalt techniques
be made overt,” and for change to result as the
without being guided by gestalt theory). Yet, one
client increases self-acceptance and develops
of the greatest advances in both theory and re-
new cognitive schemas (Greenberg et al., 1993,
search for gestalt therapy has come from the
p. 191).
work of Leslie Greenberg and colleagues. Their
Although the two-chair technique can look
systematic presentation of the effectiveness of
relatively simple when applied by an expert, the
two-chair work and their rigorous empirical
task actually requires extensive skill, including
analyses of two-chair work has significantly en-
detailed knowledge of potential techniques, sen-
hanced the confidence in and applicability of ge-
sitivity to nonverbal cues, and the ability to deal
stalt therapy.
with “resistances” and to track process (Fagan et
Gestalt’s unfortunate fate of receiving minimal
al., 1974). This technique requires that the thera-
theoretical and empirical validation has begun a
pist and client make an agreement to work on a
critical change. Greenberg and colleagues have
split (Fagan et al., 1974) and that the client de-
both modified and clarified the arcane orthodox
velops both sides of the conflict through dialogue
gestalt ideas in their “process-experiential ap-
(Greenberg et al., 1993). Greenberg noted that
proach,” which combines Rogerian humanism
many clients begin with the “harsh, self-critical,
with Perlsian techniques. Greenberg’s marriage
blaming” part of themselves (Greenberg et al.,
of the two orientations has resulted in the birth of
1993). Then, the two-chairs are used to further
a well-defined and well-researched integrated
distinguish and separate out the conflicting as-
therapeutic orientation.
pects of the self (Greenberg et al., 1993).
Research has shown the two chairs can be de-
Two-Chair Work for Conflict Splits
scribed as the experiencing chair and the other
Greenberg’s formal analyses of gestalt thera- chair and that the metaphorical individuals pres-
pists suggest that two-chair work is most often ent in the two chairs undergo different transfor-
used when the client expresses a split (Clarke & mations during therapy. During the work, the ex-
Greenberg, 1988; Elliott & Greenberg, 1995). A periencing chair deepens in depth of experiencing
split is “a division of the self process into partial and inner exploration and uses an expressive

184
 Gestalt Therapy

voice (Clarke & Greenberg, 1988), while the fective states, and the constant use of caution and
other chair is “filled with the person’s ‘shoulds,’ due care in the application of these techniques
negative self statements and attributions” and cannot be emphasized enough. This work calls
uses an externalizing and lecturing voice (Clarke for exceptional fortitude on the part of the thera-
& Greenberg, 1988). pist and may contribute to secondary traumatiza-
A primary goal for the therapist during two- tion, confusing countertransferential responses,
chair work is to help the client keep the partial and unexpected transference–countertransference
aspects of the self separated, which can aid in dynamics for neophyte therapists or therapists
conflict resolution and integration (Clarke & who have poor psychological boundaries. In ad-
Greenberg, 1988). The therapist’s direction can dition, gestalt techniques may be contraindicated
improve the client’s attention to inner processes for patients with organic conditions, severe cog-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

and may help raise the client’s awareness of what nitive disorders (in which loosening emotional
This document is copyrighted by the American Psychological Association or one of its allied publishers.

he or she is experiencing in the moment. For expression results in chaotic rather than struc-
example, the therapist may ask the client to focus tured thought processes), impulse control
on a particular nonverbal behavior that is inter- difficulty, severe personality disorders, socio-
rupting the experience or may ask the client to paths, and psychotic patients (Saltzman, 1989)
exaggerate statements or voice inflection to in- or for those who need crisis intervention (Elliott
tensify affective experiencing (Daldrup, Beutler, & Greenberg, 1995). The techniques obviously
Engle, & Greenberg, 1988). In summary, the have limited applicability and are vulnerable
principles that guide the therapist’s work include to iatrogenic harm if not used carefully. Despite
(a) clearly separating out partial aspects of the the dangers involved in using these techniques,
client’s self, (b) encouraging client attention to recent empirical work suggests that they prove
and heightening of emotional experiencing and to be quite useful in facilitating deeper
expression, and (c) assessing the two-chair work emotional experiences and in reducing various
and cognitive and emotional changes with the symptomatology.
client (Daldrup et al., 1988).
Fagan et al. (1974) cautioned unskilled, un- Two-Chair Dialogue: Empirical Findings
trained therapists about the use of this technique.
Neither the therapist nor the client knows what The two-chair technique has been compared
may unfold during the work, and it often includes with several other therapeutic orientations and
the expression of deeply felt, painful emotions has received some empirical validation for the
(Fagan et al., 1974). Fagan et al. (1974) wisely reduction of conflict splits, indecision, marital
suggested that therapists should (a) not use the conflict, and interpersonal difficulty. Analogue
technique unless they have had personal experi- and experimental studies conducted thus far have
ence with the technique, (b) be ready for “explo- compared two-chair work with client-centered,
sions or strong emotional responses,” and (c) cognitive–behavioral (e.g., problem solving), and
know their patients well enough to know how to experiential (e.g., focusing) methods. These stud-
provide follow-up support; not resolving an in- ies did not compare the effectiveness of the ge-
tense conflict can be damaging for fragile pa- stalt approach versus another approach but in-
tients. These cautions are rarely cited in the cur- stead investigated the specific efficacy of the
rent literature but are absolutely essential points two-chair approach versus other methods for a
that need to be made. specific problem (e.g., conflict split or decisional
As a trauma therapist, I caution the use of these conflict). (For an overall comparison of gestalt
techniques with trauma survivors who can have therapy to other orientations, see Greenberg et al.,
primitive, visceral emotional eruptions that can 1994, which provides a meta-analytic review of
lead to regression, retraumatization, or dissocia- 37 studies.)
tive episodes. It is not always the initial accessing In the 1980s, Greenberg compared two-chair
of repressed emotion that is dangerous but rather dialogue with client-centered empathic respond-
the therapist’s encouragement to heighten and in- ing for resolving a conflict split in college stu-
tensify the felt emotion and the expression of that dents (N ⳱ 16; Clarke & Greenberg, 1988;
emotion that may lead to volatile situations. The Greenberg & Rice, 1981). Empathic reflections
use of clinical judgment, preparation of the pa- were used as a comparison because theory sug-
tient, titration of patient exposure to intense af- gests that empathic responses also increase client

185
 Wagner-Moore

experiencing (Clarke & Greenberg, 1988). responding plus focusing techniques. Greenberg
Greenberg’s analogue study showed the two- and Clarke also reasoned that the use of another
chair technique led to greater depth of experienc- directive, engaging technique would serve to con-
ing and greater change of awareness than the em- trol therapists’ expectancy effects (Clarke &
pathic techniques (Greenberg et al., 1994). These Greenberg, 1988). When the two-chair method
studies were not well controlled and were done was compared with focusing plus empathy, re-
with college students with relatively benign prob- sults showed that the two-chair technique, ap-
lems; the results have limited external validity plied to a split, produced significantly greater
and may not generalize to severely disturbed psy- depth of experiencing than focusing plus empa-
chiatric populations. thy (Clarke & Greenberg, 1988). The two-chair
The results of this study were replicated and and the empathy plus focusing resulted in greater
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

expanded by Greenberg and Dompierre (1981, as symptom reduction and shifts in awareness in the
This document is copyrighted by the American Psychological Association or one of its allied publishers.

cited in Greenberg et al., 1994; N ⳱ 16, subjects experimental subjects versus the no-treatment
used as own controls). This study used an outpa- controls (Clarke & Greenberg, 1988). These early
tient population and implemented two experi- studies, comparing gestalt to empathic modali-
mental treatment sessions including either two- ties, are largely analogue, have relatively small
chair or empathic responding (each patient re- sample sizes, and are most appropriately consid-
ceived either empathic responding or gestalt first ered initial pilot studies in the field, given that
and then received the complementary technique they have methodological flaws.
second). Participants in the two-chair treatment Subsequent experimental research has con-
condition reported greater depth of experiencing trasted the two-chair approach with cognitive–
and more shifts in awareness; two-chair partici- behavioral modalities (Clarke & Greenberg,
pants also reported greater conflict resolution im- 1986; Johnson & Greenberg, 1985). The results
mediately after the treatment and at the 1-week of these two well-controlled, methodologically
follow-up (Greenberg et al., 1994). In terms of improved studies indicate that the two-chair tech-
behavioral change, the gestalt group reported nique is useful for reducing indecision and im-
greater change and goal attainment than the com- proves intimacy and marital adjustment while re-
parison group; participants reported no difference ducing symptomatology. Both the two-chair and
in discomfort in the two groups (Clarke & Green- cognitive–behavioral methods appear to be supe-
berg, 1988). The replication of previous findings rior to wait-list control groups for reducing inde-
to an outpatient population slightly improves the cision and improving marital conflict (Clarke &
external validity of the results. However, the re- Greenberg, 1986; Johnson & Greenberg, 1985).
sults of the latter study are suspect because the The primary goal of Clarke and Greenberg’s
participants were receiving additional treatments (1986) study was to determine whether the gestalt
(at least eight prior sessions) that were not part of method, which focuses on affective experience,
the experimental design. There is no way to know was as effective as cognitive–behavioral ap-
if the subjects in the gestalt treatment group were proaches, which focus less on affect and more on
significantly different from the other treatment cognitive processes. Forty-eight subjects were
group from the beginning; the use of statistical pretested, posttested, and randomly assigned to
covariates (e.g., pretest) were not used to resolve three groups:
this uncertainty.
At least three studies consistently found that 1. cognitive–behavioral problem-solving
the two-chair technique was superior to empathic group
responding for increasing depth of experience 2. two-chair dialogue group
and shifts in awareness (Greenberg & Clarke,
1979; Greenberg & Rice, 1981; and Greenberg & 3. no-treatment, wait-list control group.
Dompierre, 1981, all as cited in Clarke & Green- The dependent measures were scales measur-
berg, 1988). Greenberg and his colleagues began ing degree of undecidedness (modified Scale of
to hypothesize that the more active component of Vocational Indecision; Osipow, Carney, &
the gestalt work was largely responsible for the Barak, 1976, as cited in Clarke & Greenberg,
group differences. Greenberg and Higgins (1980, 1986) and a scale measuring changes in subjects’
as cited in Clarke & Greenberg, 1988) decided to decision-making stage (modified Assessment of
compare the two-chair technique with empathic Career Decision Making; Harren, 1979, as cited

186
 Gestalt Therapy

in Clarke & Greenberg, 1986). The target prob- lieved that the problem-solving approach could
lem for all participants was kept constant and be most useful after the client had explored the
involved an emotionally meaningful “interper- affect states underlying his or her indecision; this
sonal conflict related to a [career] decision” remains an empirical question.
(Clarke & Greenberg, 1986, p. 12). The therapists A study with an identical goal was conducted
responsible for the treatment conditions were by Johnson and Greenberg (1985), comparing the
trained, doctoral-level psychologists; the thera- effectiveness of problem-solving skills (consid-
pists were trained only in the treatment they ad- ered a cognitive–behavioral approach) and expe-
ministered (they were also matched to the treat- riential treatment interventions (considered a
ment they typically practiced) and followed clear combination of the gestalt two-chair and Rog-
treatment protocols (Clarke & Greenberg, 1986). erian empathy techniques) in resolving marital
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

In addition, blind raters reviewed audiotapes of conflict. Forty-five couples were randomly as-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

the sessions to check for fidelity of treatment signed to either a wait-list control group or one of
implementation; all raters agreed on the delivery the two treatments. Six trained therapists, with an
of the specified treatment (Clarke & Greenberg, average of 4 years of experience, conducted eight
1986). An advantage of this study was its delib- 1-hour experimental treatment sessions. Treat-
erate attempt to guard against treatment biases ment validity was ensured with clear protocols
and to ensure the effectiveness of the treatment and frequent treatment monitoring and rating
manipulation. (Johnson & Greenberg, 1985). Dependent mea-
Results of the study showed the two-chair sures included tests of emotional styles (used to
technique was more effective than problem solv- check for group equivalence), couples’ alliance in
ing or no treatment, for reducing indecision therapy (e.g., measure of bond between therapist
(Clarke & Greenberg, 1986). This finding is and client, agreement with therapy tasks), marital
unique in that it shows a main effect for treatment adjustment (includes affectionate expression, co-
on a direct measure of indecision, rather than on hesion, satisfaction, etc.), target complaints (sub-
measures that indirectly correlate with indecision jects rate amount of change on the presenting
(Clarke & Greenberg, 1986). Similarly, the re- problem), degree of goal attainment (client re-
sults are impressive in that the problem-solving ports how well previously specified goals were
approach is specifically designed for ameliorat- met, e.g., worse than expected, less than ex-
ing indecision by having the client focus on the pected, etc.), and relational intimacy (e.g., emo-
problem to be solved, rather than attending to tional, social, sexual, intellectual intimacy).
underlying emotions (Clarke & Greenberg, Results indicated that the strength of the work-
1986). The problem-solving approach was indeed ing alliance between the therapists and couples
better than the control group, yet the gestalt two- and therapists’ effectiveness were equivalent for
chair intervention resulted in overall greater the two treatment groups (Johnson & Greenberg,
decisiveness. 1985). The two treatment groups scored signifi-
This is an impressive empirical study, with in- cantly better than controls on measures of goal
teresting results and only minor methodological attainment, marital adjustment, intimacy levels,
flaws. The results have been replicated in at least and target-complaint reduction (Johnson &
one other study (Clarke & Greenberg, 1988). The Greenberg, 1985). However, the emotion-focused
study lacks generalizability, owing to sample treatment was superior to the cognitive–
characteristics; out of 48 subjects, 37 were female behavioral treatment on measures of marital ad-
and only 11 were male. Given gender differences justment, intimacy, and target-complaint level;
in emotional expression, it is possible that two- these differences remained at the 8-week follow-
chair work is more effective for women than up (Johnson & Greenberg, 1985). These results
men. Perhaps the greatest caveat in this work is suggest that emotion-focused work can be a
that researchers do not know if increases in de- means of positively transforming dyadic relation-
cisiveness correlate with decision implementa- ships (Johnson & Greenberg, 1985).
tion or behavioral change (Clarke & Greenberg, This study’s strengths include random assign-
1986). People may self-report greater decisive- ment to treatments, checks on treatment imple-
ness but may still be lacking essential communi- mentation, and consistent level of therapeutic al-
cation and negotiation skills necessary for suc- liance across groups. Combined, these factors
cessful change. Clarke and Greenberg (1986) be- suggest that the results are due to the treatment

187
 Wagner-Moore

manipulation rather than spurious variables. Re- (1983) and colleagues found similar results in an
search on the two-chair technique suggests that it analogue study; they compared the effectiveness
has promise but that additional research is needed of rational emotive therapy and the empty-chair
to clarify the applicability of the technique. It is technique for anger reduction in college women
important to consider that the technique has been (measured with both self-report and physiologi-
empirically supported largely for ameliorating cal examination). They found both treatment
conflict splits. These are useful data that help techniques reduced systolic blood pressure and
support the theory behind the techniques. In an- self-report of anger as compared with controls
other sense, researchers know that two-chair (Conoley, Conoley, McConnell, & Kimzey,
work is specifically designed for conflict splits 1983). The rational emotive group had a trend
but do not know if it has greater applicability. toward being more effective than the empty-chair
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Clarke and Greenberg’s (1986) study comparing group on the self-report measure (Conoley et al.,
This document is copyrighted by the American Psychological Association or one of its allied publishers.

gestalt to a cognitive–behavioral approach for de- 1983). However, research by Paivio and Green-
cisional conflict represents a step in this direc- berg (1995) demonstrated that the empty chair
tion, because they compared the efficacy of the dialogue led to greater resolution of unfinished
two approaches for a problem typically targeted business as compared with a psychoeducational
by, and designed for, cognitive–behavioral tech- group.
niques. Future research should pursue the ques- At this point, the data on the empty-chair tech-
tion of how applicable two-chair work can be for nique for anger reduction and unfinished business
additional clinical problems. are inconclusive. Although some studies suggest
The most recent empirical research has inves- that two-chair work is superior to other tech-
tigated the gestalt empty-chair technique for “un- niques, other studies suggest that there is no dif-
finished business.” This technique facilitates the ference between two-chair work and other tech-
expression of unresolved feelings toward a sig- niques. The replication of existing studies and a
nificant other. The client is encouraged to direct greater volume of research in the field will help
unresolved emotions, such as anger, toward an clarify the efficacy of these techniques.
appropriate target; for instance, this could involve
having the client redirect internalized anger to- Conclusion: Current Status of
ward the person with whom he or she is angry. Gestalt Therapy
Through this process, the client is able to express
previously interrupted feelings more fully and is The era of Perls’s clinical artistry has passed,
able to identify suppressed needs (Greenberg et and gestalt’s techniques have been demystified
al., 1994). Resolution is achieved when the client by recent empirical investigations. Despite the al-
has intensified emotion, expressed his or her luring nature of Perls’s work, it historically has
needs, and shifted his or her view of the signifi- failed to present either a coherent theoretical
cant other (e.g., either holding the other account- model or data supporting it as an empirically vali-
able or having more understanding of the other’s dated treatment. This flaw in theory and research
point of view; Beutler, Engle, Oro-Beutler, & has had detrimental consequences. Today, gestalt
Daldrup, 1986). journals, gestalt therapists, and Perls’s only text
A small number of studies support the efficacy on gestalt theory are nearly impossible to find
of empty-chair work for unfinished business. and relatively few graduate-level institutions
Beutler and colleagues constructed a biological have a systematic method for teaching gestalt
model that suggested that the inability to express psychotherapy in their curriculum.
intense affects puts people at risk for both depres- Gestalt therapy remains, at its very best, mar-
sion and chronic pain (Beutler et al., 1986). They ginalized in the eyes of practitioners of other
found empty-chair work for anger led to the re- theoretically eloquent and empirically validated
duction of anger and a decrease in subjectively treatment protocols. With Perls’s death, there was
reported physical pain (Beutler et al., 1986) and a loss of valuable techniques that are indeed use-
that empty chair work was as effective as an edu- ful psychotherapeutic interventions. Without
cation group for the reduction of depression and Perls’s continuous advocation for gestalt therapy,
chronic pain (Beutler, 1988, as cited in Green- there were few individuals able to propagate and
berg et al., 1994). teach the effective aspects of gestalt therapy,
Conoley, Conoley, McConnell, and Kimzey namely, a pinch of the eclectic mix of various

188
 Gestalt Therapy

intellectuals (the analysts, the existentialists, the psychotherapy (pp. 123–139). Oxford, United Kingdom:
Zen Buddhists, the theologians), and a heavy Oxford University Press.
FAGAN, J., LAUVER, D., SMITH, S., DELOACH, S., KATZ, M., &
dose of Perls—his character, his intellect, and WOOD, E. (1974). Critical incidents in the empty chair. The
eccentric personality style. Without Perls’s cre- Counseling Psychologist, 4, 33–42.
ative, self-righteous certitude propagating gestalt GREENBERG, L., ELLIOTT, R., & LIETAER, G. (1994). Research
ideology, other investigators have, out of neces- on experiential therapies. In A. Bergin & S. Garfield,
Handbook of psychotherapy and behavior change (pp.
sity and perhaps faith, turned to the realm of em- 509–539). New York: Wiley.
pirical investigation. GREENBERG, L. S., & RICE, L. N. (1981). The specific effects
Fortunately, some aspects of recent empirical of a Gestalt intervention. Psychotherapy: Theory, Re-
research suggest that Perls’s distinctive blend of search, and Practice, 18, 31–37.
charisma, arrogance, and creative, yet haphazard, GREENBERG, L. S., & RICE, L. N. (1997). Humanistic ap-
proaches to psychotherapy. In P. L. Wachtel & S. B.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

style did yield enduringly valuable innovations in Messer (Eds.), Theories of psychotherapy: Origins and
This document is copyrighted by the American Psychological Association or one of its allied publishers.

technique. When used, as described above, in evolution (pp. 97–129). Washington, DC: American Psy-
conjunction with relational approaches, and with chological Association.
due respect for the power of these techniques, GREENBERG, L. S., RICE, L. N., & ELLIOTT, R. (1993). Facil-
itating emotional change: The moment-by-moment process.
therapists and researchers learn that the gestalt New York: Guilford Press.
work Perls did with and for his clients, despite his JOHNSON, S. M., & GREENBERG, L. S. (1985). Differential
many flaws, was efficacious and has promise for effects of experiential and problem-solving interventions in
even the most skeptical, evidence-based, contem- resolving marital conflict. Journal of Consulting and Clini-
porary clinicians. cal Psychology, 53, 175–184.
K OGAN , J. (1976). The genesis of gestalt therapy. In
C. Hatcher and P. Himelstein (Eds.), The handbook of ge-
References stalt therapy (pp. 235–258). New York: Aronson.
MILLER, M. V. (1974). New introduction to gestalt therapy
ALEKSANDROV, A. A. (1997). Combining psychodynamic and verbatim. The Gestalt Journal, 12(1), 5–24.
phenomenological approaches in psychotherapy: Ways to MILLER, M. V. (1994). Introduction to Gestalt therapy. In
improve effectiveness. International Journal of Mental F. S. Perls, R. F. Hefferline, & P. Goodman (Eds.), Gestalt
Health, 26, 21–29. therapy: Excitement and growth in the human personality.
BEUTLER, L. E., ENGLE, D., ORO-BEUTLER, M. E., & DALDRUP, New York: Gestalt Journal Press.
R. (1986). Inability to express intense affect: A common PAIVIO, S., & GREENBERG, L. S. (1995). Resolving “unfinished
link between depression and pain? Journal of Consulting business”: Efficacy of experiential therapy using
and Clinical Psychology, 54, 752–759. empty chair dialogue. Journal of Consulting and Clinical
CLARKE, K. M., & GREENBERG, L. S. (1986). Differential ef- Psychology, 63, 419–425.
fects of the Gestalt two-chair intervention and problem PERLS, F. (1968). Gestalt therapy verbatim. Moab, UT: Real
solving in resolving decisional conflict. Journal of Coun- People Press.
PERLS, F., HEFFERLINE, R. F., & GOODMAN, P. (1951). Gestalt
seling Psychology, 33(1), 11–15.
therapy: Excitement and growth in the human personality.
CLARKE, K. M., & GREENBERG, L. S. (1988). Clinical research New York: Gestalt Journal Press.
on Gestalt methods. In F. N. Watts (Ed.), New develop- SALTZMAN, N. (1989). Integrating intensely emotional meth-
ments in clinical psychology (Vol. II, pp. 5–19). New York: ods with psychodynamic, gestalt, cognitive and behavioral
Wiley and the British Psychological Society. therapeutic elements: I. Emotional freedom versus emo-
CONOLEY, C. W., CONOLEY, J. C., MCCONNELL, J. A., & tional control. Psychotherapy in Private Practice, 7(1),
KIMZEY, C. E. (1983). The effect of the ABC’s of rational 57–67.
emotive therapy and the empty chair technique of Gestalt SIMKIN, J. S. (1976). The development of Gestalt therapy. In
therapy on anger reduction. Psychotherapy: Theory, Re- Hatcher & Himelstein (Eds.), The handbook of Gestalt
search, and Practice, 20, 112–117. therapy (pp. 223–234). New York: Aronson.
DALDRUP, R. J., BEUTLER, L. E., ENGLE, D., & GREENBERG, WHEELER, G. (1991). Gestalt reconsidered: A new approach
L. S. (1988). Focused expressive psychotherapy: Freeing to contact and resistance. New York: Gestalt Institute of
the over controlled patient. New York: Guilford Press. Cleveland Press.
ELLIOTT, R., & GREENBERG, L. S. (1995). Experiential therapy YONTEF, G. M., & SIMKIN, J. S. (1989). Gestalt therapy. In
in practice: The process-experiential approach. In B. Bon- R. J. Corsini & D. Wedding (Eds.), Current psychothera-
gar & L. E. Beutler (Eds.), Comprehensive textbook of pies (pp. 323–361). New York: Peacock.

189